Podcasts about cloud code

Keith talks with data-driven investor Neal Bawa, the "mad scientist of multifamily," about why apartment values have dropped 20%–30% while single-family prices have stayed resilient.  They break down how interest rate shocks, the homeowner lock-in effect, and a wave of new multifamily supply are reshaping returns for today's investors.  Keith and Neal also dissect the build-to-rent model—who it really serves, how apartment oversupply is pressuring its rents, and why pending legislation could upend the space.  Neal closes with a specific, data-backed timeline for when multifamily rents and values may finally turn the corner, giving listeners a concrete roadmap instead of vague market guesses. Resources: Grocapitus Website - https://www.grocapitus.com Multifamily U's Free eBook: Location Magic - https://multifamilyu.com/lp/location-magic-ebook/ Multifamily U's Investor Club – https://multifamilyu.com/club Episode Page: GetRichEducation.com/609 For access to properties or free help with a GRE Investment Coach, start here: GREmarketplace.com GRE Free Investment Coaching: GREinvestmentcoach.com Get mortgage loans for investment property: RidgeLendingGroup.com or call 855-74-RIDGE  or e-mail: info@RidgeLendingGroup.com Invest with Freedom Family Investments.  For predictable 10-12% quarterly returns, visit FreedomFamilyInvestments.com/GRE or text  FAMILY to 66866  Unlock truly passive real estate income—visit flockhomes.com/GRE today to see if your properties qualify for a 721 exchange with Flock Homes. To get in the best physical, mental, and professional shape of your life, go to DanielThomasHind.com and apply for Daniel's intensive 1-on-1 coaching for burnt-out entrepreneurs and executives. Will you please leave a review for the show? I'd be grateful. Search "how to leave an Apple Podcasts review"  For advertising inquiries, visit: GetRichEducation.com/ad Best Financial Education: GetRichEducation.com Get our wealth-building newsletter free— GREletter.com  Our YouTube Channel: www.youtube.com/c/GetRichEducation Follow us on Instagram: @getricheducation Complete episode transcript:   Keith Weinhold  0:00   Keith, welcome to GRE. I'm your host, Keith Weinhold. The single-family real estate market is steady, but with apartment building values down 20 to 30% since 2022 when will the multifamily Armageddon end? We ask our qualified guest, and how will slowing birth rates in immigration affect real estate? And more today on Get Rich Education. You know, Mid South Home Buyers, that top Memphis turnkey provider. I learned that a secret weapon behind their explosive growth is more than just you buying their properties, it's an executive coach for nine years now, their CEO, Terry Kerr, and his COO, Pat Nix, have worked privately with a coach who I've now learned from too, and he doesn't market himself online anywhere. After 12 years behind the scenes, that coach is now making himself available exclusively for GRE listeners. His name is Daniel Thomas Hind. If you're a hard-charging business owner or investor who wants to get in the best shape of your life, physically, mentally, and professionally, you can fill out an application for a free consult. This is private one on one coaching for those willing to go to uncommon lengths to achieve uncommon results. Thanks to Daniel, we've all become better leaders, better operators, and better men. It started by showing up for ourselves. Now it's your turn. Go to Daniel Thomas hind.com H I N D, that's Daniel Thomas hind.com and sign up before Spotsville Flock homes helps multifamily owners exit the operator grind, whether it's your six plex or a 50 unit apartment, through a 721 exchange. This defers your capital gains tax. It's a strategy long used by institutions. Now you can swap tenants and toilets for passive income and zero management. Request your initial valuations. See if your property qualifies at flockhomes.com/gre That's F L O C K homes dot com slash G R E.   Neal Bawa  2:13   You're listening to the show that has created more financial freedom than nearly any show in the world. This is Get Rich Education.   Keith Weinhold  2:29   Welcome to GRE from Valencia, Spain to Valencia, California, and across 188 nations worldwide. America's favorite shaved mammal on a microphone is back with you for another wealth building week. I'm Keith Weinhold, and you're listening to Get Rich Education. The world's biggest problems are the world's biggest businesses. That's not a coincidence, and that's why we discuss housing here. And there's been a chronic shortage of affordable housing last month at a commencement speech, Harrison Ford, yes, the guy that played both Han Solo and Indiana Jones, talked about how a fulfilling life has both passion and purpose. Passion is what gets you out of bed in the morning, purpose is what helps you sleep at night, you and I. We can bring this mindset to our lifestyle, to the business we do, and to our investing. Treating tenants well is what helps real estate investors sleep well at night. While we're doing well, we can be doing good too. Multifamily syndicators keep failing, going out of business, and losing all of their investors' money due to mortgage rate resets. It just keeps happening. What this really means, that these groups that pooled together investor money to buy apartment buildings, largely that were set up in 2022 and earlier keep blowing up almost fully due to the fact that interest rates reset higher. Some of them had a fixed rate for five years. Well, rates spiked four years ago, and that's why a lot of them have yet to blow up, and these apartments have lost so much value that no one will refinance them, you know. Even if that apartment operator increased the net operating income over the years, even if rents went up, it doesn't matter. So, you still haven't heard the last of it. Do you remember a couple years ago, when a lot of people in the apartment space, they were saying just stay alive till 25 and that nonsense, like if you keep your head above water until 2025 oh well, then rates are certainly going to fall, and everyone's going to be okay. Well, 2025 is long gone.    Keith Weinhold  5:01   Mortgage rates haven't fallen in any significant way, so that survive until 25 thing or whatever mantra derivative people used that was a farce, like I've said on the show here for years. You cannot predict interest rates, so I didn't make the call that they were going to go up or down at all, because you can't predict them, but so many people said, oh, rates will fall substantially by now, no way, you just can't make that assumption, you've got to take history over hunches, and all of that, a lot of those multifamily deals 100% depended. depended on refinancing at favorable rates, and that's exactly why they failed. A surefire way to look foolish is to predict interest rates. We'll talk more about the multifamily Armageddon with today's guest. I also want to get into what's called the 21st century road to housing act, because that became one of the most hotly debated housing policy provisions this year. And what this is, is a Senate bill, and it would require certain large institutional investors that develop these bills to rent single family communities. It would force them to sell those homes to individual buyers within seven years. So, in other words, what a big firm could do is build a neighborhood of rental homes, lease them for up to seven years, but they couldn't hold on to them any longer than that. They couldn't hold them indefinitely as rentals, this bill is not aimed at you, the individual investor. It is aimed at big institutions, and what I mean by that is that's generally defined as owning 350 or more homes. That's what we're talking about here. Small landlords and mom and pop investors are not the target, it targets corporate portfolios, and this means groups whose names you've probably heard of, like Blackstone, First Key Homes, Progress Residential, and Invitation Homes. They are some of the heavyweights that the government is looking to clamp down on, so whenever you hear someone talk about big Wall Street landlords, that is who they're talking about. Now, some groups are pretty worried about the 21st Century Road to Housing Act, like the NHB, that's the National Association of Home Builders, and a lot of multifamily groups are concerned, and why is that? Well, the effect is it could dramatically reduce new housing production.   Keith Weinhold  7:44   See, a big institution like First Key Homes or Blackstone, they wouldn't want to even get into this business anymore. They wouldn't want to build big build to rent communities anymore if they have to sell them all within seven years. See, they want to buy and hold for the long term, kind of like what you and I are doing, because you and I know that owning a group of selective buy and hold single family rentals is a really profitable place to be, but so if they don't want to build, then that creates a reduction in supply, which could make prices go up, and then obviously hurt those trying to afford their own home. Well, that would defeat the purpose of this whole thing. I mean, my gosh, this always seems to happen when government gets involved. So, the 21st Century Road to Housing Act could limit supply, which is the exact opposite of its intent to get first-time home buyers into their first home, and if this passes, it does have bipartisan support. This lower supply, then yes, indeed puts upward pressure on prices. Just amazing. So then it could actually go on to help the everyday mom and pop investor, like you and I, that already owns property, the individual at last check, though they're looking to pass a version that still restricts some of these giant institutions from getting into build to rents, but yet it does not have that seven year sale requirement. What's really important to remember here is that Washington, they're looking to stifle big Wall Street players from the rental market, which could reduce supply. They're not targeting individual investors. The context that's important is that these groups, they own 10s of 1000s of homes, they don't own hundreds of 1000s, and they don't own a million, so it's a really small percentage of the housing market, whatever direction policy breaks, then the headlines that it creates are just greater in magnitude than the effect on the market is. It's an important frame of reference here. Let's meet this week's guest. This week we're welcoming back a guest that we haven't heard from in a year or two in real estate circles. He is popularly known as the mad scientist of multifamily. He's quite an in-demand speaker. He has a $500 million multifamily portfolio that he essentially shares with over 1300 investors. He's sharp, a good educator, and a straight shooter. That's why he's here. It's a warm welcome back to Neal Bawa.   Neal Bawa  10:32   Thanks for having me on the show again. It's delightful to be here, and so many interesting things to talk about in the world these days.   Keith Weinhold  10:38   There really are.. I don't know if we can get it all in, Bawa is spelled B A W A. Neal, I want to get to your future housing market outlook later. How you think the future looks, including when multi families quasi Armageddon might end. But first, you're known as a data driven real estate guy. Tell us about that, and how being data driven makes you profitable.   Neal Bawa  11:03   I see concern, and I'll tell you why. The single family and multifamily market have been atrociously incredibly divergent since the first quarter of 2022 They have not tracked yet each other at all, even though if you look at the last 50 years, they tend to track each other. So you know, 2008 was a Armageddon for single family, Armageddon for multifamily, and they both sort of came up in 2012 2013 and then they had a really good time until Covid.   Keith Weinhold  11:30   Yeah,   Neal Bawa  11:31   but the second quarter of 2022 is when Fed started raising rates, and since then we've sort of slid - multifamily has gone down in terms of pricing between 20 and 30% depending upon the metro, you know, and depending upon whether it's new construction, new construction assets have gone down more than 30% and existing assets that are filled up have gone down by 20 to 30% depending upon the metro. So, metros that have a large amount of supply, closer to 30% decline in value, the metros that have less supply probably closer to 20% decline in value, right.   Keith Weinhold  12:03   Demand demand has been pretty resilient. It's more of a supply story.   Neal Bawa  12:06   It's a huge supply story, right. So, if you look at, you know, occupancy, essentially what's happened is there was so much supply that came in that really people started on those projects in 2022 maybe they didn't start a construction until 2023 they didn't finish construction until 2025 so they started leasing up in 2025 They had to give offer concessions two months, sometimes three months free, and so that pushed down the rents in 2025. And they're not done, because you typically can't rent an apartment in six months. If it's brand new, it's going to take you about 18 months to rent it, and sometimes 24 months, and so it's affected our rents in 2025 it's affecting our rents in 2026. Now it's unlikely to affect it in 2027 but we'll go there, you know, at a later stage. But at the moment, we, what we've seen is negative rent growth in the United States for multifamily for the last 12 to 15 months, and what I think is going to be negative rent growth in Q of this year and Q2 of this year, so Q1 was negative, Q2, which we are in now, is likely to be negative or flat now. Single family, on the other hand, has gone in a different direction, which has been very difficult to understand, and I believe it's taken me a while to really understand this, but I think I've finally figured it out. Single family prices are not down since 2022 which makes no sense at all, because the average mortgage in the United States today is almost double, almost double, not quite double, but almost double of what it was in at the beginning of 2022 when interest rates were about 3.3 3.4% Right now we're sitting around, you know, six and a half percent interest rates, so not quite doubled interest rates, but they've obviously gone up a fair bit, and as a result, your average, you know, mortgage has almost doubled, but home prices haven't dropped, which makes no sense if you really think about it, because home prices are a factor of demand, and they're also a factor of people's ability to pay, so if all of a sudden within four years you're paying, the mortgage is doubled, then less people are going to be able to buy, but it stayed up, the market has stayed up, and the biggest reason it stayed up is because of what is known as the lock-in effect. So, the US market typically has a million new homes every year, and there's more than a million existing homes that are transacted, right? So, it's an open market, it's a perfect competition market, but it hasn't been perfect competition for the last four years, because so many people locked in ridiculously low interest rates.    Neal Bawa  14:28   Perfect example, in 2021 and 2022 I have a 15 year mortgage at 1.75% If I sell my house back to myself, my mortgage quadruples, quadruples, right, because it goes from 1.75% to six and a half percent, so I can't even imagine even think about leaving my home, right, because it's just such a perfect loan. Most people don't have anywhere near 1.75% but there's lots of people with more mortgages in the 3% three and a half percent, and 4% range that basically can't go anywhere, and because those homes are not coming into the market. The last three years the market has had this unusual not enough supply factor, and that's been keeping prices up. That is ending. That is ending, because what we've been tracking is the percentage of homes in the United States that have low mortgages. Low is simply defined as anything under four and a half percent, and that percentage is going down each quarter, because you know divorces happen, deaths happen, you know people move for jobs, and so every time that happens, that locked in rate goes away, because you sell your home and move on, and so for a while that lock in effect was predominant, it was controlling everything, but as time has gone on, interest rates were higher in 2324 2526 For also almost four years have passed since the rate started going up. So each quarter the percentage of homes in the US that have these low interest rates has slowly moved down, and we're almost back to a normal timeframe.   Neal Bawa  15:53   And this is causing the single family market to not have a conniption, but we're starting to see a balancing of the market, where it's not just a buyer's market anymore, in some places it's actually seller's market, some places it's a buyer's market. So we're now starting to see home prices drop in number of markets in the United States. I can't say that they've dropped in super majors, but we're seeing a flattening out effect of home prices in most metros in the US, and there should be a flattening effect. Just to be blunt, I mean, obviously I own a bunch of single-family homes, so I just wanted them to keep going up for selfish reasons. But if you think about it, we had huge home price growth in like 30 plus percent in number of years, 2021 22 and even 23 and during those years, salaries only went up by two to 3% a year. In one year, they went up by 4% and rents also went up like crazy. There was a 2021 was 15% rent growth year. So, at some point, there had to be an adjustment, and we are in that period of adjustment where single family prices are basically flat on a national basis. Yes, going up in the San Francisco Bay Area because of AI, and going up in a couple other technology-heavy metros because of AI, but otherwise fairly flat, and I don't expect that to change for the next year. So, my forecast is next 12 to 18 months, home prices in the US are going to be flat on a nominal basis, they're going to be down on an inflation-adjusted basis, but you know, because of the Iran, more inflation's three and a half percent, so home prices should go up three and a half percent. So, if they stay where they are, well, they're really dropping three and a half percent.   Keith Weinhold  17:29   Yeah, before this year began, I released our forecast, it was for 2% nominal home price appreciation in the one to four unit space for the US this year, and I still like how that looks. There's so much to unpack with what you just talked about. In my view, there's nothing unusual at all that when mortgage rates rose sharply a few years ago, that home prices rose as well. Why? Because actually, that's what usually happens, which is counterintuitive to most people. In all of our lifetimes, residential real estate prices have only fallen significantly one time, that was around 2008 due to a number of unusual circumstances. The only thing that's a bit different this time is, of course, how fast rates increased in 2022 and 2023 and people wondering if residential real estate prices could still keep up, and they certainly have, but yeah, you brought up this dichotomy, this bifurcation about how the apartment market and the one to four unit space kind of separated from each other in 2022 or 2023 That's what's so interesting.   Neal Bawa  18:36   I do want to point out a couple things, though, and I don't want to be a Pollyanna here and talk about negative stuff, but I think that there's big difference between 2008 and that timeframe and where we are today, and that difference is, and it has multiple parts. Not all of your audience is aware of this. Until about 2012 the United States had very reasonable birth rates. You know, we were one of those countries that had avoided the debacle that Japan, Korea, China, and a number of other countries are seeing South Korea being the absolute worst, where basically they were producing one baby per generation, where you need about 2.2 babies just to kind of keep your population where it is, right, and the US was unusually high in that, and that we were still above that threshold, which meant that our population would continue to grow and not fall. Now, there was two reasons our population was growing: One, we had more than 2.2 babies per household, and second, we had a very significant amount of legal and a very significant amount of illegal or undocumented immigration. Right, so we had both of those pipelines today. All three of those have flipped, so the United States now basically looks like Korea or China or Japan in that every household is producing about one and a half babies, which means that our population growth, which hasn't stopped yet, because it takes a while for these things to catch. Up is likely to stop, like it's, and at some point decline again. Luckily, we're not there yet. The US is a fairly young population, unlike Japan, which is one of the oldest populations in the world. So, it'll, we'll still continue to see population growth, but there is no doubt. And you can ask Chat GPT, right? How has population growth in the United States slowed over the last 20 years.    Neal Bawa  19:22   Make me a graph, and it will make you a very nice graph, and you'll very clearly see there's a slowdown in population growth. The second part is both documented and undocumented immigration. It's my estimate that since this administration took over, somewhere between half 1,000,001 million people have left the United States. Now it's very difficult to get an actual number, as you can imagine. A number of these people were undocumented, so we didn't really know how many there were to begin with. And a number of them, when they left, they also left by an undocumented rate, that you know, path. So we've lost a bunch of those people, and also the people that have stayed in the country, we've lost a number of them in the workforce. Here's a perfect anecdote, Keith. About 33% of the construction workforce in the United States was undocumented, one in three. In Texas, as much as 40%   Keith Weinhold  19:45   Yeah, that's huge.   Neal Bawa  19:45   It's very significant. Number of those people don't show up for work anymore. I don't think they've left the US, at least I don't think so. But they don't show up for work anymore, because that's how they get caught, right. So, what we've seen is that the construction workforce in the United States has become been decimated over the last 12 months, and the impact is much greater in the second half of 2025 than the first half. Why? Because even though they wanted to do ICE enforcement, they just simply didn't have enough agents, enough facilities, enough judges. When the second half of last year, they sort of started catching up on that, hiring more agents, getting more facilities, getting more judges, and so we started to see a real challenge there. I have properties in 10 markets in the US, and what I can say is about seven of those markets, mostly Southern markets, I am beginning to see dropping occupancy related to this phenomenon. I'm seeing a reduction, and so markets like Georgia and Texas, Florida are more hit than my northern markets like Idaho. I haven't seen any impact at all, but these southern markets, multiple properties, multiple metros, I'm seeing this - people, mostly of Spanish, Mexican origin, not renewing leases. I don't know what they're doing. I don't know if they're sleeping in their cars. I don't know if they're basically just, you know, staying with mom or staying with, you know, some other family. But I'm seeing a very, very big pullback in my leases tied to this, and occupancy is dropping in those markets that are heavily Hispanic. And so I'm seeing the impact of that on landlords, but I also know that there's an impact on the US at all, and overall demand on rentals, whether it's single family or multifamily. This is a significant impact, because I don't think that the Republicans are going to make a U-turn on this. I don't want to get political, but you know, stating the obvious.   Keith Weinhold  19:45   Yes, United States had its biggest birth year in 2007 when there were more than 4 million babies born. The average age of the first time homebuyer today is 40 years old. If that holds true, that peak would take place in 2047 And then, yes, to your point about changes in immigration, yes, it sounds like a potentially a reduction in demand with what you're talking about, with some vacancies, and also maybe a reduction in supply when you have fewer construction workers to build these places as well, we're talking about building properties. Neal, I want to talk to you about the build to rent space. Somewhat is build to rent better than traditional real estate? I think that's what we really want to know. And for those that don't know, build to rent means when you construct a property where from day one that construction project is built for a tenant, not an owner occupant. I see a lot of pros and cons there. Can you talk to us about the trade-offs between build to rent and traditional real estate?   Neal Bawa  19:52   Yeah, if you think about it, it's a really terrible word, built to rent, because if you think about the word built to rent should be apartments, right, but actually doesn't mean apartments, right? So, built to rent actually means single family or town homes that were built to rent out, right? And then you're like, why don't they just said built to rent apartments and town homes? Well, you know, was too long an acronym, and we suck at acronyms anyway. But BTR, or built to rent, is essentially building single family or town homes, but specifically building them to rent, and it doesn't include any apartments at all, right? And the reason why the BTR market was growing in the last five or six years is that roughly 18 million American families can no longer afford to buy starter single family homes, you know, and by starter I mean, small old single-family homes. That's how Americans usually started, you know, in their 20s and 30s. They would buy these homes, some of them, but they would fix up, and then they over time, in their 30s, late 30s and 40s and 50s, they would upgrade, and then at starting the 50s, it would flatten out, and then the 60s, they would start to downgrade, right? That's been a typical thing that's happened in America for 56 5070, years. Well, that is, cannot happen anymore. And it broke in 2022 until 2022 It was a normal cycle beyond 2022 because interest rates almost doubled, and the mortgages almost doubled, but the incomes only increased by 10 to 20% There became this orphaned generation of Americans, roughly 18 million families, that simply cannot afford to buy that starter home, and they are now forever renters. They don't know it. They think that they're going to catch up at some point, but five minutes with an Excel spreadsheet, I could prove it to them that they're not going to catch up.    Neal Bawa  25:35   Maybe one in 100 families would see a very large increase in income, and that would result in them catching up, but for the most part, as a group, these 18 million families, they're forever enters as a group that didn't exist before 2021 right. It's entirely because of this outrageous increase in mortgages, while not seeing a drop in home prices, that led to this, and so those orphan families, they actually earn pretty well, so these are families that make 70, 80, $90,000 in mid markets. They make over $100,000 if they're living on the coasts or in expensive markets, and they still can't buy that, you know, starter home. And so they don't want to live in apartments. I have lots of apartments, old ones, new ones, and I want these people to live there, but they don't want to live there, and so they've been looking for an option, and that option has been developers like me building communities of 200 300 townhomes or single family homes with a small little yard, and then basically from day one, instead of selling them, renting them out, and then once you're done renting out the whole community with 200 tenants, then you sell that to an apartment company. You know, there's lots of apartment companies in the US that have 100,000 units. Well, they want to buy these because the turnover is lower. So, what happens is most of these town homes and single-family homes for rent. Families come in, and they typically rent for three to five years before they move, whereas in on my apartments I lose 40% of my tenants each year. So, if I have 200 tenants, I lose 80 of them every year, and I have to basically go back, clean up those units, deal with the vacancy. But when I have townhome communities like my Idaho Falls townhome community. I lose a tenant at roughly every four years, and so, as you can imagine, profitability goes up when turnover goes down, right?   Neal Bawa  27:31   Because you don't have that cost of turnover and vacancy, and so eventually those large landlords that are holding 100,000 units figured out, I like this, what Neal Bawa is doing, he's building these 200 townhomes, I want to buy these from him when they're rented. I don't want to build them, I don't want to lease them up, I just want to buy them when they're stabilized. And so BTR became that name for that marketplace where developers would build townhomes and single families, rent them out, and then sell them to institutional, and it was some—   Keith Weinhold  27:56   People think of fabulous institutionalization of the starter home.   Neal Bawa  28:00   And in many ways it is, because what happened is, for a while, these institutional players, like Blackstone and BlackRock, they were like, we are just going to go out and buy 50,000 single-family homes, and that's going to be the institutionalized. Well, that worked really well if you bought in 2008 2009 2010 2011 because you got them bought them at a discount, but when they started buying them in 2015, 16, 17, 18 at ever higher prices, they didn't make any money. So the vast majority of these public funds that were created to buy large amounts of single family have failed if they've purchased anything in the last seven or eight years. If they bought before that, they made huge amounts of money. Family homes are so expensive that basically buying them for rental did not make sense, so these companies have now pivoted to saying we'll only buy communities that have 100 or 200 or 300 of these homes, because then we get the benefits of having centralized leasing, centralized property management, centralized maintenance, and I don't have homes spread all over the metro, they're all in one place, and I can make more profit from that. In theory, that's been good, and you might think that I'm bullish on BTR, but I'm actually today bearish on BTR for one single reason. About seven months ago, Republicans started talking about a bill - I don't know what the name of the bill is, but what this bill does is it forces builds to rent developers like me within seven years of building the property to sell all of the homes in that property to single family tenants, not to Blackstone, not to Blackrock, but to single family tenants. Hasn't passed yet, but it passed the Senate with an 8910 vote, which means that both Democrats and Republicans wanted to vote for this. If it passes the House, and because Donald Trump himself is very heavily opposed to it, he's made it very clear he doesn't like this. He's a developer, obviously. It hasn't passed the House yet, but if it passes the house, that will destroy the build to rent market. No one will ever build build to rent, because the worst possible thing is I build this, and within seven years I have to actually sell it to individual buyers. If I do that, my banks are going to hate me and not give me loans to build BTR anymore. Obviously, there's going to be some grandfathering to the communities that I'm building now, or maybe even build the ones that I'm building in 2027 maybe grandfathered. It usually is, because you know, Congress never does anything retroactively, and they give you a year or two, but if it passes, it's doomsday for BTR. I hope it doesn't happen, but that's the way it's looking, because it's bipartisan. Bipartisan bills are more likely to pass   Keith Weinhold  30:40   Now for the mom and pop investor, the individual investor build to rents have obvious appeal due to your point about the lower turnover, lower maintenance costs on a new build, lower insurance costs often on a new build, and then there's the tenant appeal to a new build as well, but of course there is that investor downside. I think a lot of investors are aware of their thin initial cash flow that they're going to have on build to rent, but you know, Neal, another downside with build to rent, I think a lot of investors don't look at is, hey, just how many of these things are they building? Are they building 500 of them? Do I have some overbuild risk if I buy into this community that could suppress occupancy and rents for a while.   Neal Bawa  31:21   What we've seen is that when Built to Rent started out in 2017-2018 it was its own asset class. It wasn't competing with apartments, it wasn't competing with single family rentals, it was just its own thing. However, in the last two or three years, as more and more apartments flooded the marketplace, we had a glut. It moved away from that. It basically started getting affected, and the rent started falling, just like any other portion of the market. You know, think of it as three portions of market. There's the built to rent, which I described, you know, brand new single family homes, town homes per rent. There's the apartments, both brand new and existing, and there's the single family rentals, right, which there are millions of. What we are seeing now is it's become one market, right? All of them are affecting each other, and the apartments, which have a huge amount of glut, there's a massive amount of new apartments that have come in in the last two years, are really pushing the rents down for single family, they're pushing that rents down for BTR. So, at this point, what I would say to people that have this concern, Keith, is simply look at incoming apartment supply, because if you're in a marketplace, and I'll give you examples of really good markets that are crushed right now. If you're in a market that has a lot of incoming supply, whether you buy a single family rental, a quadplex, a 50 plex that's an apartment, or 100 unit BTR, you're going to suffer for rent growth if you have a lot of incoming supply in 2026 and that is across the board in every market in the US. Huntsville, Alabama is, in my opinion, one of the most interesting markets in the US for 5 year, 10 year growth, right?    Neal Bawa  32:54   If I had to say you don't need a loan, it's just your own cash, no investors, where would you put money in? It would be at the top of my list, not at the very top. Idaho Falls is definitely the number one market in the US in my list, but Huntsville is up there. But right now, do you know what rent growth in Huntsville is? Minus 2% negative 2% Why? Because there's 6000 units coming into a market that's, you know, 1/5 or 1/10 the size of Phoenix, right. It's 1/10 the size of Dallas, but it has half the units of Dallas or Phoenix coming in, and so rent growth is negative there. So, what I would say is today absolutely everyone that is an investor should understand that we live in the magic world of AI, and you should be talking with Chat GPT about incoming supply for any market that you're interested in, and using that to make your decisions, because all of these markets merged, BTR, new apartments, old apartments, single family, everything has emerged in the last 24 months, where they're all affecting each other, and if there's too much supply of any one kind, it's affecting all of the other markets, and that's the message that I have. And none of this is like you have to go buy a $25,000 software like Costar today. Chat GPT is your costar.   Keith Weinhold  34:11   You're listening to Get Rich Education. We're talking with the mad scientist of multifamily, Neal Bawa, where we come back, including what he thinks about recovery for the beleaguered multifamily market. I'm your host, Keith Weinhold. What if you got your mortgage loans the same place I get mine? You sure can at Ridge Lending Group, NMLS 42056 They provided GRE listeners with more loans than anyone, because Ridge specializes in investment property. They'll help you build a long-term plan for growing your real estate empire with leverage. Start your prequal, and even chat directly with President Caeli Ridge. While it's on your mind, start at ridgelendinggroup.com that's ridgelendinggroup.com    Keith Weinhold  34:56   Let me ask you something: if you've worked hard to build wealth, is your money positioned to actually support your goals? A lot of accredited investors leave capital sitting in cash because it feels safe, but inflation and missed income opportunities can quietly erode its value. Freedom Family Investments offers freedom notes for investors seeking structured income backed by real estate. It's a straightforward approach built on real assets, not speculation. In full disclosure, I'm an investor myself. What I like is that their team walks you through how it all works, so you can decide if it aligns with your portfolio and income goals. Every investment carries risk, and nothing is guaranteed, but with a track record of consistent on-time investor payouts, they built real credibility. Go to freedomfamilyinvestments.com to book a clarity call, or text family 268 66 That's Family 266 866    Speaker 1  36:00   This is the star of the A E Show, The Real Estate Commission. Todd Rollette. Listen to Get Rich Education with my friend Keith Weinhold, and don't quit your daydream.   Keith Weinhold  36:20   Welcome back to Get Rised Education. We're talking with Neal Bawa, a really sharp multifamily syndicator who's also highly data driven. And Neal, tell us more about the beleaguered multifamily market that had those aforementioned problems really cropping up in 2022 and we had a lot of supply and spiking rates. What does it look like for the path to recovery for the US multifamily market?   Neal Bawa  36:45   Luckily, demand is strong, and even though occupancies have dropped, typically the multifamily market, the large multifamily market in the US, tends to be between 95 and 96% occupied. Okay, and right now we're on 93% so that all that incoming supply means that about 7% of our apartments in the US are empty at the moment, we're trying to fill them, and we are seeing that occupancy drop, not across just new apartments that are leasing up, but also drop in class B and class C. We've also seen a huge increase in concessions, so I studied this quite obsessively, and I can tell you that 2026 in some markets is the recovery year, but not across the board in the United States, and the reason for that is sentiment. Once renters get used to huge amounts of concessions, it's like a drug, it takes a little while before you wean those renters off of those drugs, and so there's that hit right now. Every renter program,   Keith Weinhold  37:44   Everyone wants their freebie for good.    Neal Bawa  37:46   Yeah, exactly. It's like, hey, what, you're not giving me two months free? Hey, what, you're not even offering me one month free? It takes a while for that expectation to happen, because there's such a huge amount of concessions in the US. So, to me, there are a few markets, usually the smaller markets or very fast growing markets, where there's a recovery in 2026 but otherwise 2027 The first half of 2027 is recovery. The second half of 2027 is fast rent growth in a lot of markets. Why? Because remember, interest rates have been high since 2023 A lot of projects were started in 2022 went into construction in 23 came to market in 25 and 26 Lease ups are happening in 25 and 26 By early mid 27 these are all leased up, right? The second half of 2027 there isn't a lot of delivery in any of these big markets, because to deliver in the second half of 27 you would have started construction in that second half of 2025 and I counted those permits market by market. There's just not a lot, because by that time everyone knew that projects were not getting funded, everyone knew that interest rates were high, so there wasn't a lot of supply of new starts in the apartment market in the second half of 25 so there's not going to be a lot of delivery in the second half of 27 and all of the existing stuff would have been leased by then. So 2026 is one of those years where we could still see more concessions in the second half of 2026 I still see rent growth for apartments to be flat. You mentioned single family might be a little bit higher. It tends to be a little bit higher than apartments in terms of rent growth, but I think flat rent growth for 2026 is what I'm projecting. I'm projecting small rent growth in the first half of 2027 for most markets, and then I'm projecting robust rent growth, call it 3% or greater on an annualized basis, in the second half of 2027 and I'm projecting that most markets in the US that are not seeing a population drop, so count out places like Detroit are going to see a very aggressive rent growth, four or 5% rent growth, that's aggressive in our world, in 2028 28 and 29 are shaping up to be. Supply deficit years, years where supply is well under demand.   Keith Weinhold  40:05   It's pretty easy to project completions when you just go ahead and look at starts, and really, what you're counting is the story of absorption.   Neal Bawa  40:14   Yep, and what's nice about apartments is you can actually build a single family home in about nine months, right, but you can't build apartments in less than 24 months. There's just so much permitting issues, there's so many delivery issues, fire code issues, and so we have a crystal ball on the multifamily side that we are now getting better at using. I don't think the industry was very good at this in 2022 but now we're really all obsessed with how many permits does my metro have, and how many permits does my state, and how many permits does the US have? And everyone that I know in the industry that's data driven knows that there's a massive glut now, maybe a little bit of a glutton that remaining portion of 2026 equilibrium in 27 and a huge, huge supply deficit in 28 and 29 So everything that I'm doing is based on this, and this crystal ball actually works because of that two year gap between shovels in the ground and delivery,   Keith Weinhold  41:10   and it sounds like you've recommended Chat GPT as a go-to source for investors to look into these things, that happens to be my favorite one as well, and you are well, maybe it's a bit too much to say, but it almost feels like to me pioneering with the way that you use AI. In fact, I know before our show today you were running some other things in the background that made me wonder, hey, am I talking to the real Neil or the clone Neil? I know I've got the real Neil here, but why don't you tell us about how you're using AI to make data-driven decisions in real estate?   Neal Bawa  41:40   Sure, so the first thing is that we've completed our journey with the low hanging fruit of AI. Every single person in our company is fully trained on how to use Chat GPT. Most of our research-related processes are automated. For example, 100% of our investor updates are now written by Chat GPT. What we do is we go into our property manager meetings on Mondays or Tuesdays sit down with them, beat them up, and the transcript is then taken by our team in the Philippines. They take that transcript and put it into a pre-trained Chat GPT string, it's called a custom GPT, and the string took a while to train, but now that it's trained, all it needs is a transcript. We just copy paste it in, we don't give it any instructions, and it outputs a really wonderful investor update, right. And so our updates for our investors are 99% written by AI. Of course, we'll go in and add our comments at the end of the process. So we've automated investor updates, rent comps, so you know if we are underwriting a new property today, what we do is we simply go into a Google file and copy paste the address and hit enter roughly once a minute. A software, which is written by AI - we're not coders, but the software knows how to write code - it checks the file, if it sees a new address, it goes in there, grabs the address, and then it basically goes to apartments.com rent.com realtor.com and all of these places, and checks the rents for this particular property in two mile radius. It eliminates all the ones that don't match, like you don't want to match the rents of a 1970 or 80s built property with a brand new 25 built property. Those are not comps, it's not comparable. So it basically is very careful, it keeps a radius range of two miles, and also basically is a property of the same kind, you know, like it never matches up a three story property with a 10 story property. Those don't match, one of them obviously is more of a central business district or downtown sort of thing, and so it basically grabs all of those rent comps and then puts them into a file and posts in a Slack channel. Usually it takes it about 1213 minutes to do that, and so whoever put that address in about 12 minutes later goes into the Slack channel and says, "Hmm, these are all my rent comps, right? And boom, now you're basically, you have all these ready rent comps. So, what we've done is, we've automated a significant portion of what we are doing with both our property managers and inside the company with acquisitions and things like that, we're also scraping massive amounts of data from the Bureau of Labor Statistics website, which we just couldn't deal with that data before, and building very beautiful, very interactive dashboards. We don't use Chat GPT for that. We find for dashboarding a tool called Claude, which is by a company called Anthropic, is much better, so we have currently over 150 interactive dashboards that Claude has created that update in real time and give us access to data. If anything, I find that we are in this incredible time where decision making has become much easier, as long as you spend time with these tools. So, in our company we have an absolute mandate that no one has broken for the last year. One year per day, people must program, and by programming we mean issuing common language instructions to tools and build dashboards and build software that automates our work. Have we laid off anyone because of this? I mean that. Be the next obvious question. The answer is no, because it's made it easier for us to serve a much larger audience, so it's easier to grow your company. We just are not hiring anyone, and we haven't hired anybody for the last 18 months, so we have a hiring freeze, but at the same time all of our people are employed because they're they're now much more valuable. So everyone in our company is now a programmer, and even though that sounds weird, it's completely true.   Neal Bawa  45:24   Every single person in our company writes code, and they write code by talking with Cloud Code or talking with Chat GPT, and then Chat GPT, of course, does the actual code writing, but people have become very, very good at answering questions and saying, "I want a dashboard like this, turn these radio buttons into drop boxes, and give me the last month, and last three months, and last 12 months, and do this, and do that, and connect this, and I also want to host this on a server, but I want to make sure that only I can see it. I need a password added. Imagine 1000 of these conversations happening in our company every day. Yeah, that's interesting. And what you just described   Keith Weinhold  46:00   there at Gro Capitas is somewhat of a microcosm for what's happening in the broader economy, where we've been in this low high or low fire environment for quite a while. Well, Neal, as we're winding down here, we recently had a new Fed chair come in. It seems incomprehensible to me that there could possibly be any rate cuts. I don't know how we could responsibly make a rate cut with all these inflationary layers. We had the pandemic, and then terrorists, and then the Iran war, and the energy shocks, and all these bottled up supply chains. What are your thoughts with regard to the Fed?   Neal Bawa  46:29   I still think that we'll get one rate cut, and that rate cut will be based on political pressure. So, for the first time ever, I have seen the Fed break into factions, so if you look at the latest Fed meeting, which happened, you know, there was dissent, there were two clear factions, so the Fed is becoming less data driven and more faction driven, and I think that one of the factions, which obviously wants rate cuts to go down, is going to triumph at some point later in the year, but until we get past the incredible increase in inflation because of the Iran war, I don't think that faction is going to win. Right, there's three or four people in that faction, that's not enough votes to get past the others. So I'm predicting no rate cuts until Q4 of this year. If the Fed was entirely logical, there should still not be a rate card in Q4, but I think it'll happen because there's political pressure.   Keith Weinhold  47:25   The preservation of independence is key. Neil Bhawa, this has been great, and a lot of people learn from you. You're a brilliant educator, as well as what you're doing in the multifamily space, and a lot of other places. So, if someone wants to connect with you, learn more about what you do. What's the best way for them to do that?   Neal Bawa  47:43   So we built a website called Multi Family University. It's completely free. There is no subscription. There's no upsell. We do not have an educational product, but what we do is each year we have 8-12 webinars that we create with their extraordinarily good looking thanks to the use of AI. Yay, and we share them with an audience, and usually between 5000 and 1000 people attend our webinars each year, of which roughly 1% become investors with us. The rest, the remaining 99% just continue to get free access to data, and we cover every imaginable real estate topic: Single family, multifamily, industrial hotels, self storage, Airbnb, and even controversial topics outside of real estate, like climate change or impact of climate change and impact of AI. So you know, multifamily university is the best place you can go to, multifamily you.com/club It's a free club, and it's free forever.   Keith Weinhold  48:42   Neal, it's been valuable to our audience. Thanks so much for coming back out of the show.   Neal Bawa  48:46   Thanks for having me.   Keith Weinhold  48:53   Oh, a terrific, wide-ranging chat with Neal. There, yes, this interesting 2022 divergence between single family and multifamily, the slowing birth rate, and how that won't really catch up with real estate in a big way for perhaps 20 plus more years. How single family rentals beat multifamily on the basis of tenant retention, and a lot more that we covered there, and he's got a good data driven timeline for apartments being back in favor by 2027 and 2028 After the interview, Neil and I chatted some more off Mike, and he would like to come back on the show next year. We're probably going to have him, because we have a lot more to talk about at that time. We can see if the multifamily market is really healing. Also, did you pick up on this? I wonder why, for his own home he would get a 15 year mortgage at 1.75% interest, so I'll have to ask him about that. That's surely a fantastic interest rate, but a 15 year loan rather than a 30 year that maybe he could have gotten at two and a half percent at the time. Well, 15 year probably. Is not the best use of capital, because it increases your equity position rapidly. When instead, those dollars could have been out in the market earning an actual return somewhere else. But he's a smart guy, he must have an answer. We can talk about that at that time. We've got a lot of terrific shows coming up here on the GRE podcast, specific learning episodes, where it's just me teaching you, as well as new guests and returning guests too. Until next week, I'm your host, Keith Weinhold. Don't quit your daydream.   Speaker 2  50:35   Nothing on this show should be considered specific personal or professional advice. Please consult an appropriate tax, legal, real estate, financial, or business professional for individualized advice. Opinions of guests are their own. Information is not guaranteed. All investment strategies have the potential for profit or loss. The host is operating on behalf of Get Rich Education LLC exclusively.    Speaker 2  51:03   The preceding program was brought to you by Your Home for Wealth Building, getricheducation.com.  

We're announcing AIEWF speakers this week! Take the AI Engineering Survey!Today's guest Ethan first joined us for the LS Paper Club as the lead on NVIDIA Cosmos World Model, but then joined xAI and built Grok Imagine in 3 months:He comes back on Latent Space with some nuclear hot takes: that Video Models primarily get their intelligence from LLMs, not from training on video data, and that the next frontier for truly interactive, realtime, long-horizon world models is to work on LLMs (perhaps Interaction Models as well…)Put it this way: In the near term, the next Sora won't be a better video model, but a video agent.Generative Media may more closely follow the evolution of AI coding which went from focusing on one-shot output performance and cost, to multiturn reasoning and planning models for agents and systems that can plan, edit, test, debug, and submit PRs.At a certain point, coding models got so good that the only significant next step to improve performance was handling the orchestration of these models.Now as the performance of video models increases significantly across realism, consistency, & prompt adherence while becoming more cost efficient, the next evolution of video generation may also be systems that can plan, generate, edit, critique, and iterate across an entire creative task. In this episode, Ethan joins swyx and Vibhu to unpack what it actually takes to build frontier image and video systems: data, VAEs, diffusion transformers, audio-video alignment, inference speedups, and the hidden cost of storing and moving massive video datasets. From building NVIDIA's Cosmos world model to joining xAI as Grok Imagine was being built from zero to one, Ethan He has been at the center of some of the most important work in video generation, multimodal models, and real-time world models.We go deep on Grok Imagine, how a small xAI team shipped its first multimodal video model in three months, why iteration speed matters more than almost anything in model development, and why many of the biggest gains come from fixing tiny bugs in data and training pipelines. Flipbook: The future of VideomaxxingVideo agents are almost a sure bet to be the trend in the coming year. We end with a glance at what's beyond video agents:Flipbook caused a minor sensation this year when it was released, but most treat it as a fun demo. Ethan takes it very seriously — with the speed and cost of inference coming down every year, the future of custom video JIT UI is closer than you think. We talked about why videogen models may become the front end of AI, how generative UI could replace traditional HTML/CSS, why world models need to be real-time, interactive, and long-horizon, and why the future of video generation may depend more on language models and agents than on diffusion alone.We discuss:* Why fast iteration mattered more than meetings* Why small training bugs can drive huge model quality gains* Why coding models may make compute the bottleneck again* How image and video models are trained with synthetic captions* The role of VAEs and latent space in frontier video models* Why image models are the foundation for video models* The tradeoff between temporal compression and real-time interactivity* Flipbook, Neural OS, and the future of generative UI* Why future interfaces may go from user intent to pixels* The hidden cost of training video models: storage, egress, and GPU hours* How step distillation and consistency models (like OpenAI sCM) makes video inference orders of magnitude faster* Grok Imagine 0.9 and large-scale audio-video generation* Why audio-video alignment is harder than text-video alignment* Ethan's definition of world models* Reference-to-video, video extension, and long-context video generation* Why xAI's research communication undersells Grok Imagine* How xAI culture shaped the speed of development* AI watermarking, SynthID, and detecting generated media* Why prompt rewriting matters for video models* Grok Imagine Agent and the rise of video agents* Why language models may unlock better video generation* Robotics, physical AI, and embodied world models* Why Ethan left xAI and shifted focus toward LLMs* Self-managed context, memory, and the next frontier for language modelsEthan He* LinkedIn: https://www.linkedin.com/in/ethanhe42* X: https://x.com/EthanHe_42Timestamps00:00:00 Introduction00:01:25 From NVIDIA Cosmos to xAI00:03:24 Building Grok Imagine from Zero to One00:10:07 How Image and Video Models Are Trained00:18:53 Video Compression, VAEs, and Real-Time Tradeoffs00:22:10 Generative UI, Flipbook, and Neural OS00:32:10 The Cost of Training Large Video Models00:37:04 Distillation, GANs, and Fast Video Inference00:41:21 Audio-Video Generation and Grok Imagine 0.900:48:34 What Makes a World Model?00:55:51 Reference Videos, Long Context, and Video Memory01:00:11 xAI Culture, Research, and First-Principles Building01:09:45 AI Safety, Watermarking, and Prompt Rewriting01:13:10 Video Agents and AI-Assisted Creation01:27:32 Why Language Models Unlock Better Video01:31:15 Robotics, Physical AI, and Embodied World Models01:32:38 Why Ethan Left xAI01:34:16 Self-Managed Context and the Future of LLMs01:38:43 Ethan's Career Path and Closing ThoughtsTranscriptIntroduction: Ethan He, Latent Space, and the Path to xAISwyx [00:00:00]: We're here in the studio with Ethan He, most recently of xAI. Welcome.Ethan [00:00:10]: Thank you. Glad being here.Swyx [00:00:11]: We're also here with Vibhu. you were first coming to us or joining the latent space world because you were working on Kosmos at NVIDIA, and you did a paper. We loved it. you presented it as well, so thank you for doing that.Ethan [00:00:23]: I've actually, I also presented the MoEs twice at latent space.Swyx [00:00:29]: How did you actually hear about us? Did we reach out to you? Is that how it worked?Ethan [00:00:33]: No, actually, I-- the community. Like I realized, oh, there is this online community that people talk about AI and also learn from each other through papers every week through the Paperclip. It's very nice.Ethan [00:00:49]: I learned a lot.Swyx [00:00:49]: I think three years stop. We haven't stopped even on Christmas and New Years. many weeks I want to stop but it keeps going.Vibhu [00:00:58]: No, that was good. I think you had posted that you worked on a paper, and I was “Oh, very cool. We have Paperclip. Present then.”Vibhu [00:01:04]: But I might have reached out to you after.Swyx [00:01:05]: you-- because it's an amateur club, right?Swyx [00:01:08]: so it's very unusual and but we have sometimes paper authors come by and actually explain the paper. Today we just did, the poolside paper, which was apparently very good.Vibhu [00:01:18]: Came out yesterday.Vibhu [00:01:19]: pretty interesting, right? Fully open. They talk about everything, systems. So it's a good one. We'll, we'll recommend people to read it.Swyx [00:01:25]: Bring us up to speed on your transition to xAI, ‘cause I actually don't even know when you joined. just like tell the, tell the story about the sort of transition.From NVIDIA Cosmos to xAI: Scaling Video and World ModelsEthan [00:01:34]: Before xAI, I was working on Kosmos world model as in-- at NVIDIA. So Kosmos is, it's a giant video foundation models that can-- that aims to simulate the world and for-- it serves as a foundation of-- for all of the roboticists to build on top of. There, once I built the Kosmos one, I realized as this thing also has a scaling law similar to language model, we need to scale up the video models further. that's, that's why I realized I need to move to somewhere with much more compute resources. That's how ISwyx [00:02:13]: Than NVIDIA?Vibhu [00:02:14]: The GPU rich came themselves.Vibhu [00:02:19]: And timeline-wise, when was Kosmo? It was pretty early, right? It was open world model, open paper, everything.Ethan [00:02:25]: It was end of twenty-four.Vibhu [00:02:28]: End of twenty-four.Ethan [00:02:30]: Then at mid twenty-five, I moved to xAI. At that time-- I joined about the time when xAI was about to build video models and in multi-model models. There were no infra, no data, and no model, and it just-- as a few engineers, we built it in three months and released the first model, Grok Imagine zero point nine.Ethan [00:02:55]: And since then, I keep working on video models and move more from training and to post-training of the video models. For example, like a reference to videos, kind of like the cameo feature and, video extensions. And, before I left, I worked on a world model, leading a small team to focus on the real-time long horizon video generation.Building Grok Imagine From Scratch in Three MonthsSwyx [00:03:24]: Can you give like a rough roadmap of okay, you're on a brand-new team. Grok previously was only text, or they partnered with BFL for their image gen stuff. What do you-- what are the building blocks, right? You have compute, data you can procure somewhere. Like just what are like the sequence of things that people should think about when you're setting up a new team?Vibhu [00:03:43]: actually even deeper, not just data you can procure. You guys had to go through getting the data too, right? So you shipped it pretty fast, but yeahSwyx [00:03:51]: three months is likeVibhu [00:03:52]: From everythingSwyx [00:03:52]: actually like very surprisingly fast.Ethan [00:03:55]: One thing I say like thanks to my experience at NVIDIA, ‘cause first time when we were building Kosmos together, we built it, for about a year. So this is like the second time I do it. Roughly have an idea, what to do. I say the most important thing is the talent. Everyone were very strong and clever, very close with each other towards a common goal. So that speed up things a lot. So you reduce the communication bandwidth among people, and everyone can work towards the same goal. It's, it's like every day there's not that much meetings on the calendar, like maybe like a, like a sync a day, and after that it's, it's just all building. It was pretty fun at that time.Ethan [00:04:47]: And another thing is that xAI has very strong foundations of like data inference, model inference, and the supporting there can help the model develop a lot. When I look at, training models, I don't so actually the top important thing is like how many, how many iterations can you do, per day? and the more iteration can you do, you can, you can train the model much faster. So if you have very strong infra and you have a lot of compute, you can, you can train these models in very short period of time. That can give you a much larger buffer to, for errors, and it also gives you the opportunity to spot more bugs.Iteration Speed, Compute, and Debugging Model PipelinesSwyx [00:05:46]: What is an iteration? Is it like a few hundred steps or what are youEthan [00:05:50]: Let's say just the train-training the model, like from acquire new data and maybe design new algorithms and train a new model, maybe at smaller scale orSwyx [00:06:01]: So cycle time for like any hyperparam that you're searching.Ethan [00:06:04]: Cycle time and tune to like eval this model. Is this model better than my previous iteration?Ethan [00:06:11]: SoSwyx [00:06:11]: So it's like before you, someone had already set this up that you can iterate very quickly.Ethan [00:06:15]: I think the foundation there is extremely good forDeveloping and research models.Ethan [00:06:23]: And often I find is it-- this is kind of boring, but like a lot of the improvements does not come from new algorithms. It comes from finding small bugs here and there in the data pipeline, in the, in the model training pipeline. Those give, those give the biggest boost to the model quality.Vibhu [00:06:46]: It's interesting, right? So you say it's like small team, less communication bandwidth, but also a lot of quality is like find little bugs. It seems counterintuitive, right? You have a lot of people, you can iron out more of those, but it's interesting to see the other side, right?Swyx [00:07:00]: I also wonder, have you-- do you try using LLMs to look for bugs? I don't know.Ethan [00:07:05]: I remember at that time it was mid two thousand and twenty-five, so it's the coding model wasn't quite there yet. I remem- I remember like December two thousand and twenty-five, it was extremely good. Yeah, I've been, I've been using it at that time. It's, it's helpful. sometimes it produce codes that are kind of difficult to maintain, even though like the first time it built something extremely fast. But it gave the, like a spaghetti code, thousands of lines that I couldn't maintain, and the LLM itself couldn't figure out what's, what's wrong and how to improve on top of it. But now I find it much better. Yeah, I want to bring up another point here is now coding models are much more efficient and can help us implement stuff much faster. Compute might become a bottleneck again because previously, like if you want to train a new model, say you want to generate new synthetic data and then or write a new algorithm, it might take a few weeks. And during that period of time, you don't-- you might not have experiments to run. But now you can build that thing within a few hours, then you can immediately train a model.Ethan [00:08:24]: Now you have to have enough compute to try all of the ideas. So compute might be the bottleneck of iterating speed again.Swyx [00:08:36]: yeah, I actually, honestly, I think it's like kind of a stressful job because you're “Well, I should be trying everything, and if I'm not, then I'm not doing my job well.”Vibhu [00:08:48]: there's also the stress of you're eating thousands of GPUs per hour, which is very expensive and, compute can go to other researchers.Swyx [00:08:56]: You got the daddy Elon toVibhu [00:08:57]: You got daddy Elon.Ethan [00:08:59]: It wasVibhu [00:09:00]: But there's still finite amount of compute, like you want to use it, you want to use it well, you want more of it.Ethan [00:09:06]: That was quite stressful indeed. Yeah, I think one thing is the-- with coding models now, like a lot of these jobs can be automated, which is much better. A second, it's a, it's a marathon, so you got to maintain good health and, a regular schedule.Vibhu [00:09:28]: It's, it's hard to hear that when you shift from zero to nothing in two months.Swyx [00:09:32]: and, I think obviously the culture at xAI is very famously, people work very hard. one thing I did want to dive into, in our-- in the notes that you, that you sent ahead of time, you had specific comments about the cost of Video Gen training. presumably this is on the Colossus-1, right? the two hundred megawatt cluster. Any whatever you want to just share on that.Vibhu [00:09:54]: I think there's, there's three things we're talking about, right? So there's Video Gen, there's also the Image Gen model that you put out. Do you want to like complete the, okay, so zero to one, you have a few months. Just what are the stages of create Image Gen model?Swyx [00:10:06]: Oh, yeah, maybe I got distracted.How Image and Video Models Are Trained: Synthetic Captions, Tokenizers, and VAEsVibhu [00:10:07]: Sorry. and then, from there's Video Gen, there's Audio Gen. Would love to get into those next. But what is that first few months like? So small team, a lot of bugs, iterations, but what does it look like? Do we take something off the shelf? Do we just get data compute? What's, what's the few months like? How do you go to state-art Image Gen model? How do you just start?Ethan [00:10:28]: I cannot comment specifically how xAI did, but it's, it's a quite standard process. I can draw some, examples from Cosmos. So mainly it's building a video model, you actually need to build a image model first. And building these two models, the data you need is a hundred percent synthetic pair of language and image or language to video. Because on the, on the internet, actually, the videos don't naturally associate with text. So you can say, oh, like on YouTube, you have the title and you have the description and the commentsSwyx [00:11:11]: TitleEthan [00:11:11]: of a video, but usually they're not relevant to the video itself. And say maybe like the video is a natural scene of mountains or something, and the title is, I'm so happy today.Ethan [00:11:26]: So they have they have no correlation at all. So the first step is to, you have to generate synthetic pair of language with the videos. So you gather videos from the internet, and you use a VLM to caption the videos. So that part, here's a question, like how do you, how do you gather VLM to begin with? So if there's noSwyx [00:11:55]: You, so you fuse the model, right? LikeEthan [00:11:57]: Say if there's no like VLM exists, like how do you generate the text to the beginning, right? It's, it's impossible.Swyx [00:12:04]: I see.Ethan [00:12:05]: In the beginning, it's like you ask human to describe the video as detailed as possible.For example, you ask them to describe everything, like all objects, all characters, and all interaction and dialogues in the, in the videos. So that's in the protocol of Cosmos labeling. We require the objective we give to the labelers was that you have to describe the video as detailed as possible, such that a blind person hears a blob of text can reconstruct what the video is like from their head.Swyx [00:12:43]: Video or image? You're talking about images.Ethan [00:12:44]: Video or image, either one of them.Vibhu [00:12:47]: This was pretty common when we went from clip and DALL-E, right?Vibhu [00:12:51]: It's all training on really detailed captioning of images. So same is applied to video, but insteadEthan [00:12:57]: same appliedVibhu [00:12:57]: of using multimodal model to pass in video images and write rich descriptions, you can alsoSwyx [00:13:04]: I think there's this traditional perspective of supervised, or, very highly human curated thing. I feel like there's a unlock with unsupervised, right? Where like you have enough to bootstrap that you can just throw common corpus on it or, whatever. like unsupervised vision and language pairing, right? Like where you just have, interspersed image and text and it just learns. To me, that is the VLM breakthrough that is different from the clip, different from the LM era.Ethan [00:13:36]: It's interesting to see that you kind of need both data.Ethan [00:13:41]: For example, for theSwyx [00:13:41]: You need it to bootstrap it up. YeahEthan [00:13:43]: for the generative model training, there's also usually like a small percentage of unlabeled data. So the model is instructed to generate a video without any text instruction. That can also help the model generalize. So after this stage of generative synthetic pair, so, one important common step is to train a compressor or a tokenizer of the image or videos. So because, if you train-- If you can technically, theoretically train image or video models on pure pixels, but the problem is that the, it's, it's a lot of tokens. So like one image, it's, a thousand by a thousand, it's like one million tokens, one million pixels. It's impossible to train transformer on that. So it's, you need to train a tokenizer, which can go from image to latent space and latent space back to image.Swyx [00:14:45]: That's why we named the podcast.Swyx [00:14:48]: But, basically, you're talking about vocabulary science.Ethan [00:14:50]: so vocab.Swyx [00:14:51]: And so, what is, what is imp-- like a million is impossible?Ethan [00:14:54]: In generative models, the vocab is continuous. It's a continuous space. We can think about like you map an image to a vector. It's a, it's a fixed length vector. It's sixteen or forty-eight, something like that. And then you map that vector back to the image space. And the mapping is, has-- The mapping is patch-based. So you say you haveEthan [00:15:22]: a sixteen by sixteen patch and you match, you map that patch of pixels into this latent space.Swyx [00:15:29]: We've covered thisVibhu [00:15:30]: This is like the vision transformersSwyx [00:15:32]: VAEs,Ethan [00:15:33]: VAEs.Vibhu [00:15:34]: You basically compress your input, you do your generation, you're reasoning all that generation in smaller dimension, and then you project back out.Swyx [00:15:43]: VAE is a form compression, but I think the for me, the patching thing is from VIT, right?Ethan [00:15:48]: You can make those.Swyx [00:15:49]: Literally the, yeah, the paper is titled like sixteen by sixteen is all you need. something like that. and then I think also, people make a lot of comparisons with this kind of patching with convolutions.Swyx [00:16:02]: Which is you're, you're kind of re- reconstructing the old paradigm with the new.Ethan [00:16:05]: Actually, in VAEs, there are, there are both convolution networks and transformers. You can actually do both.Ethan [00:16:14]: After this VAE, so what you've got is you've got latent space tokens and you've got the language tokens. So now the training of the diffusion transformer, usually generative models use diffusion transformers. It is actually quite standard. It's, it's very similar to how you train a language transformer models. It's not that much difference. It's just the tokens, the visual tokens in, visual tokens out. The only difference is there's a denoising process. So you train the model to unmask some of the noise. So you add, you add random noise to the visual tokens, and then you train the model to remove those noise to generate the clean tokens. Any inference, the model can iteratively remove noise from a hundred percent noise.Swyx [00:17:12]: And then there's also, to speed things along on the tech tree of diffusion, there's CFG, and then there's, there's also, latent diffusion that, there's, there's someone in there. I think, somewhere along the line, obviously, like stability and all these other guys, pioneered a lot of this, architecture. I don't know if you want to get into that or just, or do the video side up to you.Bootstrapping Video from Image Models and Temporal CompressionEthan [00:17:37]: After you train such model, such image model, the reason it's a, it's a foundation for video models is that image models are cheaper to train, and they have much denser connection between language and text. So, sorry, language and images. For example, you train a billion, you train on a billion images, and there's a mapping from the text to the image. And the cost to train the same, like the, a billion, a billion text to a billion videos, that's much more expensive because videosNaturally have more tokens than images. Because the diffusion models, their understanding of, language purely come from this mapping. So if you don't have enough mapping, so if you only train on like a ten million videos or something, there-- you might not see enough language tokens in your training, so your model does not understand human intention enough. So that's why you really-- you train-- you first train this image diffusion models, and then you bootstrap the video model from there.Swyx [00:18:53]: One thing I did want to ask, because I-- actually, I think you're, you're the first per-- video model person I've ever talked to, I think. we've, we've like talked to Luma and all those folks. There's all these tricks in video compression where basically frame by frame there's not that much difference, so actually you don't have to regenerate or save the whole frame, right? but I think MP4 compression or something else like that.Swyx [00:19:16]: is it tempting to use that? Or as far as I can tell, everyone just treats it as, “No, we would just generate every frame.” Is that roughly the state-art?Ethan [00:19:27]: There are a few different approaches. Let's say first, like you want to just directly use MP4 compression and use that as the tokens for the transformers to train, right? So people actually have tried that, but the main challenge is the latent space for the MP4 tokens were not, were not very comprehensible for the models. It's, it's extremely hard to train on that. And there's aEthan [00:20:01]: So that's why they created VAEs, which creates more continuous, latent space, so the models can understand that latent space and learn from it much easier. Even within the VAEs, there are different difficulties of the latent space. So you can imagine something the simplest, the most naive VAE is like you have an image, and you just shuffle all of the images into a, into a vector. So you don't need to train any VAEs, right? But that latent space is extremely hard for models to train on top of. That's why there are some debate on like how do you compress the tokens. So you mentioned like you can compress frame by frame. Also, you can compress, the temporal dimension.Ethan [00:20:52]: The difference is if you compress the temporal dimension, you get a much higher compression rate. Because there's temporal redundancy between frames, because, this frame and the last frame, likely they are mostly similar, so there's only some small difference. for example, I think in 12.1 VAE, they have like a eight by eight by four compression rate. So the four temporal tokens are compressed into one tokens. That can save a lot of, save a lot of the context length. If you do it frame by frame, you have to do maybe like eight by eight by one. Your context length will be four times larger. That being said, the benefit of the frame-- per frame compression, we might come back to this later, is, real-timeness and interactivity. ‘Cause if you, if you strain the output of the model, frame by frame, you can-- the model can respond to any user request immediately. So if you have like a temporal four compression, four times compression, thenSwyx [00:22:06]: It might be laggyEthan [00:22:07]: there's a lag there in nature.Swyx [00:22:10]: So you're very pilled on this. let's just go ahead and bring it up ‘cause we have the visual prepared anyway. There's some frontier applications of real-time video gen. So Flipbook is one of the examples that went viral recently, right? What is Flipbook?Real-Time Generative UI: Flipbook, Neural OS, and Diffusion Front EndsEthan [00:22:23]: Flipbook is kind of like a web brow- web browser. You can see like it has the web bro- browser UI on top. The difference is all of the UIs are generated by generative image model in real time, and anything here are fake. But you can, you can explore inside this wor- this imaginary world. Say like we-- here we have engineering the Great Pyramid. Like the model generates this for us to understand how it works, and if we want to navigate around and understand further, we can click on some of the, some of the description here, and the model will generate a new page, new subpage describing the details we want to know about.Swyx [00:23:14]: So it's basically kind of we're playing a video, but it's pausing for our next interaction, and then it just plays the next thing based on our interaction.Swyx [00:23:23]: Which is kind of cool.Vibhu [00:23:25]: and you kind of decide your story. So this was, how do you make a pyramid? levering technique seemed interesting, right? It shows how do you take Okay, I want to know what is thisSwyx [00:23:35]: The demo, the demo tweet had more animation between frames.Vibhu [00:23:38]: I think it's just skipping,Swyx [00:23:39]: Oh, it's just skipping a lot of frames.Ethan [00:23:40]: they also have a video modeVibhu [00:23:42]: It takes a lot. There's a lot of peopleEthan [00:23:42]: but, a lot of people are using it.Ethan [00:23:45]: So it's not available.Vibhu [00:23:46]: There's a live video stream. We can try,Swyx [00:23:50]: So this is an example of the kind of future that you see at the extreme. We don't-- we're obviously not in it today.Swyx [00:23:56]: But in a world where inference is completely free this is better than generating code and text?Ethan [00:24:02]: So this is, this is a final state of where Viva will be at for word model, I think. Imagine internet doesn't exist, and then you type in google.com. Like what should, what should, what should a model show you?the model can imagine something, and this is what the model imagine. And these web pages, they completely do not exist. So I think as the inference costs come down, we are going to have generative UI for everything. If you think about how the coding model works, so they write code for a web page, and they render the code might be con- converted into binary, and the binary render the pixels on the screen. So we in machine learning, every time we have some breakthrough, obviously it's, it's more intuit. So why don't we have like user instruction to the pixel directly? So the generative UI will be user intention to the pixels directly. And say like even if I want email, let's say everyone have the same interface, but I want, I want it slightly different. I want the email to show to me like a TikTok, so I can swipe left and right for the emails. And or maybe you want something else. We can have completely different things. Or like I have I'm looking at, Instagram stories, and I don't like the Like button. I always may click it. And, generative UI resolved it. So it's going to be a revolutionary replacement of the interface. So in the future, we might have much more powerfulEthan [00:25:50]: LLMs and coding models running behind the scene. And in the, in the front-end, the diffusion model will actually be the front-end to show stuff to you. That's how I imagine it.Swyx [00:26:02]: Diffusion front-end, deterministic back-end.Swyx [00:26:04]: Something like that. I find that very expensive, but,Vibhu [00:26:08]: I find it interesting you called LLMs writing code on the back end deterministic, but okay.Swyx [00:26:14]: you write it onceVibhu [00:26:15]: Compare it toSwyx [00:26:16]: And then you execute.Ethan [00:26:17]: If you think about the cost, say, let's say H100 costs $1 per hour, and if you use this eight hours a day and thirty days, so, every month you're paying this two forty, you'll actually not wanna pay for that. That's even more expensive than Cloud Code Max. But if you think about the compute costs come down like two times every year, and I think the future will likely arrive like within few years.Vibhu [00:26:49]: It's everything, right? compute cost comes down, compute gets faster, model gets smarterEthan [00:26:54]: More efficientVibhu [00:26:54]: model gets smaller.Swyx [00:26:55]: I don't know why you say two times, ‘cause I think it's like 100 times. In language models, it is roughly one hundred to a thousand times every twelve to eighteen months, for the same given level of LMSys, ELO.Vibhu [00:27:08]: That's a net of everything, right? That's model performance alongside compute. So different than just compute costs come down. But, a very interesting future.Swyx [00:27:19]: So the web designers will have to shout out that accessibility is an issue, right? how do you deal with screen readers or whatever. But yes, this is higher bandwidth storytelling than anything you can possibly generate with code, right? So I think that's the rough idea.Ethan [00:27:34]: And I'd like to add a little bit that so human naturally have the maximum bandwidth when we are looking at things, look at videos, and we also have maximum output bandwidth when we are talking. So in the future, it might be something like we talk to AI models, and the AI model responds back with a generative UI. So that would be the maximum input and output bandwidth to interact with AI models before neural link happens.Vibhu [00:28:06]: And it's also very custom, right? Some people are very visual, some people are not as visual, right? They prefer the text. But the best thing about generative UI, right, it can also be text.Swyx [00:28:17]: There's another project that we wanted to highlight, which is the Neural OS. Kinda similar idea, but here you're literally operating, simulating an operating system with a video model.Swyx [00:28:27]: and you can play Doom, you can do Firefox. I find this like mildly less impressive, obviously, because it's an OS that I can run.Swyx [00:28:37]: But here everything is imagined.Vibhu [00:28:40]: I was, used to the Command+W to close the Firefox tab. It didn't crash. That's why I saidSwyx [00:28:45]: It's too immersive.Vibhu [00:28:46]: It's, it's too immersive for me.Swyx [00:28:47]: Too immersive.Vibhu [00:28:48]: I wanted to close the tab.Vibhu [00:28:49]: But yes, I can play generated diffusion.Swyx [00:28:51]: this is shockingly fast.Swyx [00:28:54]: Because I remember there was a demo about like maybe one to two years ago. Someone tried to do the first-person shooter with a image model. There was no consistency. It was very slow. But here it looks like realistically it's-- this is Doom.Vibhu [00:29:07]: I think there's two sides to that, right? There's okay, what is running a game? The heavy part of it is actually the game engine, all the lighting, all that stuff, the graphics. This is just kind of video, right? Like we've solved consistency. This is still, it looks like a few years old image generation. There's some temporal consistency, but it's, it's kind of just images stitched together as frame video. But it's a good visual representation to pi- to picture the future you wanna see, right? that's, that's what I see in these more so.Ethan [00:29:38]: This reminds me of how the video models gets better and better. So Neural OS is kinda if you just look at it feels like it's just a crappy version of the, like the Windows we could have, right? And, but the difference is, so the model, this model is overfitted on the existing operating systems. It can generate nothing different than that. But it's actually also similar to video models. So when we are training these video model, image model, we train them on internet. There's no imaginary supernatural stuff on the internet. But once we train this model, you can prompt the model to generate something supernatural that have never existed in the data set. So if you train your Neural OS or neural computer on the standard screen recordings on the entire internet. The model can imagine completely new interface to interact with the computer.Swyx [00:30:43]: This is one of those things that is magical to me. usually generalizing out of distribution is bad, but somehow we have learned some kind of internal world model that you say, this plus, but it looks like rainbows and butterflies, it'll do it and it will kind of make sense.Swyx [00:31:03]: So yeah, that's kind of cool. Yeah, I don't know if there's any comment more on there. I do, I do wanted to, I did wanted to touch a little bit more on the model architecture stuff, which I think you were getting. It's, really fascinating. We don't get a chance to talk about this enough. So one of the papers that we covered, we've covered every annual, segment anything release. and I don't know if you follow-- you're a computer vision guy, so youEthan [00:31:26]: I knowSwyx [00:31:27]: . So they did memory attention, which is kind of interesting. And I always think, anything where you can, across the temporal dimension, keep some consistency, I think it's, very fascinating, and I don't know if Basically, does that-- the CV side bleeding into video gen side, I think is underexplored, right? we talk about it for labeling, but actually you can borrow the architecture itself.Ethan [00:31:50]: There's, there's also complete different approaches, right? you brought up the term world model, so we went from video model to world model. There is diffusion, but there's also other approaches that people are doing. So maybe we get into those after as well,?Swyx [00:32:03]: He has a whole definition of world models and stuff. I feel like we threw a lot at you. Whatever you want to comment on.Why Video Models Are Expensive: Storage, I/O, and Training ScaleEthan [00:32:10]: I think one thing that we should actually comment back on is okay, so we were talking about the steps to train image gen to video model. One thing we don't see as much of is okay, you brought up the delta in training data, right? SoEthan [00:32:24]: you won't have as much a video model might not generalize, but what is the cost of training a large video model? So we know for LLMs roughly, okay, even like the poolside thing that came out today, right? It's a Gemma level model trained on roughly forty trillion tokens at this many H200s over this much time, right? You can see what is the exact cost of that. So how many GPU hours over how much H200 costs? So how do we do the back-end math of, same thing for video models, image models. How do you, how do you kind of break that down? I can share some back-envelope calculation. So surprisingly, video models is-- the cost is very-- is comparable to language models and obviously the largest scale is language model, maybe like a medium scale to language models. I said just storing the videos alone, it costs a lot. You can, you can maybe look up on AWS or something.Ethan [00:33:20]: You really, say if you have a billion videos and let's say, let's just say like each video, like five megabyte, then you need five petabyte to just store those videos. And also remember we talk about you use a VAE to compress the videos, and you also need to store, typically you need to store those continuous feature, in-- also in your storage. That's also comparable size with the videos themselves. So just storing these videos and the features is tens of petabytes alone. And,Swyx [00:33:58]: I just, I just looked up the calculation. Five petabytes on S3 Standard is one hundred K per month.Ethan [00:34:05]: AndSwyx [00:34:05]: It's comparableEthan [00:34:05]: and you needSwyx [00:34:06]: AndEthan [00:34:06]: And then like tens of petabytes, two hundred K. And even more expensive is you have the ingress and egress.Swyx [00:34:13]: Oh, yeah.Ethan [00:34:14]: Like you-- through the internet. You have to just to download those videos, I believe it's, it's more expensive on AWS than just storing those videos.Swyx [00:34:25]: Storing, yeah.Ethan [00:34:25]: And each training runs, you probably need to pull them once. If you train multiple times, it's, it's even more than that. So it's like just storing the network, those costs is just, it would be a few, a few millions per month to just storing everything, not to mention the GPU cost.Ethan [00:34:45]: AndSwyx [00:34:45]: my side tangent, the compute rental, like GPU rental is very efficient. There's one side, okay, you can be XAI and build your data center. Should we not just build our, storage compute as well? LikeEthan [00:34:57]: Of courseSwyx [00:34:57]: cloud cost compared to just,Ethan [00:34:59]: You save so muchSwyx [00:35:00]: store. Yeah, exactly.Swyx [00:35:01]: Especially with like egress and stuff. So.Ethan [00:35:04]: That's a good idea, but it also comes to-- there are some of its own challenges.Swyx [00:35:09]: Of course, of course.Ethan [00:35:10]: like people who build the GPU data centers, they might not expect this much, storage. And yeah, people build storage, typically they just build it somewhere with just CPUs.Swyx [00:35:23]: I just looked it up. Five-- AWS only charges for egress, not ingress. Tier five for five petabytes is two hundred and thirty K.Ethan [00:35:32]: Even more expensive than the storage.Swyx [00:35:34]: But storing is per month, right? You check in, then you cannot check out. so it's so cool. It's okay. So there's that side.Ethan [00:35:41]: So the TLDR, my backhand mathSwyx [00:35:42]: Data is larger than you think. Yes.Ethan [00:35:44]: my backhand math of GPU hours times GPU cost is also very much, I'm missing some storage.Swyx [00:35:49]: You're also-- you're basically like also more IO bound than normal training.Swyx [00:35:55]: Yes. ‘Cause like data loading, so caching everything, it becomes super important.Ethan [00:36:00]: So in Cosmos, we did a lot of optimizations to make it not IO bound. So, speaking of the training, actually training the model, the GPU cost, if you look up like the open source model, how big these video models are, I think like LTX has nineteen B parameters. That's a dense model. And people are also exploring, MoEs, so it might be twenty B active and, like a hun- hundreds B, total. So that's, that's even-- that's similar size as medium-sized LLM models. And if you, if you look at number of tokens-Uh, we disclose that in Cosmos. It's also like tens of trillions of tokens on the visual tokens. So putting this together, the cost of, training these video models, it's actually comparable with LLMs. Not to mention, the infra is slightly different from LLM, so it might be less efficient to train these models.Inference Speedups: Step Distillation, Consistency Models, and GANsSwyx [00:37:04]: Do you get the benefits of traditional diffusion speed-up? So for, images, there's LCM, LoRAs for, fine-tuning. There's, there's a lot of stuff that's beenEthan [00:37:15]: Flow matching.Swyx [00:37:16]: there's flow matching. There's a lot of stuff that's been done. there's some overlap that applies to diffusion on the inference side and stuff or?Ethan [00:37:23]: so the difference-- the inference side is a completely different story.Ethan [00:37:28]: I think for the training side, it might be a little bit hard to reduce that cost. And for the inference side, the biggest gain is from the distillation of these models. You can-- It's called step distillation, slightly different from knowledge distillation in LLMs. So you-- Typically, for flow matching models, you need like 100 steps or something. Like a distortion model even need even more, like 1,000 steps to generate a good image or video. A step distillation is try to learn to generate fewer step from the model itself. It's kind of like now we-- you use the full model to generate in 100 steps, and then you take a model that only generate 10 steps and let that model to learn from the perfect one.Ethan [00:38:25]: why this workSwyx [00:38:27]: Strong to weak seemingly.Ethan [00:38:28]: It is. It's kind ofSwyx [00:38:29]: DistillationEthan [00:38:29]: kind of like strong to weak. the-- from the modeling perspective, the strong model, the teacher model is trying to model the image and videos of inter-internet, and that distribution is extremely complex. But the step distilled model is just trying to learn from the teacher. The teacher is a model, and the size is fixed, as the distribution is much simpler than the whole internet. That's the intuition I have why step distillation can work. So usually these models serve in productions, they only run in a few steps. In Cosmos, I believe we have, we have like four step and eight steps. If you do some simpler task, image-image translation, it can even run in fewer step, like one step in Cosmos Transfer.Swyx [00:39:22]: I think this is the same intuition that guides a lot of the consistency model work. I sent you a link for, SCM. I don't know if you covered that. To me, that was actually one of, the most impressive papers I've ever seen from OpenAI.Swyx [00:39:34]: That this is the unifying grand concept of consistency models. I don't know if you have any comments on this.Ethan [00:39:41]: So there are, there are a few different approaches,Swyx [00:39:46]: Oh, yeah. Here it is.Swyx [00:39:47]: Two steps versus twenty or 100 steps, whatever. It's already done.Ethan [00:39:52]: So there are, there are a few different approaches, for example, consistency model, and there are also Actually, we shouldn't forget GAN. So GAN, actually, that was, that was the OG ofSwyx [00:40:05]: OGEthan [00:40:05]: step distillation ‘cause it trained just one step to begin with. So actually, a lot of, uh-- For example, there's a distribution matching distillation which use, which uses GAN, as one of the laws for distillation. It-- GAN just tells you, “Hey, generate an image,” and thenEthan [00:40:31]: it has a discriminator to tell, is this image real or not? So the model, the model just need to learn one of the distribution, not the full distribution. Because in training, the model is asked to reconstruct the ground truth image from the internet, which is extremely hard. And in-- When you're training GAN, it's a step process. It's just a, “Hey, you generate image. Does this image look as real as the image from the internet?” Which is a much simpler task. And, yeah, combining a lot of these approaches together, people typically do that, like consistency model and distribution matching and GAN, and we can get these few step models.Audio-Video Generation and Time AlignmentSwyx [00:41:21]: Then there's one step I wanted to add, which is audio and video.Ethan [00:41:26]: So, Grok Imagine zero point nine, I believe it's, it's a first audio video transmodel deployed at a large scale. SoSwyx [00:41:39]: And that was your first model?Ethan [00:41:40]: that was, Grok Imagine's first model. It's, it's audio video, joint generation. I think the hard part is, the modality alignment, ‘cause before this transmodel, we have, we have text to video alignment. We have this, correspondence between text and video. Typically, most of the VLMs, they understand images and videos. Video's very rare, and they don't understand audio mostly. And if you look at the audio generation on the LLM side, you can talk to them perfectly fine, but if you ask them to sing a song or something, it typically is not very good. Also, they don't have, they don't have music either. The hard part is thatUh, actually audio has two component. It has like a discrete component, a continuous component. The discrete component is like the language.Ethan [00:42:44]: So when we speak, it's just, someSwyx [00:42:47]: It's an ASR issue, yeah.Ethan [00:42:49]: It's, it's text token with some characteristics, I would say.Ethan [00:42:54]: But musicSwyx [00:42:56]: I think the speech guys would disagree with this.Swyx [00:42:57]: Like disfluencies and then,Vibhu [00:43:00]: There's tones you can get angry.Ethan [00:43:01]: Well, I say largely.Ethan [00:43:03]: the mu- but the music is completely different. It's, it's very continuous, and you cannot model them like discrete tokens in language models. this is like the hard part for models is, not to mention we have to align text, video, and audio together.Ethan [00:43:26]: SoVibhu [00:43:26]: How?Ethan [00:43:28]: So significant-- some significant challenges are like-- So first, like we talk about as the VLMs, they cannot understand most of them cannot understand audio.Ethan [00:43:39]: So you have to have some way to do the synthetic data generation for audio. You have to caption the model, and that involve, that involve synthetic data and human data effort a lot. And not just surprisingly, most of the LLMs are very bad at recognizing, like the beat, tone, and the details of the of music. They can, they can give some general prediction of which song is this, but it's very hard to describe the details of the music. like we mentioned in image generation, like you have to describe image as detailed as possible so that someone blind can reconstruct that. So here is like someoneVibhu [00:44:32]: DeafEthan [00:44:32]: someone deaf can reconstruct how the music sounds like without actually listening to it. Maybe you can think of it need to have the-- or they call the script.Vibhu [00:44:49]: Subtitles, yeah.Ethan [00:44:49]: You gotta have all the details of the music, and the dialogue.Vibhu [00:44:55]: So is the challenge there typically stuff like music and audio, or is it just Like is there a baseline? Okay, there's enough data where we can understand, narration, conversation, but there's nuances in audio that's where you hit all the data issues or is it just from stage zero, you just do it all right?Ethan [00:45:15]: So one important thing is like the alignment. So the model, the model has to know like the video and audio, the, uh-- it has to have a time-based alignment, like at which time step the video and the audio token correspond to each other. But we actually don't have this kind of alignment for most of the other modalities. If you think about like text and image, text and video, they are loosely aligned. So you can, you can have a description of what's going on in the video, but you don't have to exactly, You typically don't have exact description, oh, at, time step one second like what happened?Vibhu [00:46:02]: It's veryEthan [00:46:03]: At time step two second what happenedVibhu [00:46:03]: coarse. Yeah.Swyx [00:46:05]: So what was the ideal time step? You have to oblate it, and then it's like four seconds or something.Ethan [00:46:09]: So that comes down to how you design the model to, for the model to be aware of as a time, as a time modality. So the model is like a time aware. And that's something pretty unique if you think about LLMs. So if you ask LLM to complete a task, say they, uh-- you ask them and they will say, “Oh, this task will probably take twelve hours to complete,” and they come back in one hour. Say “I've already spent two days on this and I've exhausted everything.”Ethan [00:46:47]: So the LLMs them-themselves, they don't have a sense of time there.Vibhu [00:46:53]: I actually don't think that's just them not having a sense of time. I think it's somewhat based, right?Vibhu [00:46:58]: Like you tell someone, “Okay, go work on this feature. Go implement this,” there's a general understanding you would have of how long that would take without LLMs working at LLM speed, right? So you think back like two years ago, if I tell you to like build me like a new front end for latent space, have a search bar, have all this, you'll estimate that it'll take a few days, right?Vibhu [00:47:19]: So you tell an LLM, “Go build this.” It'll take me a few days. But I think it's somewhat grounded as opposed to them not having the best-- Not saying that they have a great understanding, but I think that example is like you can see where it comes from, right? You're trained on all over the text.Swyx [00:47:35]: They're, they're trying to estimate what a human would say.Vibhu [00:47:37]: because that's what the, that's what the data kind of represents. It's not themEthan [00:47:41]: It came from the corpus on the internet. People have a estimate of how much time.Vibhu [00:47:45]: And not even just in direct like training samples, right? Just your world understanding of tokens of how long stuff takes, right? Go read a book. It'll take you a while, right?Vibhu [00:47:56]: Even if you do nothing but read a book, it takes a few days. So yeah, LLM, I read it took me a few hours.Vibhu [00:48:01]: It'll take me a few hours to go through this research. But this is a tangent.Swyx [00:48:05]: Somewhat, yeah.Swyx [00:48:06]: This is a train of thought I haven't really expressed until now is, which is basically like a full world model must also be recursive, meaning that the participant in the world model must also be aware that they have a world model. which is like this whole recursive thing down the, down the line. but yes, and that the world model can be wrong and that they need to update it and blah. Yeah. We've, argued this on the, newsletter as well, that there needs to be sort of recursive or adversarial world models.World Models: Real-Time, Long-Horizon, Interactive VideoVibhu [00:48:34]: just, to ask, how do you define world model?Swyx [00:48:38]: Oh, yeah, let's go there.Ethan [00:48:40]: SoVibhu [00:48:40]: So just for context, we talked about, video generation, and then there's a-- if you say there's a distinction between world models, what's your, what's your definition? How do you see the two?Ethan [00:48:53]: So disclaimer, I'm not going to debate, what is world model. Yeah. there are many definitions, so I'll just talk about my definition. Since I came from the multi-model, multi-model domain, so mainly talking from video. So world model is like real-time interactive long horizon videos. So there are three parts. so we-- let's talk about them one by one. So the so interaction, so we just, we just look at Facebook and neural computer. So the interaction part of it, so you, world model can allow you to interact with them through keyboard, mouse, and maybe also voice. So these all is-- all is a modality. You can, you can interact with the model, and the model should respond reasonably. Second part is real time. So once you, once, say, you move your mouse, if, say, the world model generate a game, how fast can the game respond? So if you're like professional CS: GO players- -my say, oh, you have to respond- He's beginner within sub ten milliseconds or- Yeah even less. So that's not most of the- No, sixty FPS. Let's go. Oh, three hundred FPS. Oh, five hundred FPS. Wait. okay, yeah. I didn't do the math, but yeah, okay. Uh- Yeah, three hundred FPS, that's a three millisecond. So you have to respond- Oh, s**t. Okay. YeahEthan [00:50:29]: within a millisecond. Most of the video models cannot do that. Yeah. And, but if you, say, if you have a video model that is, say, like a digital human, the response time might be more generous. Maybe typically, for real-time voice interaction, it's like two hundred millisecond. So that's, that's much more generous. But even two hundred millisecond is pretty, it is pretty tricky, ‘cause remember we mentionedEthan [00:51:01]: you have this, temporal compression coming from the VAE. So if you, if you don't compress the temporal dimension, your sequence length is going to explode. So if you want to have this real-time, real-timeness in your model, you have to do is one context problem. And the third part is long horizon, ‘cause we-- if you're not going to just play with, video games just, a few seconds, most video models only a few seconds. We're going to play with minutes, hours. The model have to be able to generate long-form content.Ethan [00:51:42]: So putting these three together, it's, real-time, long horizon interactive videos. I think the final state will be, for example, like a video, a video version of Playbook, where you can, you can interact with, a neural computer. You move your mouse, and you click on the generative interface, and it will reply to you through pixels- generating in real time. But getting there, it's, it's a very long way to get there. So one of the first step, at Grok Imagine, where I led a small world model team there, was to build video extension. So, video extension- it's the first step of interactivity. Yeah. It's, it's the first step. Yeah. So it's the first step- You have it here, video editing, yeah. Yeah. Yeah. So the first step is because, this unlocks long horizon videos. Typically, for most of the video generation models, you give it a prompt or an image as an initial frame. You generate video, that's it. That's just, one time, done. And some creators would try to, use the last frame as a first frame for the second video. It can-- sometimes it works, but if you do it a few times, it says the quality would decrease. And- It doesn't have that context- Yeah over the full video, so the temporal- Yeah, exactly. Yeah, ‘cause you only gave it the last frame, of course, right? Yeah. Exactly. And- it's actually a pretty fun hack. if you've seen like- Oh, no, he's saying something better. Yeah. And for example, like Vue, I remember Vue 3 has like a second context of the last video. It is slightly better than using the last frame, but it has the same problem-- similar problem that it, the quality would decrease. if you extend a few times to, one minute, the video quality would look much worse than the first video. Second, another problem is that the model doesn't have long-range knowledge of, what's happening before. Say, if they generate some dialogue, some, two people speaking, and their voice might change, over some time, especially if the second conditioning, it does not cover the previous context. So these are the core challenges. So the Grok Imagine video extension, it has historical context of all of the previous generated videos. It can, It has, it has the context of, who is speaking and what objects have appeared and everything, having that to generate the next video. So if we naively do this, you can imagine, just, put all of the previous history video tokens into the context. The context lens will easily explode. Especially for video models, that can be like a few, a few million context, I would imagine- context lens. Yes.Yeah.Swyx [00:54:58]: Let's run with that.Ethan [00:54:59]: for example, like in Cosmos, I think just five seconds of video is like a fifty K or sixty K number of tokens. So like if you do, if you do fifty second, that's a five hundred K tokens. If you do longer than that, easily explode. This long horizon, problem was the first step we're trying to solve world model. It turns out people, yeah, people love video extension. Like a lot, a lot of the creators love using video extension to create longer form videos. This is the part I liked that you have a, you have an intermediate step toward the final goal instead of just a straight shot to the final version very much.Swyx [00:55:48]: But I can see you have a strong vision of where we want to end up.Long Context, Redundancy, and Efficient Interactive VideoVibhu [00:55:51]: Does it seem like it's an efficiency issue? okay, we're at a few million tokens context,. If you draw the parallel to language models, we had very short context, two thousand, eight thousand, then, you scale it up one million, ten million. sure, there's effective context, but at the end of the day, it's just what's it worth? sure, there's a whole training data side. In video, it might be slightly easier ‘cause we have a hundred million token video, right? Just take a movie with the full context there. Like is this efficiency from an inference standpoint that like it's expensive, but we know how to solve it? Or like why is this not the approach? So like my broader point was on your second point of world models, you say it needs to be interactive and live, right? You should be able to play a game and see the interaction live. So one thing I see with research is a lot of what you actually serve is different than what you build, right? So we talked about distillation. You train big model, you distill it, you do quantization, speculative decoding. We do all this stuff to serve it efficiently. Should we not just have a solution, like a world model that can interact well, do inference optimization, serve it, distill it secondary, so make it real time after you solve it? So like a-- another parallel is say, continual learning, right? What we need is someone to solve it and show it works inefficiently. Give it a few years, people will make it efficient. Same thing with regular attention, right? It worked. Over a few years, people have different forms of attention, and we've scaled it to be efficient at log context,? So kind of two things there, right? One is it seems like it works. You've scaled it. Can we not just scale it a lot more efficiently over time? Do we need a separate approach if this works? And same thing with interaction, right? if we can get it done, like if we can solve some way that it works, we can solve making it more efficient from an inference standpoint later.Ethan [00:57:53]: that's actually a very good point. So in videos, there's actually a lot of redundancies. So we solve a lot of the pixel redundancy from VE, but there's more redundancy in long range and long horizon videos. Say, if a character appear in the first clip and then it disappeared, it only reappear at the end of the video, you probably don't need the-- the context, like in the middle of the generation. So you only need that character, where you need. So that's why, I helped build another feature. It's a reference video.Vibhu [00:58:36]: Is it here?Swyx [00:58:36]: is it the same model release or different one?Ethan [00:58:39]: It's a different one.Ethan [00:58:41]: You probably need to search onSwyx [00:58:43]: I'll find itEthan [00:58:43]: X reference to video.Ethan [00:58:46]: So reference video allow you to like upload up to seven images as condition and generate the video. Say, if like I want-- it can, it can be characters or objects or even scenes. Say like I want, I want condition on, Sean's selfie and holding a bladeSwyx [00:59:07]: We have a dogEthan [00:59:08]: or whatever.Swyx [00:59:08]: We put the dog in the thing.Ethan [00:59:09]: you can put them there and the video models will generate the video from and copies the context over. So that can solve a lot of the problems there, like the long context problem. It doesn't need to have a very long context, but it's-- I feel like it's an intermediate solution. The modelSwyx [00:59:29]: It's cheating.Ethan [00:59:30]: the model should be able to like selectively know, where should I draw the references. So say if I want to generate a movie, I generate it autoregressive, like a ten second at a time or something. And now this character appear, I can look back to where it first appear and, bring that back. Yeah, this one, I put the references. Yeah, that's, Optimus, Einstein myself, Annie.Vibhu [01:00:02]: Oddly enough, I used Grok Search to find it, and it pulled your LinkedIn post. But yeah we found it.Ethan [01:00:08]: Interesting.Vibhu [01:00:10]: ButxAI's Underrated Work, Culture, and WatermarkingSwyx [01:00:11]: this is a problem. This is not your fault, but like XAI doesn't communicate all this work that you do very well because they just have the model release and then that's it. But actually, these details are very good.Swyx [01:00:22]: As far as I understand, everything you just described is state-art, like no one else has done it.Vibhu [01:00:30]: A lot of-- yeah, I have a lot moreSwyx [01:00:32]: And then, and then you just put this blog post with the cookies. I'm this is not enough,?Swyx [01:00:37]: but I, obviously this is like the high level numbers that people want to know. But no, okay, soVibhu [01:00:42]: And I wonder, like part of that is also some labs don't share research into what happens. And ifSwyx [01:00:50]: No, but this is literally bragging about how good they are, right?Swyx [01:00:54]: Like, why would you not say that you are capable of extending with full context? this is not a secret sauce. This is like we did the work. yeah, I don't know.Ethan [01:01:02]: different labs have slightly different communication styles.Swyx [01:01:07]: Anyway, if anyone from XAI is listening we are always happy to help you tell your story. Yeah, okay, so you did references, and I think, I think kind of the point you're, you're making is it is sort of like a kludge, right? this is-- you can do seven, but what about 100?Swyx [01:01:23]: Right? Then you need a completely different thing.Ethan [01:01:26]: So I think it's-- this is, a mechanism to, select the context from the history, and you might not put the entire history into the context. for example, there's a paper called Frame Pack, which haveEthan [01:01:41]: a heuristic that the latest history, the last one second, I put the entire history, and the history before that, I would, compress it and makes the video smaller. So they follow this pattern, this build overall pattern that the maximum sequence length is fixed. So the further you are from the current frame, you have a smaller image. So this is just a heuristic. I think it can be more automatic. The model is aware like which history part of it can be select. So this part of the research is actually being actively, worked on by a lot of people. It's also quite interesting. I feel this is actually, this part of long context is a little bit ahead of the LLM part.Ethan [01:02:31]: So for example, like in LLMs, if you-- so contexts keep growing. Let's say if you call tool and the tool call history is extremely long, that's still in context, and keep growing, keep growing. Even if you switch the topic to something else, the whole context was there. There are some agentic harnesses that help you to, say, prune the tool results and, prune Like when you, when you query a file, only show like the top 200 lines or something. Those were very heuristic-driven.Swyx [01:03:08]: For listeners, we did a write-up on the cloud code, leak where there are eight different kinds of pruning, including like you prune the tool results and all that. So you can, you can read up on that kind of thing.Ethan [01:03:17]: I think, one breakthrough in continual learning might be like a way to automatically, manage its own context.Swyx [01:03:27]: These are all heuristics, and they will be replaced by machine learning.Ethan [01:03:30]: InterestinglyVibhu [01:03:32]: TheEthan [01:03:32]: the same thing is being researched in both LLMs and video models.Vibhu [01:03:36]: The interesting thing is also like in the paper you showed, it's actually happening at the model level, right? Compared to like language models, sure, we have base attention, but we'll do our own compression, we'll do our own pruning, which is separate from model error.Vibhu [01:03:49]: Eventually, it all just boils in, hopefully.Swyx [01:03:52]: I think this is a form of like attention, but like also know sort of reasoning attention. I feel like that's different than normal attention.Swyx [01:04:03]: Does that, does that make sense?Ethan [01:04:04]: It's, it's different in the sense that attention, not to mention, set sparse attention aside,

AI w e-Commerce to dziś temat numer jeden, ale w praktyce w firmach to tylko 20% całej układanki. Reszta - czyli 80% - to zwykła, dobrze poukładana automatyzacja procesów. Bez niej żadne AI nie zadziała.W tym odcinku moim gościem jest Mikołaj Brunka z Nocodework - były właściciel e-Commerce'u, który dziś wdraża automatyzacje i sztuczną inteligencję w wielu polskich firmach.Mikołaj opowiada o:➡️ trzech kryteriach, po których rozpoznasz proces wart automatyzacji➡️ tym, jak obsłużyć 70-90% zapytań mailowych w pełni automatycznie➡️ powodach, dla których voiceboty wciąż są złym pomysłem➡️ wyborze między n8n, Make a Cloud Code - i pułapce, która generuje rachunek na tysiąc złotych za jedną noc➡️ najczęstszym powodzie, dla którego wdrożenia automatyzacji upadają (i nie jest on techniczny)Posłuchaj, zanim wydasz pierwszą złotówkę na wdrożenie automatyzacji w swoim sklepie.

О чём говорили: - Code with Claude 2026: конференция-продукт, полупустой зал для девелоперов и бан Лёши в Azure - Copilot перешёл на оплату по реквестам вместо токенов - Антропик проиграл OpenAI — или всё-таки нет - OpenClaude: нужен ли он и почему его не настроить без сеньора - Доклад Fiona Funk: код подешевел, а bottleneck теперь — это человек - Shift to left, боль автотестов (CarPlay, Raspberry Pi, ворота) и спор про документацию - Экономика AI в Enterprise: инвестиции, хайп и кого теперь нанимают - Конец халявы: смерть «-p» и pragmatic usage с 15 июня - Голосовой режим, удалённое управление и докер-войны - Авторежим, bypass permissions и плагин SuperPower - Worktrees и боль параллельного запуска проектов - Thinking Lever: effort, thinking-токены и Opus 4.7 на Extra High - Кэширование токенов, ToolSearch и паттерн Advisor - Copilot at GitHub Scale: Harness и кэш на уровне организации - Vibe coding с Борисом Черни и Джаредом Самнером: Bun переписали с Zig на Rust - Memory and Dreaming: как агенты консолидируют память во сне Тайминги: 00:00 Интро 00:54 Code with Claude 2026 и бан в Azure 03:03 Copilot: сертификация и оплата за реквесты 06:37 Конференция-продукт и Antropic против OpenAI 12:19 OpenClaude: нужен ли он и как его готовить 17:47 Доклад: Running an AI Engineering Organization 26:05 Старые процессы, документация и стоимость AI 36:30 Кого нанимают в AI-команду 41:01 What's new in Cloud Code: смерть «-p» и pragmatic usage 50:18 Голосовой режим и удалённое управление 1:00:32 Авторежим, bypass permissions и SuperPower 1:06:55 WorkTrees и боль параллельного запуска 1:14:16 Доклад: Thinking Lever и thinking-токены 1:25:04 Доклад: кэширование, ToolSearch и паттерн Advisor 1:36:44 Доклад: Copilot at GitHub Scale 1:44:08 Vibe coding с Борисом Черни и Джаредом Самнером: Bun 1:51:31 Доклад: Memory and Dreaming for Self-Learning Agents Нас можно найти: 1. Telegram: https://t.me/proConf 2. Youtube: https://www.youtube.com/c/proconf 3. SoundCloud: https://soundcloud.com/proconf 4. Itunes: https://podcasts.apple.com/by/podcast/podcast-proconf/id1455023466 5. Spotify: https://open.spotify.com/show/77BSWwGavfnMKGIg5TDnLz Нас можно найти: 1. Telegram: https://t.me/proConf 2. Youtube: https://www.youtube.com/c/proconf 3. SoundCloud: https://soundcloud.com/proconf 4. Itunes: https://podcasts.apple.com/by/podcast/podcast-proconf/id1455023466 5. Spotify: https://open.spotify.com/show/77BSWwGavfnMKGIg5TDnLz

Nick agreed to personally set up your Orgo in a 15 min call: https://startup-ideas-pod.link/orgo_ai I sit down with Nick from Orgo to break down exactly how to run a one-person AI agent business that can realistically clear a few million dollars a year. Nick walks through the offer, the verticals worth chasing, the full software stack, and the live setup of an agent that manages other agents. We focus on tactics over theory, with specific tools, pricing, and the playbook for landing customers as a solopreneur. By the end, anyone with solid AI fluency will have a clear path from offer design to fulfillment. Timestamps 00:00 – Intro 02:54 – Designing the AI Agent Business Offer 06:38– Selling an AI Employee, Not an Agent 07:26 – Industries to Target (and Two to Avoid) 14:54 – Content Is Overpowered and How to Get Customers 17:51 – The Customer-Facing Tool Stack 20:49 – Building Agents Stack 25:51 – Model Picks: GPT 5.5, GLM 5.1, Kimmy, Opus 4.7 27:08 – Nick's Stack 28:14 – Why Obsidian Is the Second Brain Layer 30:22 – Live Walkthrough: Spinning Up a Cloud Computer in Orgo 33:53 – Cloud Computers vs. Mac Minis 38:37 – Building Agents and Structuring Workspaces for Customers 43:56 – Watchdogs, Observability, and Reliability 45:28 – Closing Thoughts on the Solopreneur Era Key Points Sell unlimited agents, unlimited usage, and unlimited support to remove friction; most customers actually use one to three agents. Avoid healthcare and finance to start; focus on legacy verticals like marketing, law, insurance, manufacturing, wholesale, and real estate. OpenClaw agents go for around 5K a month; Hermes agents can go for 10K a month. The full stack: Granola, Trello, Loom, Superhuman, Asana, Codex, Hermes, Orgo, Composio, Agent Mail, and Obsidian. GPT 5.5 is the recommended default model for tool calling; GLM 5.1 and Kimmy work for lighter tasks; Opus 4.7 fits long-horizon coding. Use agents to set up other agents — pair Cloud Code or Codex with MCPs like Perplexity, Context7, and X MCP for live docs. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND NICK ON SOCIAL Youtube: https://www.youtube.com/@nickvasiles Instagram: https://www.instagram.com/nickvasilescu/ Personal Website: https://www.nickvasilescu.com/

En el episodio de hoy, nos alejamos de la resaca de la feria para meternos de lleno en el desarrollo con IA y las novedades que están sacudiendo el ecosistema WordPress. 🤖 Cloud Code y el nacimiento de «Speed Clean Bot» Miguel Ángel nos cuenta su idilio con Claude Code y cómo esta herramienta le ha permitido desarrollar de forma frenética. ¿El resultado? Un bot de Telegram llamado Speed Clean Bot. ¿Qué hace? Le mandas un audio (incluso esos interminables de WhatsApp) y el bot hace magia: limpia el ruido de fondo, mejora la calidad como si tuvieras un micrófono profesional y, lo mejor, lo transcribe a texto usando Whisper. El objetivo: No es solo leer audios sin escucharlos, sino automatizar la creación de posts en WordPress a partir de notas de voz. La IA no nos quita trabajo, ¡nos da más porque ahora podemos hacer cosas increíbles! 🔌 ¿Qué es eso del MCP (Model Context Protocol)? Mariano nos trae el término de moda: MCP. Inventado por Anthropic (los creadores de Claude), es básicamente un protocolo que estandariza cómo una IA interactúa con un sistema. En WordPress: Gracias a los nuevos «AI Building Blocks», WordPress ahora tiene su propio MCP. Esto significa que ya no tienes que explicarle a la IA dónde está cada función; el protocolo le enseña directamente cómo publicar artículos, optimizar el SEO o gestionar comentarios. Adiós a la configuración manual: Es como una evolución de la REST API, pero diseñada específicamente para que los modelos de lenguaje «entiendan» las tripas de tu web. 🌍 Actualidad: De la WordCamp Asia a Polonia No podíamos cerrar sin un poco de salseo de la comunidad: Matt Mullenweg y las críticas: Se comenta el revuelo en la WordCamp Asia sobre la actualización de WordPress.org y el sistema de medición de contribuciones. Próxima parada: WordCamp Europe: Miguel Ángel ya ha sido «fichado» como voluntario (justo el día de su cumple). Nos vemos en Polonia, entre pierogis (empanadillas típicas) y charlas sobre el futuro del CMS. 🛠️ Herramienta recomendada: My WordPress Mención especial a My WordPress, el servicio que permite tener un WordPress persistente directamente en el navegador, superando las limitaciones de volatilidad que tenía WordPress Playground. ¿Te ha gustado el episodio? Si quieres que sigamos experimentando con bots, protocolos y empanadillas polacas, no olvides suscribirte y dejarnos tu valoración. ¡Nos escuchamos en el próximo capítulo! Métodos de contacto Enviadnos vuestras preguntas al grupo de Telegram. Apuntaos al canal de Youtube del podcast https://www.youtube.com/potenciapro Si nos queréis decir algo directamente lo podéis hacer a @potenciapro , @materron, @mpc, o en el grupo de Telegram Y si eres muy muy muy fan del podcast Echa un vistazo a cómo nos puedes ayudar en https://potencia.pro/se-prosperoso/

263 | Ein überraschender neuer Player im AI-Rennen, warum will SpaceX Cursor kaufen und müssen alle Startups sterben, wenn Corporates sie kaufen? Samuel und Alex diskutieren die Tech-Themen der Woche.Partner dieser Folge:ebay.comPrüft, wie Live-Commerce euer Business voranbringen kann. Keine Verkaufsprovision für neue Live-Seller in den ersten 6 Monaten. Mehr Infos: ebay.de/startliveMach das 1-minütige Quiz und finde eine Geschäftsidee, die zu dir passt: digitaleoptimisten.de/quiz.Kapitel(00:00) Intro(01:29) DasTelefonbuch.de - under the radar over the top(08:14) Finds of the week: Offline is the new shit und SpaceX kauft Cursor (vllt.)(23:31) Sprach-Post von Optimisten - Mirja(38:12) Ankerkraut - was wollen Corporates mit Startups?(48:35) Das Große Digitale Optimisten Linkedin Quiz(52:25) Geschäftsidee von Samuel: Outfraction(1:00:06) Geschäftsdiee von Alex: Urban Sports Club für KinderSo erreichst du uns:Sprachnachricht senden: https://www.speakpipe.com/digitaleoptimistenEmail schreiben: alexander@digitaleoptimisten.deLearningsGanzheitliche AI-StrategieEndgame der AI-Entwicklung wird als holistisches Ökosystem aus LLM, Code-Tools und Hardware beschrieben. Die Diskussion nennt Beispiele wie Claude Code, Cursor, XAI und SpaceX, um diese Verknüpfung zu illustrieren. Hypothese: Wer AI über digitale und physische Systeme hinweg integriert, könnte Wettbewerbsvorteile erzielen.Fractional Ownership als neues ModellDie Idee, teure Ausrüstung per Fractional Ownership zu finanzieren, wird konkret diskutiert und auf Kategorien wie Camping- und Outdoor-Ausrüstung, Kletter- oder Wanderausrüstung sowie Rennräder übertragen. Beispiele nennen anteilige Beteiligungen von 10 Prozent bis 20 Prozent an Vermögenswerten wie Ferienhäusern oder Luxusuhren; Upfront-Kosten könnten pro Anteil bei ca. 500 Euro liegen, plus jährliche Servicegebühren. Die Plattform soll Käufer bündeln und zusätzlich durch Abonnements laufende Einnahmen generieren, was laut Diskussion etwa 20 Prozent des Umsatzes als Subscription-Revenue bringen könnte.Retrieval-gestützter Wissensbot aus TranskriptenEine Hörer-Idee (Mirja) schlägt vor, einen Retrieval-basierten Bot zu bauen, der Transkripte als Wissensbasis nutzt, um Folgewissen abrufbar zu machen. Kernidee ist, eine App zu bauen, die auf die Transkripte der Folgen zugreift und Fragen zu einzelnen Episoden beantworten kann, idealerweise ohne Halluzinationen. Die Umsetzung erfolgt über Cloud-Code, Kontext-Import der Transkripte und eine einfache Benutzeroberfläche, um relevante Passagen schnell nachzuschlagen.Unternehmens- und Markenakquise im AI-ZeitalterEs wird diskutiert, dass große Konzerne wie Nestlé Marken wie Ankerkraut kaufen, um unorganisches Wachstum zu generieren, und dass Gründer später oft den Verkaufsweg wählen. Der Fall Ankerkraut (Rückkauf durch die Gründer 2026) wird als Signal genannt, dass Markenwert durch Übernahmen kippen kann und Shareholder-Value-Denken eine treibende Kraft bleibt. Für Gründer und Unternehmen bedeutet das: Exit-Planung, M&A-Strategie und das Verständnis der Rollenspiele in Kapitalmarkt-basierten Wachstumsstrategien frühzeitig berücksichtigen.KeywordsClaudeCursorSpaceX XAINestlé Ankerkraut RückkaufClaude Coding-Tools Einsatz im B2B UmfeldBenzinpreisvergleich Telefonbuch.deFractional Ownership Camping AusrüstungSecond Order Investing Anwendung MärkteMiroFish Tool ReaktionsszenarienOffline AI ModelleRetrieval SystemeOpenAI Konkurrenz ClaudeKI Coding Tools Markt

Today, we check in a year after the first Unsupervised Learning x Latent Space Crossover special to discuss everything that has changed (there is a lot) in the world of AI. This episode was recorded just after AIE Europe, but before the Cursor-xAI deal.Unsupervised Learning is a podcast that interviews the sharpest minds in AI about what's real today, what will be real in the future and what it means for businesses and the world - helping builders, researchers and founders deconstruct and understand the biggest breakthroughs.Thanks to Jacob and the UL production team for hosting and editing this!Jacob Effron* LinkedIn: https://www.linkedin.com/in/jacobeffron/* X: https://x.com/jacobeffronFull Episode on Their YouTubeWe discuss:* swyx's view from the center of the AI engineering zeitgeist: OpenClaw, harness engineering, context engineering, evals, observability, GPUs, multimodality, and why conference tracks now reveal what matters most in AI* Whether AI infrastructure has finally stabilized: why “skills” may be the minimal viable packaging format for agents, why infra companies have had to reinvent themselves every year, and why application companies have had an easier time surviving model volatility* The vertical vs. horizontal AI startup debate: why application companies can act as the outsourced AI team for enterprises, why some horizontal companies still matter, and why sandboxes may be the clearest reinvention of classic cloud infrastructure for the AI era* The “agent lab” playbook: starting with frontier models, specializing for your domain, then training your own models once you have enough data, workload, and user behavior to justify the cost and latency savings* Why domain-specific model training is real, not just marketing: how companies like Cursor and Cognition can get users to choose their in-house models, and why search, domain specialization, and distillation are becoming more important* Open models, custom chips, and alternative inference infrastructure: why swyx has turned more bullish on open source, why non-NVIDIA hardware is suddenly getting real attention, and why every 10x speedup can unlock new product experiences* What it means to sell to agents instead of humans: why agent experience may mostly just be good developer experience by another name, why APIs and docs matter more than ever, and how pretraining-data incumbents are compounding advantages in an agent-first world* Why memory and personalization may become the next big wedge: today's models mostly reward frequency of mentions, but in the future, swyx expects product choice to be shaped much more by personalized memory systems* The state of the AI coding wars: why coding has become one of the largest and fastest-growing categories in AI, how Anthropic, OpenAI, Cursor, and Cognition have all ridden the wave, and why the category may still have more room to run* Capability exploration vs. efficiency: why the industry is still in a token-maxing, experiment-heavy phase where people are rewarded for spending more rather than less* Claude Code vs. Codex and the strange stickiness of coding products: why first magical product experiences may matter more than expected, and why the bigger mystery may be why only a few names have emerged as real winners so far* What the end state of the coding market might look like: two major players, a longer tail of niche products, and possible disruption if Microsoft, Mistral, xAI, or the Chinese labs push harder into coding* Where application companies still have room against the labs: why frontier labs are trying to expand into verticals like finance and healthcare, but still leave space for focused companies that own the workflow and the last mile* Why coding may be a preview of every other AI market: the first category to truly go parabolic, the clearest example of foundation model companies colliding with application companies, and a template for how future vertical AI markets may develop* Why AI valuations now feel unbounded: from billion-dollar ARR products built in a year to trillion-dollar market caps, swyx and Jacob unpack how the AI market has broken traditional startup intuitions about scale and durability* Consumer AI vs. coding AI: why ChatGPT's consumer category may have plateaued on frequency and product design, while coding continues to feel like a daily-use category with real momentum* The next product frontier beyond coding: consumer agents, computer use, and “coding agents breaking containment,” with swyx's thesis that 2025 was the year of coding agents and 2026 may be the year they begin to do everything else* Whether foundation models are really killing startup categories: why swyx is less worried for early founders, more worried for mid-size startups and traditional SaaS, and why building something ambitious may now be the best job interview for a frontier lab* AI vs. SaaS and the internal culture war around adoption: the tension between AI-native employees who want to rip out expensive software and skeptics who think quick AI-built replacements create fragile systems* Why traditional SaaS may be under real pressure: swyx's own experience spending six figures on event and sponsor management software, the temptation to rebuild it cheaply with AI, and the broader question of whether teams will trust custom AI-native replacements* Biosafety, security, and frontier model access: why swyx raised biosafety at a dinner with Anthropic's Mike Krieger, why Krieger argued security is the bigger issue, and what restricted model releases reveal about Anthropic vs. OpenAI* The era of giant models: why 10T+ parameter systems may only be a temporary rationing phase before bigger clusters arrive, why labs may increasingly keep their most powerful models private for distillation, and why scale alone no longer feels like a complete answer* Memory as the slowest scaling factor in AI: why context windows have improved far more slowly than people hoped, why million-token context still has not changed most real workflows, and why memory may be the key bottleneck for the next generation of systems* What swyx changed his mind on in the past year: becoming more bullish on open models, more convinced that the top tier of agent startups behaves very differently from the median AI company, and more optimistic about fine-tuning and specialized model adaptation* “Dark factories” and zero-human-review coding: the next frontier after zero human-written code, where models not only write the code but ship it without human review, forcing companies to rethink testing and verification from first principles* Why RL and post-training may matter more than people assumed: even if the resulting models get thrown out every few months, the data, workflows, and domain-specific improvements persist* Synthetic rubrics, Doctor GRPO, and multi-turn RL: why reinforcement learning is becoming much more domain-specific and multi-step than many people realize, opening the door to much deeper customization* The next frontier after coding: memory, personalization, and world models, including why swyx thinks world models matter not just for robotics or gaming, but for giving AI something closer to lived understanding* Fei-Fei Li, spatial intelligence, and the Good Will Hunting analogy: the idea that today's LLMs may know everything by reading it all, but still lack the lived experience that turns knowledge into a deeper kind of intelligenceTimestamps* 00:00:00 Intro preview: AI coding wars, startup pressure, and market structure* 00:00:28 Welcome to the Latent Space × Unsupervised Learning crossover* 00:01:17 What AI builders are focused on now: OpenClaw, harnesses, and infra* 00:04:33 Why AI infra is harder than apps, and where startups can still win* 00:06:39 Should companies train their own models?* 00:09:28 Open models, custom chips, and the new inference race* 00:11:25 Designing products for agents, not just humans* 00:16:49 The state of the AI coding wars in 2026* 00:19:27 Capability exploration, token-maxing, and why coding is going parabolic* 00:21:41 What the end state of the coding market could look like* 00:23:50 Where app companies still have room against the labs* 00:27:02 Why AI valuations and market swings feel unprecedented* 00:28:56 Consumer AI vs. coding AI, and why sticky products still matter* 00:32:28 What the next breakthrough product experience might be* 00:32:53 2026 thesis: coding agents break containment and eat the world* 00:35:27 Are foundation models wiping out startup categories?* 00:37:33 AI vs. SaaS, vibe coding, and internal team tensions* 00:40:01 Biosafety, security, and the politics of restricted model releases* 00:42:19 Giant models, compute constraints, and the limits of scale* 00:44:30 Memory as the real bottleneck in AI* 00:44:57 Why swyx changed his mind on open models* 00:47:44 Dark factories and the future of zero-human-review coding* 00:49:36 Why post-training and RL may matter more than people think* 00:51:50 Memory, world models, and the next frontier of intelligence* 00:53:54 The Good Will Hunting analogy for LLMs* 00:54:21 OutroTranscript[00:00:00] swyx: Isn't that crazy? That number is just mind boggling.[00:00:03] Jacob Effron: What is the state of the AI coding wars today?[00:00:05] swyx: We're in a phase of sort of like capability exploration. The general thesis that I have been pursuing now is that the same way that 2025 was a year coding agents 2026 is coding agents breaking containments to do everything else.[00:00:16] Jacob Effron: Do you worry about the foundation models just getting into a bunch of these startup categories?[00:00:21] swyx: Mid-size startups. Yes.[00:00:23] Jacob Effron: What do you think the end state of this market is[00:00:25] swyx: for the market structure to, to significantly change? There would be[00:00:28] Jacob Effron: today on unsupervised learning. We had a, a fun episode and what's really become an annual tradition, a crossover episode with our friends at Latent space.Swix and I sat down and we talked about everything happening in the AI ecosystem today. What we thought of the various changes at the model layer, what's happening in the infra world, the coding wars, and a bunch of other things. It's a ton of fun to do this with someone I really respect and another great podcaster in the game.Without further ado, here's our episode. Well switch. This is, uh, super fun to be back with another unsupervised learning, uh, latent space crossover episode.[00:01:02] swyx: Yeah,[00:01:02] Jacob Effron: I feel like a lot of places we could start, but you know, one thing I always find fascinating, uh, about the way you spend your time is you obviously are like at the epicenter of this engineering movement and community, and you run these events and conferences and put on these.Awesome talks and, and I think just have a great pulse on the zeitgeist of what's going on.[00:01:16] swyx: Yeah.[00:01:17] Jacob Effron: Maybe to, to start just what are the biggest topics people are thinking about right now?[00:01:21] swyx: Yeah, so I just came back from London, uh, where we did a IE Europe and we're doing roughly one per quarter now, which Yeah, you've[00:01:27] Jacob Effron: really up[00:01:27] swyx: the, hopefully[00:01:28] Jacob Effron: up the, up the pace.[00:01:29] swyx: It's trying. We're trying to match AI speed, youknow?[00:01:30] Jacob Effron: Yeah, exactly. The tops would be completely different, I imagine. Uh,[00:01:33] swyx: yeah. You know, I definitely curate the tracks, like you can see what I think. When you see the track list and the, the speakers that I invite, obviously Open Claw is like the story of the last four or five months, and then be, be just below that.I would consider harness engineering, context engineering to be two related topics in agents and rag. And then there's a long tail of Evergreen stuff like evals, observability, GPUs, uh, and uh, LM infra and just general, just in general. We also have other updates on like multimodality and, uh, generative media, let's call it.Um, but I definitely, the, the first three that I mentioned are top of mind people. Yeah.[00:02:13] Jacob Effron: I think harness is particular like, so interesting. Um, you know, there was this tweet from Harrison Chase, the, the lane chain, CEO, that, that caught my eye recently where he said, you know, it finally feels like we have stability, uh, around the infrastructure for, uh, you know, around ai.And I think what. He basically was implying his like, look over the past two, three years as a company at the epicenter of AI infrastructure, it was a bit like playing whack-a-mole, right? You were constantly moving around with, however, the building patterns were evolving[00:02:36] swyx: for Harrison for sure. Right? Like he's basically had to reinvent the company every year since he started Lang Chain.Right? It was Lang chain, Ang graph and LP agents and like, uh, I think he's like one of the most nimble, adept sharp people about this. Yeah. Yeah.[00:02:49] Jacob Effron: Saying now, now is finally the time stability[00:02:51] swyx: this. Yeah.[00:02:52] Jacob Effron: Yeah. Um, do you buy that or what have you kind of make of that take?[00:02:56] swyx: I think that. It, it's very expensive to say this Time is different sometimes, but when you're just writing code, like it's actually okay to just like try to make a call and I think it may not even matter if this call is right or not.Like I just don't even care that much because you can be right on a thesis, but if you don't, you don't figure out how to monetize the thesis, then who cares if you said something first that said, um, it does feel like, for example. Uh, we went through a lot of different ways of passion packaging integrations up with, uh, with agents.And it feels like we've landed at skills, which is like the minimal viable format. Yeah. Which is just a markdown file, uh, with some scripts attached to it, and I don't see how it can be more simple than that. And so there is some justification for. The stability around harnesses. I feel like there may be more adaptation with regards to maybe like the real time elements or subagents or memory or any of those like agent disciplines, let's call it in, in agent engineering.Uh, but if, if the thesis is that, okay, you just want agents are LMS with tools in the loop with a file system, what they can do. Retrieval with, with skills and all these like standard tooling that now seems to be relatively consensus then probably. That makes sense. Um, I just think like there's no point trying to stake your reputation on this thesis that we're there because if it changes again, just change with it.It's fine.[00:04:33] Jacob Effron: Yeah. It's always, you know, I've always been struck by how that is. Much more challenging for infrastructure companies and application companies. Like obviously I think, yeah. You know, on the application side you've seen, you know, Brett Taylor from Sierra Max, from Lara. Like, they're like, look, we build, you know, what's ahead of the models and we're willing to throw everything out every three months, you know, as the models get better and better.Exactly. Yeah. But the thing you at least have there is you have. Uh, you have an end customer, right? That's like decently sticky. Um, you know, they will mostly stick, you know, they'll, they'll give you a shot at least of, of building these things. What I've always found more challenging, uh, at, at the kind of like, you know, reinvent yourself every three months of the infrastructure layer, it's like, you know, developers are definitely a, a pickier audience maybe than an accounting firm or, uh, you know, a bank.Yeah. And so it's definitely a, a, a more challenging position to be in to, to have to constantly reinvent yourself.[00:05:17] swyx: Yeah. Yeah. Yeah. And, and like when they turn, it's like. Very complete. Like, they'll leave to like the, the hot new thing, uh, because there's like no defensibility, I guess. Like e even, even if you are a database, like, uh, people can migrate workloads off databases.Like it's, it's a, it's a known thing. Uh, so I think like basically what we're talking about is the vertical versus horizontal, uh, debate in, in AI startups. And uh, the way I think about it also is just that like when you are. Um, Lara, when you are a bridge, like you are the outsource AI team, right? You, you are, your job is to apply whatever state ofthe art AI methods.[00:05:55] Jacob Effron: Yeah. Like this translation layer between model capabilities and your[00:05:57] swyx: own customers. Yeah. To, to the end customers and like, well, if they didn't have you, they would've to hire in house and they're not gonna hire in house so they have you. And like, I think that's like a reasonable, like very robust to any whatever trends and, and discoveries that people make in, in the engineering layer.I do think like there is, um. It like sort of useful horizontal companies being built, but they're all. Very much like, sort of like the reinventions of classic cloud in the AI era and the, the primary one being sandboxes. Yeah. Um, which like, it's another form of compute guys, like, let's not get too excited about it.But I mean, like the, the workloads are enormous.[00:06:38] Jacob Effron: Right.[00:06:38] swyx: Yeah.[00:06:39] Jacob Effron: It's interesting, and I feel like as, as part of this, you know, the questions that folks are asking around infrastructure, there's a lot around, you know, the extent to which companies should have their own AI teams and what they should be doing in-house.And, you know, uh, I think there's questions around should people be training their own models? Should people be doing, you know, rl, uh, in-house based on the data they have? I feel like, you know, one has to evolve their takes on this every, every three months with paces. But where, where are you at on this today?[00:07:00] swyx: I think, well, I mean actually all models have gone up. Um, and obviously I'm involved in cognition and also cursors doing, doing, uh, a lot of own model training. And I think that that is some part of the, what I've been calling the agent lab playbook, where you start off with the state of the art models from, uh, from the big labs and you, uh, specialize for your domain.But once you have enough workload and enough high quality data from your users, then you can obviously train your own models and like save a lot on cost and latency and all that, all that good stuff. Um, you also get like a marketing bonus of like calling it some fancy name and putting out some research[00:07:38] Jacob Effron: from my seat.I can't tell how much of it is like actual, you know, value that's provided to the end user. And how much of it is that marketing bonus? Right. It seems some combination of the[00:07:45] swyx: I think it's both.[00:07:46] Jacob Effron: Yeah.[00:07:46] swyx: Um, no, no. There, there actually is real value. Um, and you, you know that for a number of reasons. Like one, even when it's not subsidized, people do choose it as like one of the top four or five.This is both composer two and, uh, suite 1.6 I one of the top five models. Like in a, in a fair market? In a free market, yeah. In a, in a, in a model switch. Or people do choose it and like, it's not subsidized. Like, so that's as good as it gets. Uh, but beyond that, like domain specific models, for example. For search with, with both, which both companies have absolutely makes, makes a ton of sense.Everyone says like, yeah, we should always, always do this. And honestly like, I think the infrastructure for that is becoming easier with, um, like thinking machines tinker thing as well as primary like, uh, lab stuff. Yeah, I mean like, this is one of those like reversal of the, the bitter lesson where you first bootstrap on the large models and the general purpose models to get big.And as you get very well-defined workloads that are just high quantity but not high variance, um, then you just distill down to a smaller model and run that on your own. Right. Which like totally makes sense.[00:08:50] Jacob Effron: What I'm less clear on is the kind of DIY RL use case, which I think is really mostly around, you know, improved, uh, quality for, for different things.Obviously there's probably like more efficient ways to, you know, get a smaller model that's that's faster and cheaper. And it'll be interesting to see whether. You know, obviously you had, you know, uh, two, three years ago this whole case of companies that were, you know, pre-training and claiming better outcomes in, in their domains than getting kind of cooked as each model iteration improved.You know, I wonder whether that's a, a similar story plays out in the, uh, in, in the, our all space. Yeah, for the focus on, on on pure outcomes and quality, not the cost side, which clearly your own models for cost at scale makes a ton of sense.[00:09:28] swyx: I think there are this, there are two sides of the same coin.Like you basically always want to hold, uh, quality constant or trade off a little bit of quality for a drastic decreasing cost. And that's true for everyone. Uh, one element I wanted to bring out, which is very much in favor of open models, is custom chips. So this would be cereus, but also talu. And then there's a huge range of stuff in between.This has been a huge story this past year on just like everything non Nvidia is getting bid up, including like freaking MatX is working for, which is very, which is very rewarding for me, but I think one of those things where like, oh, like the suddenly, because the number of alternative. Hard, uh, hardware is increasing and the inference that you can get is insanely high.Like, um, we're talking thousands of tokens per second instead of less than a hundred. So the trade off for qua quality doesn't hold as much anymore because the speed is so high.[00:10:24] Jacob Effron: Have you seen a lot of companies go all in on the alternative chip?[00:10:26] swyx: So cognition has Yeah. On Cerebras, uh, and, and so has OpenAIUm, uh, and so no, I don't think so beyond that, uh, and that, do you think that's like a, that's mostly, that's foreshadowing of, that's, yeah. I used to be kind of a skeptic in terms of like, okay, so what if I get my inference at a hundred to a hundred tokens per second sped up to 200 tokens per second. It's only two X faster.It's not that big a deal. Um, but when you, uh, I think every 10 x does unlock a different usage pattern. Um, and you, we have proof in Talas and, and some of the others. That you can actually, um, drastically imp improve inference speed and what happens from there? I don't even really know, like it's, it's so hard to predict when entire applications just appear at once.Yeah. Uh, and it also isn't that expensive, right? So like, um, this is one of those things where like, I, I think the, the investment cycle is gonna be multi-year. Um, and I. Would caution people to not dismiss it too, too quickly.[00:11:25] Jacob Effron: Yeah. I mean, one other like infra question I was curious to get your thoughts on is obviously it seems increasingly a lot of the cutting edge infra companies are building for agents as the buyers of their product or users of their product, right?[00:11:35] swyx: Ooh,[00:11:36] Jacob Effron: and[00:11:37] swyx: another huge theme. Yeah. Yeah.[00:11:38] Jacob Effron: And I'm trying to figure out like what. What, what do you have to do differently about selling into agents? Um, are they just the ultimate rational developers? Uh, or is there, you know,[00:11:46] swyx: no, absolutely not. Um, I think they are easily prompt, injected and, uh, very tuned towards like, basically com compounding existing winners.[00:11:57] Jacob Effron: Yeah,[00:11:57] swyx: so like if, like, congrats if you won the lottery for getting into the training data right before 2023, because now you're like installed in there for the foreseeable future. But yeah. Uh, you know, one stat that Versal, uh, CTO Malta dropped at my conference was that there are now, uh, 60% of traffic to Elle's, um, like app arch, like admin app architecture for like configuring versal applications, uh, is bought.It's not, it's not human. Uh, so like your primary customer is agents now. Um, and it's mostly co like mostly coding agents, mostly people using CLI on CP or whatever. But yeah, I mean, I think. More. I, I think step one, if it doesn't exist as an API that agents can use, it doesn't exist. Right, right. Which I think is like, uh, it's a good hygiene thing anyway, to, to make everything API available, but not as like an extra, um.Push on like products, people to not only work on the ui, um, you should probably work on the on SCLI stuff. Beyond that, I think honestly there is like, so I, I come from the sensibility of, I think everything that you are trying to do for agents experience now, which is the term that Matt Bowman and Nullify is trying to coin, is the same thing that you should have been doing for developer experience.That you should have had good docs, you should have had a consistent API, uh, that is. Mostly stateless. Um, you should have, I guess, discoverable or progressive disclosure or like search or like whatever. And so now that people have energy in like finding these customers to do that, that's great. Um, do I believe in.Extending beyond that into something like a EO, um, for gaming The chatbots? Not necessarily, but obviously there's gonna be huge advantages when people who figure out the short term wins. Yeah. And short term wins can compound.[00:13:43] Jacob Effron: Do you think these compounding advantages to like the, the pre-training data cutoff companies, like, you know, obviously over some period of time, I imagine that doesn't persist.And so as you think about like. I dunno, three, four years from now what the, you know, selection criteria end up being. Do you think it still mirrors exactly what you were saying before? Like it's exactly what you should have been doing all along to sell a good product to developers?[00:14:01] swyx: It could be, except that I think in three, four years we'll probably have much better memory and personalization.So then general a EO or GEO doesn't really matter as much. So I think whatever memory or personalization system we end up with will probably d determine what you end up choosing much more. Than, than what is currently the case, which is just frequency of mentions, let's call it. Yeah,[00:14:26] Jacob Effron: yeah.[00:14:26] swyx: Uh, so you just spa quantity and I think that's, I mean, that's something I'm looking forward to.I do think, like, like, you know, I, I think that the fundamental exercise to work through for yourself is if you start a new, um, sort of. Uh, disruptor company. Now there's a, there's a big incumbent that everyone knows, like, like superb base. Super base is like, kind of like the Postgres, like database, uh, incumbent.If you wanna start like new superb base, how would you compete with them? And I don't necessarily have the answer, but I, I, I do think like people, like resend like relatively new. I think they would start like 20, 23 and still there was, there was a recent survey where like, people. Checked what Claude recommends by default.If you just don't prompt it with anything, just say, gimme an email provider and says, resent as in like 70, 70% of each cases. Like the fact that you can get in there with like such a relatively short existence, I think is, is encouraging.[00:15:14] Jacob Effron: Yeah.[00:15:14] swyx: I do think like. Um, you do want to do whatever it is to, to like to, to get in that Very short mentions this because, um, it's not gonna be 20 of them, it's gonna be like three.[00:15:26] Jacob Effron: No, definitely. It feels like, uh, you know, probably more, more consolidation than ever. Uh, or, or kind of like, you know, uh, a winner take most market than maybe the, the, the physics of go-to market in the past. Yeah. Might have, uh, enabled.[00:15:38] swyx: The other thing also is like, semantic association is gonna be very important, uh, in the sense that like, you want to do like the combo articles where you're like, use my thing with for sale, with blah, blah.And like that all gets picked up in a, in a corpus. And so that's. Probably one thing that you, you wanna do? Well, I don't know what else. Uh, it's, it's, it's, it's one of those things where like, I think I feel, I feel I'm behind, uh, I don't know how you feel about this, but like,[00:16:04] Jacob Effron: I think AI is just everyone constantly feeling like they're behind some, uh,[00:16:08] swyx: yeah.With,[00:16:09] Jacob Effron: I wanna meet the person that doesn't feel behind,[00:16:11] swyx: but like with, with ax, right? Like, so, so like, my, my stance was that exactly what I said before, like everything that you, that you should do for agents is something that you should have done for humans anyway. Yeah. And so. To the extent that you're just getting it more energy to, to do things for agents, great.But like, uh, it's hard to articulate what new thing apart from just like more spam, um, that you should be doing. Anyway, that would be my take right now. Um, I I, I do think like there, there will be more turns at this. I think the personalization turn that is coming, um, will be big. And I don't know what that looks like because like basically we're kind of, we feel kind of tapped out on the memory side of things.[00:16:49] Jacob Effron: Yeah. I, I guess since we last chatted, you know, you, you took this role over at cognition, um, and you've obviously have a, have a front row seat to the AI coding space today. You know, I feel like coding in many ways. You know, people view it as this, like, I mean, besides being like the, the mother of all markets and this massive opportunity, I think it's kinda a preview of like, what's to come for many other spaces.Both. Yeah. You know, I feel like agents are most advanced in coding. I also feel like the, you know, competition between foundation models and application companies, you know, and, uh, mirrors what we may see in other spaces. And so maybe for our listeners, can you just lay out like what is the state of the AI coding wars today?[00:17:25] swyx: Um, it is massive, right? Like, uh, and I don't think necessarily, last time we talked about this, we appreciated the size of what[00:17:32] Jacob Effron: No, I wish we did.[00:17:33] swyx: I state of AI coding wars today, um, both opening eye philanthropic have made it their p serials to competing coding. Um, and. Tropic is like 2.5 billion in a RR just from Cloud Code.The way they recognize a RR is. Opt for debate, uh, open ai. I don't think the, a public number is known, but let's call it 2 billion as well. And then cursor is like, rumored to be 2 billion, you know? And, and those, those are like the public numbers that are known? Yeah. Um, so like huge markets that have just been created in the past one year.Like, like anthropic, just like Claude Code just recently celebrated their one year anniversary, which is, yeah, pretty nice. Um, so, and then I think, like the other thing that I see is there's, there's some other people who are like, oh, here's like the, the sort of relative penetration of, uh, Claude use cases, right?Like, and it's like coding 50% and then legal, whatever. Health, uh, it's like the, the remaining ones. And there was a very popular tweet that was like, okay, I'll look at the, the empty space and all these other use cases. If you are a new founder today, you should be betting on the other stuff because on, on a sort of catch up Yeah.Theory and my. Consider my, my pushback is the same pushback that, uh, I had on app over Google, which is like, well, well why is this time different? Like, why, if it went from let's say 10 to 50% in the past year, why can't I keep going? Uh, and like getting that wrong is actually a very painful one because you could have just did, did the momentum bet.Instead of the mean reversion bed. So I, I, I think that that is the, the state of things now that people are very, very much into psychosis. Um, they're are getting rewarded for spending more rather than spending less. And I think we're not in that phase of efficiency. We're in a phase of sort of like capability exploration.So I think people who are more crazy, who are more. Uh, creative, um, get rewarded comparatively. Yeah.[00:19:27] Jacob Effron: Well, it's interesting. I mean, it feels like behind these like token maxing, leaderboards and whatnot is this, it's like the first phase of this transition from a workforce perspective is you just gotta show your employer like, Hey, I, I use these tools.[00:19:37] swyx: Here's my nu number of tokens I cost, and that's it. They don't care about the quality. Right. It is, uh, maybe distasteful to someone who cares about the craft and, and all that. Um, but directionally everyone just wants you to go up regardless. And so, um, there it is not very discerning. It's, and it's probably very sloppy, but I think it's net fine because we're still probably underusing ai just in generally.Yeah. Um, and so I think that's like very interesting. Like we had on the podcast, uh, Ryan La Poplar from OBI, who spends a billion tokens a day. Yeah. Um, and that's for those county home, it's like something like 10,000 worth, $10,000 worth a day of API tokens. If they, they did market rates, um, and like most of us can't afford that.Yeah. But like. And, and, and probably a lot of what he does is slop.[00:20:25] Jacob Effron: Right.[00:20:25] swyx: But like, he's going to dis, he's like, if there were a new capability, he would discover it first before you because he was, he was trying and you were not trying. Right. And like, you only do things that work like, well, good for you.But like the, the people who are going to discover the next hot thing are living at the edge.[00:20:42] Jacob Effron: Right and increase in living at the edge of just having the compute budget to like run these experiments. I mean, kind of similar to what living at the edge on the research side has always been. You know, it was constrained in many ways by the amount of compute you had to run these experiments.It feels similarly on the, almost on the builder or like actualizing these tools now.[00:20:56] swyx: Yeah. The other thing that's, I mean, very obvious is philanthropic is kind of like the high price premium player. Um, that where, you know. Restricting limits or restricting model releases even is like the name of the game.Whereas Codex is like, come on in guys, use our SDK, use our login and we don't care. We're gonna reset limits. Whatever you do want to try to exploit the subsidies where you can get it. And definitely Codex is super subsidized right now. Gemini also very subsidized. Um, and. Comparatively, like, I think you should make, Hey, I guess while, while that's going on, it's not that bad to be a capabilities explorer on just the $200 a month plan from Cloud Code or from OpenAI.Um, and, uh, I I, I, my sense is that people aren't even there yet.[00:21:41] Jacob Effron: How do you think this, like, market ultimately plays? I mean, it's obviously such a big market that, you know, any slice of that market is interesting for, for anyone going after it. But I think what, what makes people so interesting in the coding market particularly is it feels like it's kind of this.Foreshadowing of what will happen in other, you know, any other kind of application market that the foundation models eventually turn to and are all their models against and gather data around. And so how do you think, you know, like does there end up being room for lots of different kinds of players or like, what do you think the end state of this market is and is that, do you think that's applicable to other markets?[00:22:10] swyx: I feel like there will be, I mean. Status quo is probably the most likely outcome, which is there are two big players and there's a small range of longer tail people that, um, fit other use cases that the, the two big players don't. That feels right to me. I think that, um, for it to, for the market structure to, to significantly change there would be, there needs to be significant change in like the economics or like the, the brand building or like the, the, the, the value propositions of the, of the companies involved and I.Haven't seen any in the last six months that, that have really changed the stories materially. So I feel like they would just keep going until something, something else happens. Something else happens, meaning like Microsoft wakes up and like goes like. Guys, we have GitHub, we have, uh, you know, we, we, we'll, we'll do something much bigger here than other, other than just copilot.Um, and, uh, that would be a big change. Um, MSL has put out a model now, and I was in a breakfast with, uh, Alex Wang, where they were like, yeah, like, we, we really, really want to go after the coding use case. We haven't done anything yet, but like, don't underestimate them. Right. Um, and, and similarly for the Chinese labs.Um, I think they're trying to go after it. Like ZAI is doing stuff. GLM uh, ZI and GLM is same thing. Um, uh, and, and so it's, so like everyone's trying to get a piece of that pie. I, I feel like the, the status quo has been pretty stable for the past, like almost a year I'll say.[00:23:39] Jacob Effron: Yeah. And is the room for the, not like, you know, for, for the application companies more on like the enterprise side or like where do the, where do the, like what surface area do the model companies leave for application companies?[00:23:50] swyx: Yeah, that's a good one. Um. It's very much evolving. Um, it, I, I, I will say because opening I did not have this, the, this level of attention on coding. Yeah. Uh, a year ago. We just don't have that much history. Right. Um, and it seems like, for example, so the big push at Open I now is the Super app. Um, is that a consumer thing?Is that like a products like. Portfolio rationalization thing, how much is that gonna take away attention from coding at the time when they actually do want to put more coding? I think it's, it's very unclear. So I do think like there's, there's all these, like in both big labs, there's. Uh, sorry. Both of the, and, and drop and, and deep minus and XAI are are separate cases.Um, they are trying to see the other time expansion areas. So cloud code for finance. Yeah. Um, uh, cloud cowork, all those, all those things. Whereas I think cursor and cognition are like comparatively just focused on coding and so I, I do think they leave space and I do think for the other verticals that also means the same thing.Right. That, uh, that they're not gonna be that. Um, intensely focused on, on, on that domain. Except for, I, I think I would mark out finance and healthcare as like the next ones, um, that they're clearly going after. Uh, I, I would say comparatively, healthcare seems more thorny. There, there, there've been some announcements about it, but like, I would respect the, the finance work a lot more just because like the, the path to money is a lot clearer.[00:25:12] Jacob Effron: Yeah, no, I mean, obviously like, I, I think, you know, maybe similar to, to the space that's being left in these other domains, you know, there's obviously. Uh, a lot that's required to actually implement these tools in enterprises, uh, versus, you know, maybe just giving them, uh, giving model access to, to folks outta the box.[00:25:27] swyx: Yeah, yeah. Yeah. So the, the agent lab thing is like, we'll do the last mile for you. Whereas I think the model labs tend to just trust the model and, and be minimalist about it. Both of them work.[00:25:38] Jacob Effron: Yeah.[00:25:38] swyx: I, I don't, I don't necessarily think one, uh, beats the other, uh, for every, for every use case. Um, all I, all I do know is that it does seem like.Uh, the large enterprises do want a dedicated partner that isn't just the model labs, which is kind of interesting.[00:25:55] Jacob Effron: We, we've been in this phase of, of pure capability exploration. And so I think nothing has been, you know, better for the large labs, right? I mean, they're always gonna be, uh, uh, the frontier of, of capability exploration.And so I think have a very good relationship with a lot of these enterprises. But ultimately over time, like. The, uh, the incentive structure of these labs is always gonna be maximal, you know, token consumption for, uh, for the end customers they work with. And there's just, I think, so few companies that have actually gotten to massive scale.Maybe coding again is the most interesting. So it's the first space that really is just completely gone, you know? Yeah. You must love it every day. Like absolutely insane. And. I think it[00:26:32] swyx: gets even. Okay. I mean, like, I think we, we say good things about crystal cognition, but the sheer liftoff of like both end UPIC and open ai.‘cause they, they, they have independent valuations. I mean, let's throw an XEI in there because it's now I ping at 1.2 trillion. That number is just mind boggling. Like I, I feel like in normal investing or normal startups, there's kind of like a ceiling market cap or valuation. Totally. That, that like you, you reach and you go like, all right, let's, it's gonna be chiller from now on.And these guys are not slow down. No.[00:27:02] Jacob Effron: Well, I also think the dynamic is fascinating about some of these later stage companies is, is, you know, in the past, I feel like in, in venture world, if you got to a certain level of scale, the question around you was really more a valuation question. And this is like why there was different phase, like, you know, types of venture people did and like the late stage growth people were just incredible at like, you know, a little bit of what's the ultimate market opportunity of this company, but also what's the right way to, to value it.Like we know it's, it's in some bands of an outcome that is like. Sure there's some variance to it, but it's like relatively understood what that bands is and then maybe you get over time surprised to the upside. Whereas any kind of like later, even the labs themselves, any later stage company, the bands of which that company might be worth right now, even in a year or two years are so massive because of how fast the ecosystem changes that it's like.Even for later stage companies, every three months could be an existential level event to the upside to the downside. Yeah. Um, and I think that, like, you are obviously seeing it in the, in the positive with code, which, you know, if you think about a company like philanthropic, you know, that. For a while, it was like unclear if they were going to have access to enough capital, um, to really stay in the, in the race, right?And then coding hit at the exact right time. They had the perfect model for it. They executed brilliantly. Um, and you know, now are, are, you know, uh, you know, one of the most valuable companies in the world.[00:28:13] swyx: Uh, at the same time, I, I don't find, I, I have zero sympathy for opening eye because they're crushing it and they're all rich.You know, this is like a high class champagne problem to have to, uh, to be number two at coding or whatever. Like, who cares? Like, you're, you're doing great.[00:28:27] Jacob Effron: Yeah. It's funny though. I can't even, I mean, you would be closer to this, uh, you know, even that you're in the AI coding space, but it's like a lot of people I talk to think Codex is just as good, if not better than Claude Code.Right. I think one thing that I've been really surprised by, and maybe, maybe Cloud Code is a better product in some ways, I'm curious your thoughts is just in consumer AI with chat GBT. You saw this big first mover advantage, right? Where admittedly today, like, I don't know, Claude Gemini. Great products.Not sure, not abundantly clear chat GBTs any better, but like. People stick with chat, GBT, it's the first thing to introduce them.[00:28:56] swyx: They stay, but they're not growing anymore. I don't know if you've seen[00:28:59] Jacob Effron: Right. But that to me is more of like a, a, a product problem than it is. They're not like, it's not like they've like lost share to someone else.My understanding is the overall problem with consumer AI today is much more of a how do you take this tool and, you know, for, for folks like us, like knowledge workers, it's like this incredible magic tool, but it's not necessarily a daily active use tool for a lot of people around the world today. And what are the like products?It's, it's kind of a category wide problem. Like in coding, for example, like. The entire space has gone parabolic. There may be some relative growth in, uh, in other consumer AI players, but it's not like consumer AI as a category is like going parabolic and they're not capturing most of that thing. I think it's actually the larger problem is much more, hey, the category has kind of hit a bit of a plateau of people haven't figured out how to bring, you know, tons more users on board.Yeah, yeah. Or increase the frequency of those users. And so it seems more of a category wide problem than it is, you know, a massive market share of change. I was gonna draw the comparison to, to the coding space where Claude Co is the first product, obviously, to introduce people to this magical experience.You know, by all accounts, codex is, is pretty damn close to as good, if not better. Um, but like still that first product, you, you would've thought that would not be a super sticky, uh, you know, product surface area. And it actually has, it turns out, I, it feels like the first lab to introduce you and experience really does, uh, keep a lot of, uh, a lot of the focus.[00:30:12] swyx: I, I think. M maybe it's like still, still early days. You know, Chad, BT is like three plus years old and Yeah. Cloud code is only one. Just turned a year. Yeah. So give it time, you know? Yeah. Like, yeah. I mean, definitely sometimes a lot of people have switched from to Codex. Maybe that will keep going. I, it's like really hard to tell.Uh, yeah. I, I, I do, I do think that. Because we are in this like, high volatility, high temperature phase. Um, the loyalty and stickiness to first movers and category creators, I don't think is as high as it might be in some other, uh, areas in our careers that we've looked at.[00:30:47] Jacob Effron: Yeah. Though, I mean, I've been surprised by the cloud code thing.I, I would've thought that, like, in many ways I always worried about the[00:30:52] swyx: enterprise. You think you would've been gone by now?[00:30:53] Jacob Effron: Not gone. But I would've, I I always worried that the, that the consumer business of these companies would be quite sticky. And then the enterprise API business. Uh, was actually like, you know, in some ways like your least loyal buyers, like they would, they would move to,[00:31:05] swyx: right, right.But, but they worked out that it wasn't the enterprise API it was enterprise product.[00:31:09] Jacob Effron: Totally. And maybe that was the, that was the secret that like, but the amount of lock-in or just default behavior that has happened in that space, uh, is, is more than I might've imagined with two products that by all accounts are pretty damn similar.Yeah.[00:31:22] swyx: No fight there. Uh, I will say I do think that Codex is still in like a catch up. Like in terms of personal experience. Um, the only thing I like out of, out of Codex is the, is like Spark and like yeah. Uh, the, I, I feel like the skills integration is a little bit better. I feel like, uh, the, the speed is a bit better.Maybe ‘cause it's in, is written in rust or whatever. Um, very minor things that you like. Almost like telling yourself rather than like objectively assessing between two, two of them. I, I, I do think, like vibes wise, I think that's going on. Um, the, the, you know, I, I feel like the, the missing questions, uh, in, in this whole debate is like, why is this so concentrated in only two names, right?Yeah. Like, um, how, where, like, where is the Gemini? You know, presence, where's the Xai presence? Um, and like they are trying, it's just they haven't made that much progress yet.[00:32:12] Jacob Effron: But what the, what the Claude Co moment does show, and it actually in some ways makes you a little more bullish on the potential for someone else to catch up because it does feel like if you're the first person to introduce some magical net new product experience, that that actually might be stickier than one might have imagined.[00:32:27] swyx: Right, right, right. Okay. Yeah.[00:32:28] Jacob Effron: And so it's, everyone can believe they have shot[00:32:29] swyx: that. What do you think that new product experience might be like? I, I, it's, it's like, and this is a failure of imagination on my part. Like, I always wonder, like, people always say this like, well, the, the thing that will save us is like being first to the next new thing.Like what is it?[00:32:41] Jacob Effron: Yeah.[00:32:42] swyx: It's like,[00:32:45] Jacob Effron: I dunno, something around like, uh, consumer agent, computer use, like hybrid. I think, obviously, I think we're like scratching the surface on the consumer side.[00:32:53] swyx: So my, my current theory is like the. Open claw is like a vision of things to come.[00:32:58] Jacob Effron: Totally.[00:32:58] swyx: Um, and uh, it's good that O open I has like the association with open claw, but by no means do they have the rights to win it.The general thesis that I have been pursuing now is that the year the same way that 2025 was the year of coding agents, 2026 is coding agents breaking containment to do everything else. Um, and so coding agents continue to still win, but because they generate software and software eats the world, so like, it's kind of like the trans.Associated property of like software, eat the world, coding agents, eat software, therefore coding agents eat the world. Um, which is like an interesting,[00:33:30] Jacob Effron: yeah, and breaking containment always an easier phase phrase in the consumer context than the enterprise one. You've seen people run these really cool, uh, experiments in their own personal lives.I think like,[00:33:37] swyx: yes.[00:33:38] Jacob Effron: Figuring out, you know, how you, obviously everyone's focused, you know, on the enterprise side now around how you create these experiences. I feel like the vibes, you know, people love to have these narratives of like, everything is completely shifted. It's like I actually, you know, open AI.Organizationally, uh, you know, volatility aside is, you know, great products, great team, great models like everyone else in the world is incentivized for there to be. Two, three more. Everyone would love more like great model companies. And so I feel like the, the natural forces of the world revolt when any one company, you know, is too much the star of the show, right?There's so many people in the ecosystem that are incentivized for that not to happen. And so I think I'd be shocked if we don't have. Uh, uh, reversion of vibes, not maybe completely the other way, but at least a little bit more equal at some point over the next six, 12 months.[00:34:24] swyx: I, I think there's just a kind of different stages when, when you talk about the world, one wanting more model companies, I talked think about like the neo labs.[00:34:30] Jacob Effron: Yeah.[00:34:31] swyx: And I mean, I don't know, is it fair to say none of them have really broken through in the past year?[00:34:35] Jacob Effron: I think that's totally fair,[00:34:37] swyx: which is rough. Um, and well, how are we gonna, how are we gonna grow that diversity in, in, in choice, like. Um, that's, this is it.[00:34:46] Jacob Effron: Yeah. It'll be really interesting to see what, what, what ends up happening with that.And you've seen, you know, folks like Nvidia, you know, very incentivized to make sure there's, there's a broader platform of, of other model providers.[00:34:57] swyx: I think, uh, I don't know people say this, but I, I, I don't think they try it hard. Nvidia tries harder to build neo clouds[00:35:05] Jacob Effron: Yeah.[00:35:06] swyx: Than neo labs.[00:35:07] Jacob Effron: Well, they try pretty damn hard to build neo Cloud, so[00:35:09] swyx: that's,[00:35:09] Jacob Effron: yeah.[00:35:10] swyx: But like, you know, let's call it like the, the core weaves of the world, much happier place in the, you know, than any neo lab built on top of them.[00:35:18] Jacob Effron: Yeah. That one might argue it's, it's easier to, to enable a neo cloud to be successful than it is. Uh, you can't will a neo lab into existence the same way you, soNvidia[00:35:25] swyx: has more direct control over it.Uh, for sure.[00:35:27] Jacob Effron: What else is kind of catching your eye today on the startup side? I mean, you worry, there's obviously this whole narrative of like, you know, the foundation models, you know, they announced a product and every stock goes down 15%. Like[00:35:36] swyx: Yeah.[00:35:37] Jacob Effron: Do you, do you worry about the foundation models just kind of eating into to a bunch of these startup categories?[00:35:43] swyx: Not really. I, I think actually like. As, uh, there's, there's, okay, there's, there's, there's the, there's the point of view of like being an investor in startups, and there's a point of view of like, do you wanna start something? And I think honestly, like the, the downside for all these is so. Minimal in, in a sense of like, the worst you do is you just get hired into one of these labs anyway.So I, I think the, the market for people who just do things and try things and try to execute in like a competent way, even if like it doesn't work out commercially, even if it just wasn't that great anyway. Like, but like that's your job interview to go into, into one of these things anyway, so, um, I don't feel that.From a, from a very, very small startup perspective, mid-size startups. Yes. Uh, I will say there's been a lot of dead, um, LM Infra, a lot of LM infra consolidation like the, the, uh, lang fuses of the world getting absorbed into, into click house. And I, I think. Like people have maybe worked out the domain specific playbook, uh, and like, I think that's okay.Um, and, and yeah, I'm not that, not that worried about, uh, okay. So, um, I, I would say I'd be more worried about traditional SaaS, like low NPSS. This is the whole AI versus SaaS debate that has, that's been going on. Uh, and, and like literally I'm going through that exact thing in my company where, so I like kind of.Thinking through this on a very visceral, visceral level, right? On one hand you have the people who say you vibe coders don't appreciate the amount of work that goes into A-A-C-R-M and like, yeah, you think you can rip out Salesforce? So did the 30 entrepreneurs before you, right? Like, like, you know, you classically underestimate the things that you don't.Deeply, no. And, and, and target audience is not you. Uh, at the same time, like we have never been able to build software so easily and customize software so easily and like Yeah, you're not gonna use 90% of the things in Salesforce. So like, yeah. What's the typical, so what have you, what[00:37:33] Jacob Effron: have you done internally?[00:37:34] swyx: So we have there the main SaaS that we do for event management and sponsor management. That's, and we paid 200 KA year for that. Not, not huge, but like chunky for, for, for my, my scale. Um, and like, yeah, I could probably spend 2000 and, and build like a custom version of that. Um, the, the, the trick has been dealing with my, the rest of my team and getting them on board.Yeah. ‘cause I'm the most ethical person on my team, but like, I can't make that decision myself. And I think in the same way I've been telling with other CEOs team leaders as well, it's like, well you can be super cloud pilled. You can be super LM psychosis and that you think that's okay, but you like you have to bring your team with you.And I think like there, the sort of widening disparity in LM psychosis in companies is causing real s real riffs because. And on one hand, on one hand, the people who are less AI native are not getting with the picture. They're not, they're actually like behind, they're actually not waking up to the fact that like you, everything you think is necessary is not actually that necessary.And in fact, exactly would be better of you if you just like held your nose and went in and when came out the other side. Yeah, only talking to agents in natural language and like your life would actually be better and you just, you're just like close-minded. There's that perspective. The other perspective is, oh, you vibe coder.You, you did this in a weekend and you got the 80% solution and now the rest of your employees. Have to pick up the rest of your s**t, right, that you, that you thought you were, you were such hot, amazing, uh, uh, at, but like, actually you didn't figure it out. And like, actually LMS are still useless at this and blah, blah, blah.So like, I think there's this huge debate going on in every company right now. Um, and like, um, you know, I have a small microcosm of it, but like, yeah, it, it's making me hesitate to, to pull the trigger. But like I will at some point, it's like maybe I've put it off for one year, but not like five. Yeah, but like, so, so like SaaS is definitely getting squeezed.Um, it does make me wonder, like, I, I do think that there's an opportunity for a more AI native, um, system of record thing that is not just Postgres. Um, or not just MongoDB, although both are very good. Maybe it's like a convex or like people Yeah. Bring up convex a lot. I don't know, like, like, I, I just feel like the sort of quote unquote firebase of, of AI apps isn't really a thing yet.Um, beyond what we have. Uh, which, which is fine. It's, it's, it's just. We could probably start in a more sort of rapid iteration cycle first before scaling up to like a Postgres or MongoDB, which are more sort of old tech. I was at a dinner with, uh, Mike Krieger, the CPO of en philanthropic, and, and he, we were just kind of going around the room going like, what are people most worried about?Yeah. And, uh, for me, uh, I, instead of security, I brought up biosafety. Yeah,[00:40:21] Jacob Effron: classic.[00:40:22] swyx: Um, actually, like I said, it was. Cliche and classic, and the rest of the table were, were like, what do you mean? Someone sitting at home can manufacture a virus that wipes out half of humanity,[00:40:32] Jacob Effron: almost like the OG Jeffrey Hinton.Like, this is why you should be scared.[00:40:35] swyx: I'm like, yeah, like the read the, you know, risk reports. Like this is like the thing. Um, I think, and Mike was just sitting there knowing he was sitting on Mythos and going like, actually it's security. Um, and I think like, um, I think the, there's, there's, part of it is.A very good marketing. Like too good. Yeah, like I would actually advise and topic to tune down the marketing because also it's, it is just a very good model and you don't have to make so many marketing claims around it. At the same time, it is not really a private model. If you give it to 40 companies.Each of whom have like 10,000 employees or whatever. Right. It's not, it's not private, it's, it's like there's bad actors in there.[00:41:18] Jacob Effron: Yeah. Hopefully, hopefully not as, uh, as bad as releasing it widely, but, uh, no, I mean, it's an interesting. You know, it's an interesting case study for how all, I mean, many model releases might, I mean, you know, this might be the first model release that looks like the rest of ‘em from from now on, right?[00:41:31] swyx: It, it, so it's, it's the, there's an overall product strategy, uh, for anthropic of like bundle, uh, you know, restrict access bundle, uh, product with model maybe.Whereas, uh, OpenAI has definitely been a lot more sort of. Philosophically aligned on like, we will just enable access everywhere and we don't know what you, what will come out of it. Right.[00:41:51] Jacob Effron: Right. Though, I mean, this current moment, uh, obviously the cynical take is also just ties to the amount of compute that both companies[00:41:56] swyx: Yeah.Right, right, right. Yeah, I think, I think that's true. I I do think like the, the, this is the, the, the scale, the dawn of like larger than 10 trillion parameter models is very interesting. I don't think it, I think it's a temporary phenomenon because we have much larger compute clusters coming online for everyone over the next like three, five years.It's, and this is like already written in, in the cards.[00:42:18] Jacob Effron: Yeah.[00:42:19] swyx: So to the extent that like, you know, will we have rationing of models, uh, above 10 trillion, uh, in like two years? I don't think so. I think everyone will have no, we'll just[00:42:29] Jacob Effron: have rationing of the next phase.[00:42:30] swyx: Right. Right. But like, that's as it should be almost like, um.My, my classic example, which I, this is just me theorizing, not anything confirmed by Google. When Google announced Gemini, they actually announced three sizes, which was Flash Pro Ultra. They never released Ultra. They only have Pro and Flash. Um, so my theory is they have ultra sitting in a basement and they just could distilling from it for, for flashing pro.Um, which like, yeah, I mean, I, I actually think that's. As it should be for any lab that they, that they do that.[00:43:02] Jacob Effron: Yeah. Just because those are the models that people actually wanna end up using. And it's just like cost prohibit.[00:43:06] swyx: It is more, yeah, it's cost. Yeah. It's, it's not the want, it's just, just, just the cost.Um, I do think, like, uh, it is interesting that, uh, for a while I was, I was considering the theory that models capped out at two, 2 trillion, and I think that's proving to be wrong. And well then if I'm wrong, how wrong? How wrong am I? Do we do 200 trillion? Do we do two quarter trillion, whatever? Um, and I don't think we have the straight answer to that, but like, uh, it's interesting that we are continuing to scale number of pers when everyone kind of assu like can see that we're not going to get like the next thousand or 1 million x from this paradigm.So like the others, like the alias of the world are working on other. Um, model architecture improvements. We need a different scaling law, I guess, because like, we're, I, I feel like people already already feel like we're tapped out on this. Like the, the end, the end state of this is we turn most of the world into data centers and like, I don't know.I don't know if we want that.[00:44:08] Jacob Effron: Yeah, I mean, uh, if the, if, if, if the return of intelligence are there, maybe, uh, maybe not so bad.[00:44:13] swyx: I, I, I think there, there's just a sheer amount of like, like un scalability that like is wrangling people's sensibilities right now. Um, especially in terms of like context lengths.Um, my classic quote is that context length is like the slowest scaling factor in, in lms.[00:44:30] Jacob Effron: Yeah.[00:44:30] swyx: Um, we, like, we took maybe. Three years to go from like 4,000 context length to a million and that's about it. Yeah. Like Gemini has had a million token context length for two years now. Um, and no one's using it.Like, so like yeah, it's memory. Memory is probably gonna be the, the biggest limiting constraint on all these things.[00:44:50] Jacob Effron: Yeah. Certainly seems that way. I guess I'm curious over the last year since you recorded last, like what's one thing you've changed your mind on?[00:44:57] swyx: I feel like I was kind of bearish on open models like last year.Um, in a sense of, like, I, I had just done the podcast with an Al[00:45:07] Jacob Effron: Yeah.[00:45:08] swyx: Of Braintrust where he, and he, I mean, you know, he has a good cross section of all the top AI companies and he says market share of open source is 5% and going down. Um, I think that's changed. I think it's going up. Um, and even if,[00:45:22] Jacob Effron: even though the capability gap does seem to be increasing.Spending on the[00:45:26] swyx: time. It's hard to tell. Yeah, it's, it's really hard to tell. ‘cause like, okay, for, for listeners, capability gap increasing is like on public benchmarks. And let's say you're comparing mythos versus like, I don't know, G-T-O-S-S or like GLM 5.1. And, um, it's, it is really hard to tell. ‘cause even if they were closing, you will also not believe that they were closing that much because it's very easy to gain the benchmarks.Yeah. So you just don't really, really know. Um, all you know is like. Uh, there's somewhat objective open router stats on like what people choose in a free market. And people do choose some of these open models in significant volume, except that a lot of them are heavily discounted. So you need to kind of like price adjust, uh, these things.So even if, even if that were true, which I, I'm not sure, like I, I, I feel like the numbers just up now instead of down. Uh, I think the. Separation between what the top tier agent labs

Early bird discounts for the San Francisco World's Fair, the biggest AIE gathering of the year, end today - prices will go up by ~$500 tonight so do please lock in ASAP!From near-universal AI tool adoption inside Shopify to internal systems for ML experimentation, auto-research, customer simulation, and ultra-low-latency search, Mikhail Parakhin joins us for a deep dive into what it actually looks like when a 20-year-old, $200B software company goes all-in on AI. We cover why Shopify has become much more vocal about its internal stack, what changed after the December model-quality inflection, and why the real bottleneck in AI coding is no longer generation, but review, CI/CD, and deployment stability.We also go inside Tangle, Tangent, SimGym, which are three major AI initiatives that Shopify is doing to make experimentation reproducible, optimization automatic, customer behavior simulatable, and search and catalog intelligence faster and cheaper at scale. Along the way, Mikhail explains UCP, Liquid AI, and why token budgets are directionally right but often measured badly, why AI-written code can still increase bugs in production, what makes Shopify's customer simulation defensible, and what he learned from the Sydney era at Bing.We discuss:* Mikhail's path from running a major Microsoft business unit spanning Windows, Edge, Bing, and ads to becoming CTO of Shopify* Why Shopify is talking more publicly about AI now, and why staying at the frontier has become necessary for the company* Shopify's internal AI adoption curve, the December inflection, and why CLI-style tools are rising faster than traditional IDE-based tools* Why Jensen Huang is directionally right on token budgets, but raw token count is still the wrong way to evaluate engineering output* Why the real unlock is not more agents in parallel, but better critique loops, stronger models, and spending more on review than generation* Why AI coding can still lead to more bugs in production even if models write cleaner code on average than humans* Why Shopify built its own PR review flow, and why Mikhail thinks most off-the-shelf review tools miss the point* How PR volume, test failures, and deployment rollback are becoming the real bottlenecks in the agent era* Why Git, pull requests, and CI/CD may need a new metaphor once code is written at machine speed* What Tangle is, and how Shopify uses it to make ML and data workflows reproducible, collaborative, and production-ready from the start* Why Tangle is different from Airflow, and why content-addressed caching creates network effects across teams* What Tangent is, and how Shopify is using auto-research loops to optimize search, themes, prompt compression, storage, and more* Why Tangent is becoming a democratizing tool for PMs and domain experts, not just ML engineers* Why AutoML finally feels real in the LLM era, and where auto-research still falls short today* Why Tangle, Tangent, and SimGym become much more powerful when combined into one system* What SimGym is, why simulated customers only work if you have real historical behavior, and why Shopify's data gives it a moat* How SimGym evolved from comparing A/B variants to telling merchants what to change on a single live storefront to raise conversions* Why customer simulation is so expensive, from multimodal models to browser farms to serving and distillation costs* How Shopify models merchant and buyer trajectories, runs counterfactuals, and thinks about interventions like discounts, campaigns, and notifications* Why category-level behavior is so different across commerce, and why ideas like Chinese Restaurant Processes are showing up again in practice* Shopify's new UCP and catalog work, including runtime product search, bulk lookups, and identity linking* Why Shopify is using Liquid AI, and why Mikhail sees it as the first genuinely competitive non-transformer architecture he has used in practice* Where Liquid already works inside Shopify today, from low-latency query understanding to large-scale catalog and Sidekick Pulse workloads* Whether Liquid could become frontier-scale with enough compute, and why Shopify remains pragmatic and merit-based about model choice* Who Shopify is hiring right now across ML, data science, and distributed databases* The Sydney story at Bing, why its personality was not an accident, and what Mikhail learned from deliberately shaping AI character early onMikhail Parakhin* LinkedIn: https://www.linkedin.com/in/mikhail-parakhin/* X: https://x.com/MParakhinTimestamps00:00:00 Introduction: Mikhail Parakhin, Microsoft, and Shopify00:01:16 Why Shopify Is Talking More About AI00:02:29 Internal AI Adoption at Shopify and the December Inflection00:06:54 Token Budgets, Jensen Huang, and Why Usage Metrics Can Mislead00:10:55 Why Shopify Built Its Own AI PR Review System00:12:38 AI Coding, More Bugs, and the Real Deployment Bottleneck00:14:11 Why Git, PRs, and CI/CD May Need to Change for Agents00:18:24 Tangle: Shopify's Reproducible ML and Data Workflow Engine00:21:19 Why Tangle Is Different from Airflow00:26:14 Tangent: Auto Research for Optimization and Experimentation00:30:07 How Tangent Democratizes Experimentation Beyond ML Engineers00:33:06 The Limits of Auto Research00:36:36 Why Tangle, Tangent, and SimGym Compound Together00:37:20 SimGym: Simulating Customers with Shopify's Historical Data00:42:47 The Infra Behind SimGym00:46:00 Why SimGym Gets Better with Real Customer History00:47:30 Counterfactuals, HSTU, and Modeling Merchant Trajectories00:51:55 CRPs, Clustering, and Category-Level Customer Behavior00:53:30 UCP, Shopify Catalog, and Identity Linking00:55:07 Liquid AI: Why Shopify Uses Non-Transformer Models00:59:13 Real Shopify Use Cases for Liquid01:03:00 Can Liquid Scale into a Frontier Model?01:09:49 Hiring at Shopify: ML, Data Science, and Databases01:10:43 Sydney at Bing: Personality Shaping and AI Character01:13:32 Closing ThoughtsTranscript[00:00:00] swyx: Okay. We're here in the studio, a remote studio, with Mikhail Parakhin, CTO of Shopify. Welcome.[00:00:08] Mikhail Parakhin: Thank you. Welcome.[00:00:10] swyx: I don't even know if I should introduce you as CTO of Shopify. I feel like you have many identities. Uh, you led sort of the, the Bing ML team, I guess, uh, uh, or ads team. I, I don't know, I don't know, uh, you know, it's, uh, people va-variously refer you as like CEO or, or, uh, I don't know what that, that, that said previous role at Microsoft was.[00:00:29] Mikhail Parakhin: Uh, that was... Yeah, my previous role w- at Microsoft was the-- I actually was the CEO of one of Microsoft's business units, which included, as I, you know, as we discussed, all the things that people like to laugh about, uh, including Windows and Edge and Bing and ads and everything.[00:00:47] swyx: Yeah, yeah. What a, what a, what a wild time.You've obviously, uh, done a lot since you landed at Shopify. Uh, one of the reasons I reached out was because you started promoting more sort of internal tooling, uh, primarily Tangle, but also a lot of people have seen and adopted Tobi's QMD, uh, and obviously, I think, uh, Shopify has always been sort of leading in terms of, uh, engineering.I think more-- it's just more recent that you guys have been more vocal about your sort of AI adoption. Is that, is that true?[00:01:16] Mikhail Parakhin: Well, I think AI tools in general are fairly recent development, uh, and we've-- Shopify, you know, at this stage of its development, we're developing AI in-in-house and other, uh, building tools that use AI and, you know, interfacing with the wider AI community, uh, you know, are on the sort of the, uh, runaway trajectory.So it just did by sort of natural byproduct. We, we talk about it more also. We just, uh, just even yesterday, Andrej Karpathy was famous in tweeting about, oh, are there some, uh, ways, uh, that, that you can organize your agents to store the data and then, uh, look up the data so that you don't have to research or, or lose context every- Yestime. And a little bit tongue in cheek, I tweeted that, “Hey, we've, we've done it much earlier, and we even have different approaches, Tobi and I.” Tobi, of course, is a big fan of QMD, and I'm more of a SQL, SQLite fan. But, uh, yeah, very similar things that we've already done here. The point is, yeah, we're very dynamic, you know, explosively growing company, and we have to be at the forefront of AI adoption, obviously.[00:02:29] swyx: Yeah. Yeah. Um, you, your team kindly prepared some slides actually that we were gonna bring up on to, uh, the screen. I think I can, I can screen share, and then we can kind of go through some of the shocking stats that maybe, maybe put some numbers to what exactly is going on. So here we have, uh- An internal AI tool adoption chart.What are we looking at here? What ?[00:02:54] Mikhail Parakhin: Yeah, this is very interesting statistics. Uh, this is number of daily active workers, you know, think of, uh, DAO, basically the active users of-[00:03:05] swyx: Yeah ...[00:03:05] Mikhail Parakhin: AI tool as a percentage of all the people in the company, right? And then- Yeah ... different AI tools. And, uh, you could see two things here is that one is the green is total.Uh, green is just total. So you could see that it approaches really % by now. It's hard not to do your job now without interacting deeply, at least with one tool. You could see another interesting thing is just as many people commented in December was the phase transition when suddenly models gotten good enough that, that everything took off and started growing.Uh, it, it was many people noticed that the thing is that small improvements accumulated into this big change in Sep- December roughly timeframe.[00:03:52] swyx: Yeah.[00:03:52] Mikhail Parakhin: The other thing I would claim you could see is that, uh, CLI-based tools and tools that don't require you to look at the code becoming more popular, and you could see, yeah, various versions of, uh, Cloud Code and Codex and Pi and internal development tools taking off.Uh, exactly, yeah, uh, and blue is our River, just internal agent for coding, where tools, uh, that require IDEs such as, uh, GitHub, Copilot or Cursor, they're not exactly shrinking, but they're not growing as fast. Like, uh, red, red line is, is the IDE kind of tools. So you could see that they're, they're not experiencing as, as fast of a growth.[00:04:37] swyx: As I understand it, basically, every employee has their choice, right? Of choose whatever tool you use, and then you're just kind of doing a, a daily sur-survey or something.[00:04:47] Mikhail Parakhin: Exactly. And, uh, we- Yeah ... the, the push is to get your job done, you can use any tool, and we effectively fund unlimited tokens for everybody.Uh, we, we do, we do try to control the models that, uh, people use, but from the bottom, not from top. Like we basically say, “Hey, please don't use anything less than Opus four point six.”[00:05:09] swyx: Oh .[00:05:10] Mikhail Parakhin: Some people, some people end up using GPT five point four extra high. Some people use Opus four point six. Um, uh, you know, uh, there are some, uh, there are plus and minuses in going for full one million context window versus not.But, uh, we try to discourage people from using anything less than that.[00:05:28] swyx: Yeah, yeah. Got it, got it. Uh, I mean, uh, that's, you know... The, the next chart here, it really kind of shows the expansion and the sort of December twenty twenty-five inflection, right? That, uh, people are using a lot of tokens. I think it's also really interesting that no one was kind of abusing it in twenty twenty-five.Like it was- Had comparatively, uh, to this year, there was almost no growth. I mean, it's still like, you know, probably, probably gave fifty percent.[00:05:56] Mikhail Parakhin: Yeah. This is just a different scale. It's still exponential- Yeah, yeah ...growth at just a different- ...rate of expansion. Uh, there was inflection point, and Sean, I would claim the, the super interesting part here is that you could see that the distribution becoming more and more skewed.Yes. The top percentiles grow faster. So that means- Yeah ...the people in the top ten percentile, they, their consumption grows faster than seventy-five and so forth. So, uh, the distribution skews more and more towards the highest users, which is... I don't know what it tells me. It's like it feels not ideal, to be honest.Or maybe it's okay. We'll see.[00:06:36] swyx: Why does it feel not ideal? Is, is it because of, um, quantity over quality, or what's the concern?[00:06:42] Mikhail Parakhin: Because take it to the limit. That means, you know, if, if this rate of separation continued- Ah, yes ...a year, there will be one person consuming all the tokens. So it's just, it's kinda strange.[00:06:54] swyx: Yeah, I mean, um, uh, I, I think internal like teaching and all that, uh, will, will help sort of distribute things more widely. But in, in the early days, of course, the people who are sort of more AI-pilled will obviously find more ways to use it than the people who are less AI-pilled. Maybe let's, let's call it that.I'll just, I'll just kinda quickly, uh, pause from the, the... You know, we will go back to the rest of the slides, but I just wanna, um, review, you know, there are a lot of CTOs of, of large companies like yourself where they're all considering some kind of token budget, right? Like I think it's something, something that Jensen Huang has been talking about, where like if your 200K engineer is not using 100K of tokens every year, like they're, they're underutilizing coding agents.Of course, Jensen Huang would say that, but like it seems a very quantity over quality approach and like some, some people are basically saying like, well, is this comparable to judging engineer quality by lines of code, right? Which we also know is like kind of flawed, but better than nothing. So I, I don't know if you have like a sort of management take here on, on how to view this kind of, uh, metrics.[00:08:02] Mikhail Parakhin: Well, I mean, you're, you're baiting me. I, I like... This is my favorite topic. Uh, if you let me, I'll probably talk for two hours on just this. I have a lot of things to say. Like I do think Jensen gotten a lot of bad press saying, “Oh, of course you're, you know, this, uh, the- ...the cake seller says you don't need enough cakes.”You know? Like, of course. Uh, but, uh, I actually, uh, think that's undeserved. I think he, he's actually right. Uh, I do think- He,[00:08:33] swyx: he's directionally correct.[00:08:35] Mikhail Parakhin: Yeah. Yeah. He's directionally correct for sure. Uh-[00:08:37] swyx: Who knows what the right number is? Yeah.[00:08:39] Mikhail Parakhin: The thing that I do Uh, want to say, and this is something that we learned through trial and error and very important is like two things.One is that it's not about just consuming tokens. Uh, you can consume tokens and, and in fact, the anti-pattern is running multiple agents, too many agents in parallel that don't communicate with each other. That's almost useless, uh, compared to just fewer agents and burns tokens very efficiently. Uh, setting up the right critique loop, especially with the high quality models, where one agent does something, the other one, ideally with a different model, critiques it, uh, suggests ways to improve it, the agent redoes it with this critique and, and so it takes much longer.So people don't like it because latency goes up. You know, they, they have to wait until this debate is happening. But, uh, the quality of the code is much higher. And another thing, just since you mentioned like, look, uh, uh, yeah, the overall budget is just like, uh, lines of codes. Lines of codes are exploding for everybody right now, or partially because AI is really mover balls, but partially just because AI can write a lot more code, you know, doesn't get tired.And so you have to have to have a very strong narrow waist during PR review. Otherwise, just the number of bugs will go through the roof. It's, uh, it's this unexpected consequence of the just volume trumping everything. I would claim by now good model writes code on average with fewer bugs than, than the average human.But since they write so much more of it, like more of it will make it into production. So you have to- You still[00:10:26] swyx: have[00:10:26] Mikhail Parakhin: more bugs. Yeah. Have to have a very rigorous PR reviews, also automated of course. But, uh, yeah, that to spend a lot budget there. Like this, this for me, for me, actually, the important metric is the ratio of budget spent during code generation versus, uh, spent, uh, expensive tokens like GPT, uh, five point four Pro or, uh, uh, Deep Think from Gemini, you know, checking on PR reviews.[00:10:55] swyx: Yeah, totally. Uh, I noticed in your chart you didn't have any review tools. Do you just use like, like let's say a Claude code to review tools? Or do you have another set of review tools like the Greptiles, the Code Rabbits, uh, Devin Reviews has a review tool. I don't know if you've had those specialist review tools.[00:11:13] Mikhail Parakhin: You are a little bit jumping on my store tool right now because the graphs I was only showing public tools. Uh, uh, the-- I haven't found a good PR review tool that, that does what I think should be done. And, uh, partially my, my thinking is because it's so... It just goes against both what people feel like emotionally they prefer and, uh, some of the, uh, you know, frankly Even business models that, that the companies run.At peer review tool, uh, time, you want to run the largest models. That means, I don't know, Codex or, or, uh, Cloud Code is not gonna cut it. You need to have pro-level models if you really want to, uh, stand the tide of bots from going into production. And you need us to spend a lot of time, the models taking turns, but you don't want, like, a big swarm of, uh, of, uh, agents.So in fact, you end up in a different dual-dualistic world where you generate not that many tokens. You, in fact, generate few tokens, but it takes f-a long time because these are expensive models taking turns rather than many, many agents trying to do many things in parallel. So that's, that's why I feel like I haven't found good tools, so we are using our own for peer review for now.[00:12:33] swyx: Yeah. Yeah. I mean, uh, I think a lot of companies are building their own, uh, especially to their needs, right?[00:12:38] Mikhail Parakhin: Mm-hmm.[00:12:38] swyx: Um, I, uh, you also have a chart here going back to the slides on, uh, PR merge growth, where we're now at thirty percent, uh, month on month rather than ten percent. Uh, and also the, the estimated complexity is going up.You know, this is productivity, right? ‘Cause y- presumably there's more stuff going into the code base and more, more features getting worked on. I'm curious about the backlog, right? Like the, the, the-- I actually don't mind a pro-level model taking an hour or two hours to review my PR, because I've dealt with humans who take a week to review my PR, right?And I keep pinging them on Slack, “Hey, hey, review my PR.” So, you know, I think there's some trade-off here where, like, it still doesn't make sense.[00:13:18] Mikhail Parakhin: Exactly. That, that's exactly m-my point. Uh, that on one hand, you can tolerate longer latencies at, uh, PR. On the other hand, like right now, the real problem is not in spending time waiting for PR.It's real problem is since there's so much more code than- Yeah ... uh, probability of at least some tests failing going up, and then you, like, keep de-failing, then you have to find the offending PR, evict it, retest it without that PR, and so deployment cycle becomes much longer. Uh, so it actually, in terms of the overall time to deploy, it's total time savings if you spend more time on a longer model, like thinking for an hour, because then, then you, you don't have to spend all that time during testing and rolling, you know, rolling back the deployment.[00:14:03] swyx: Yeah, totally. That's still worth it. You know, you don't look at the individual, look at the aggregate, and look at the, the, the change in the aggregate system.[00:14:11] Mikhail Parakhin: Exactly.[00:14:11] swyx: I'm kind of curious if, like, there's this PR mentality and, like, c-- the, the, the CICD paradigm will be changed eventually. Some people are like, obviously a lot of people want new GitHub, but I even wonder if, like, Git is the problem, right?Like, is that the bottleneck? Is the concept of a PR a bottleneck? Do you guys use stack diffs? I don't know if, uh, that's a, like, a merge queue stack diff type of thing.[00:14:34] Mikhail Parakhin: We, we use, we use Stacks, we u- we use Graphite. We worked with, uh, Graphite a lot. Uh, so we use Stack, uh, PRs. I think, uh, like that's clearly the overall CICD in general, and the interaction with the code repository right now is the, clearly the sort of the, the main issue and the bottleneck for us, uh, and highest top of mind.I would say we probably need a different metaphor or different whole design of how to process it in new agentic world. I haven't seen anything dramatically better yet. I, I think everybody right now is just trying to keep their head above the water ‘cause, ‘cause there, there's so many PRs and then everybody's CICD pipelines start creaking, the, the times are increasing, the number of bugs slipping by increasing, and you have to, have to clap on down.And so we are a little bit in this situation when we need to first stabilize that story and then start thinking, hey, what, what it could be a completely different and new world, which I haven't... I know some people working on it. I haven't seen something, like anything super compelling yet, but clearly the old thing were designed for humans will need to be morphed into something new.[00:15:53] swyx: One of the thing that I, I think about is kind of like the merge conflict is basically a global mutex on the whole system, right? And in, in hu- in human organizations, we do have something like that. It's the company standup. But like, other than that, it's like it's actually fitting for us to be somewhat decentralized, somewhat plugged into one stream of information source, but somewhat lossy.Like it's okay, you know, that, that not every delivery is like atomic consistency. Like we're not dealing with a database sometimes.[00:16:27] Mikhail Parakhin: This is a very good point, uh, because since humans don't write code too fast, you know that global mutex is not too bad. Once you-[00:16:36] swyx: Yes ...[00:16:37] Mikhail Parakhin: start writing code at the speed of machine, it becomes the, you know, the bottleneck.Then what do you do? Maybe, and I can't believe I'm saying this because I, I'm long-- lifelong opponent of, uh, microservices, and I always thought that was, like, a really bad idea. And now that you're saying it, like, maybe in new guys like microservices will make a comeback, you know, because then you, you can ship things independently in tiny things and, and the managing all that complexity automatically will be much easier.I don't know. Like, we'll s-- we'll have to see.[00:17:10] swyx: Yeah. I mean, I don't know what the Microsoft or, or Shopify thing is, but I, I read this paper from Google where they have a monorepo that deploys into microservices, right? And then, uh, the other concept that I think about a lot is the Chaos Monkey concept from, from Netflix.Being able to create, like, this robust system where, um, uh, you know, you, you have the service discovery, you have the, uh, the independent, independent microservices discovery and, and, uh, you know, probably going to be a fair amount of duplication. That's how an organic system sort of scales, uh, that, that you have that...I don't know how you call it. Slack? Robustness? Depend-- uh, d-duplication. I, I, I forget the-- I, I'm-- And this-- those-- these are not exactly the terms- Hmm ... I'm looking for, but I c-can't really think of the words. Okay. I was gonna go into Tangent and Tangle. Uh, so, uh, we, we sort of discussed the overall stats that, uh, Shopify has.Uh, but, you know, I, I think some, some pretty cool stuff that you guys are working on is your ML experimentation, uh, and your, your sort of auto tr-research training pipeline. Presumably you're much closer to this one because it's, it's a sort of personal hobby of yours. How, how would you explain them in, together?I thought we have a slide that, like, uh, has the s- the system diagram.[00:18:24] Mikhail Parakhin: Yeah. Tangle first and then Tangent as a-[00:18:27] swyx: Yeah ...[00:18:28] Mikhail Parakhin: as a thing on top of Tangle. And, uh, Tangle is the third generation, I claim, of, uh, systems of, uh, running any data processing, but a bit with a skew for ML experiments, but not necessarily. Any sort of data processing tasks where you need to iterate, share, and you have scale so that you want maximum efficiency.You know how, like, normally you would work, you would-- Imagine you're a data scientist or an ML practitioner, you would get Jupiter notebooks or, or maybe you would get, uh, you know, Pyth- your Python scripts, and you would manage the data, and you produce those TSV files, and you put them in some JFS or something.Then you would notice that, oh, it has this, uh, weird missing values. You go and write another script that, uh, goes and replaces them with, uh-[00:19:20] swyx: Ah ...[00:19:21] Mikhail Parakhin: dash S. And then, then you, then you run some, some, uh, “Oh, I need to filter bots.” And so you run some light GBM model that, uh, removes the bots. And then, then you like-- And then you, you kind of like get into shape, and then you start experimenting, and you run multiple experiments, and then you're like, “Oh my God,” like, “this experiment is worse.”You undo, and you cannot get to previous result. And like, “Ah, what did I do?” Like that. Again, then, then you finally like get everything working. Then you like start throwing it over the fence to production. You, you replicate it, those things don't work, and then sometimes you like don't notice that you forgot some feature naming and the, the features don't match.But then, like imagine you, you did everything, and then six months later you're like, have to repeat it because now there's more data, or you wanted to do another pass, and you're like, “What, what did I do?” Or like, or like, “This script crashes now,” or the, “the path has changed.” And then, then you're trying to, like you spend another month just doing ar- digital archeology on your own, you know, history, right?Now multiply that by many, many teams. Now imagine you got an intern that you wanna ramp up. Now you have to show that intern, “Oh, you know, look, here's the folder, there's the scripts, you know, ask your cloud agent to do, and then, uh, to, to figure it out.” And then cloud agent does something, and then you're, “Ah, yeah, right, right, it was the wrong folder.I forgot to tell you, I actually have this other thing I forgot myself.” And, and that's, that's the, like, the daily life we all, uh, all know it, uh, if, if you're a data scientist, machine practitioner, ma- machine learning practitioner or, uh, or even like any data managing, uh, person.[00:21:00] swyx: Yeah. So I, I used to do this, uh, f- uh, on the quant finance side, uh, in, in my hedge fund.So we did this before Airflow, and then, uh, obviously Airflow came along and, uh, then more recently Dagster, uh, I would say is like, in my mind, what I would use for that shape of problem, uh, where you had to materialize assets and create a pipeline.[00:21:19] Mikhail Parakhin: And that's, that's very good segue because... So Airflow is great, but Airflow is more about you, you have something and you wanna repeatedly run it in production on schedule.It's less about you as a team developing things and being able to share, and you grabbing the standard pipeline and saying, “Hey, I wanna change this tiny little component in the huge sea of data processing, and I don't wanna-- I wanna run ten experiments on this, and I wanna do hyperparameter optimization.”All that is very hard to do with Airflow. It's very easy to do with Tango. Tango is m- more about, it's everything about group of people Running experiments, it might be agents too nowadays. Uh, running experiments cheaply, collaborating, sharing results. Uh, you don't need to understand fully. You, you grab-- you clone somebody else's experiment or somebody else's pipeline, uh, run, uh, change small piece, run it, be, like, get it to production state, and then ship in one click.So then the... You don't have to port it into any other system to, to run in production. You can just run the same experiment. It's, it's fully production ready. And, and it's, uh, it has lots of... Again, as I said, it's third generation system. The original one was, I would claim there was Ether and then, uh, at least in my career, Ether was the first, first, uh, that pioneered this type of approach.And then there was, uh, Nirvana, which, uh, uh, at Yandex, which did kind of sec-second take on this. And now this one aggregates the, the learnings from all of those and, and Airflow as well to, to get to the state where you try it, it, it feels kind of magical. Uh, ‘cause now everything is based on content, uh, hashes.So even if the version changed, but if the output didn't change, nothing is being rerun. It's very efficient. If you... Multiple people start experiment that needs the same sort of data preprocessing, it's not repeated multiple times. It's automatically done only once. If you start ten experiments that all require, you know, some, some data preparation first as the first step, and you don't have to coordinate for that.Like, you don't have to know that other people are starting it. You now, it's very easy compos-, uh, composability, any language you can u- uh, you wanna use, and it's very visual. So you can see immediately, you can edit it easily, you can assemble small things with just even mouse clicks if you want to, and, uh, share, clone.And everybody knows also it's fully kind of static in the sense that we rerun it second time, it will exactly have the same results. Like, you will never have to do digital archeology. So full versioning and everything is also there.[00:24:06] swyx: Uh, so, so people can, uh... It's open source. Go to the GitHub repo and, and, uh, check it out.Uh, and it is also a really good, uh, blog post about it. I think all these is, like, really appealing. The, the, the, the thing that I think sells me the most about it is that, um, sort of development to production transition, right? Which I think, um, a lot of people haven't really solved that, uh, strictly, right?Like, we develop really, really well in, in Python notebooks, but then, you know, that's obviously not a sort of production ready process. I think that, like, any way in which that is solved, I think is, is very appealing. Then the other thing that you mentioned, which also raised my eyebrows, was content-based caching, which you mentioned is, is, um, you know, is ve-very much, uh, um, a sort of efficiency measure about, uh, you know, just like recalculation only on, on sort of content addressing Which I think makes sense.Uh, it surprised me that the savings could be this much, but maybe I just haven't worked at your scale where there's so much duplication, uh, that people just rerun because they change a single ID upstream.[00:25:10] Mikhail Parakhin: It does, yeah. But it's not only you rerun. The, the main savings are coming from the fact that you ran it, you got your job done, and you moved on.Then- Yeah ... somebody else in some department you don't know existed runs the same task, but on a newer version.[00:25:27] swyx: Yeah.[00:25:27] Mikhail Parakhin: Like right now, you can't, in, in most of the organizations, you can't even find out about it so that you can't even measure that you're spending that time twice, right? Here- Yeah ... if everybody's on Tango, that's detected automatically and detected that the output is the same.And then for that person, all it looks like is like experiment just suddenly moved, jumped forward, right? Uh, uh- Yeah ... so that's because, because the, there's network effect of multiple people helping each other.[00:25:51] swyx: Yeah. This is one of those things where it's designed to be a platform from the beginning rather than an individual developer's tool from the beginning, right?And, and everything's gonna streams down from there. That is the sort of Tango, uh, orchestrator, and it's, it manages jobs. We've seen a few versions of this, and this is obviously, uh, uh, the sort of, uh, unique approaches that you guys have, have, uh, figured out. And then there's Tangent.[00:26:14] Mikhail Parakhin: Yeah. And Tangent is basically an automatic auto research loop that can help and kind of do your work for you.Uh- ... you know, uh, effectively, effectively, Andrej Karpathy recently popularized it with auto research. Yes. Remember he said like he was, uh, speed running this, uh... Yeah, uh, you know the story. The, here we're basically bringing the same capability into Tango so that, uh, the, uh, Tangent can analyze it. It's just an agent that can run multiple experiments, figure out what can be changed, and keep on rerunning it, keep on modifying until, uh, maximizing some goal, some loss function, whatever you need to, to achieve.And in general, I would say if you're not using auto research-like approach in whatever you do, like literally whatever you do, then you're missing out. We saw at Shopify that taking like a wildfire, anything where you can put measurements can be done dramatically better. Our-[00:27:19] swyx: Mm-hmm ...[00:27:20] Mikhail Parakhin: uh, speed of, uh, templatization HTML, uh, completely new UX tem- uh, templatization of, uh, reducing latency for liquid themes.Uh, we-- Our, uh, search, uh, recently we moved from It's hard even, uh, quote from eight hundred QPS to forty-two hundred QPS with the same quality just by pure optimizations and not a research loop that kept running and changing code in our index serve on the same number of machines, just increasing the throughput.We, we managed to improve the quality of gisting and machine learning process. Uh, you know, gisting is the prompt compression technique that[00:27:59] swyx: allows for[00:28:00] Mikhail Parakhin: lower latency and, and lower and, uh, actually higher quality slightly. So like literally whatever different walks of life, and it doesn't have to be AI related.Uh, we, we had a reduction in, uh, storage because the agents would go and find data sets that clearly are derivative, uh, and then you don't need to store things twice. You know, we, we, we found somewhat embarrassingly that it was one of the largest tables was hashing random IDs into another random ID, and we literally- Oofput only one. So it was translating, yeah, two random IDs hashed[00:28:36] swyx: into[00:28:37] Mikhail Parakhin: each. So, so[00:28:37] swyx: it has access to the code as well, so it can, it can check the, like what, what the hell is it doing?[00:28:42] Mikhail Parakhin: So there, there cou- it could be run in two levels. You, uh, you know, at the superficial level, it could just use ex-existing components and, uh, reshuffle them.Uh, you know, like you can grab- Yeah ... uh, XGBoost, and you can grab some, some Py- PyTorch module, and then can grab some, you know, grab another tools and, and combine them. At a deeper level, since Tangle is all sort of CLI based underneath you, every, every component is a wrapped really CLI, uh, call and a YAML file, it can analyze code and create new components and, and, uh, keep on iterating as well.So, so you can, you can both have quick modifications of existing t- uh, pipelines with the, with components that are already there pre-baked, or you can create new components, uh, and-[00:29:29] swyx: Yeah ...[00:29:29] Mikhail Parakhin: keep iterating on those. So auto research is, again, this is probably the, the thing I was excited the most in the last two months happening, and we see it taking like, like totally like a wildfire.Just, uh, everybody, every day, every... well, every day, every minute, I would, uh, have somebody Slack message saying, “Oh, look how much better I made it.” And, uh, it's all throughout the research.[00:29:53] swyx: Is this democratized in some way in, in the sense that like is it your ML, uh, engineers and researchers doing this, or is it your regular PMs and software engineers also have the ability to auto-- to use Tangent?[00:30:07] Mikhail Parakhin: This is an awesome question. Like, Tango in general and Tangent in particular are extremely democratizing. Like they- Yeah ... they are the main tools for- ‘Cause I don't[00:30:15] swyx: need the details.[00:30:16] Mikhail Parakhin: Yeah. Exactly. Initially used by ML and AI engineers, but then literally, as you said, PMs are like the highest user right now is one of PMs on our org, uh, Sartak and he was, he was number one by, by usage of, of this ‘cause they're just, uh, energetic and knowledgeable, and now it, it unlocks a lot of capability where you don't have to co-change code manually.[00:30:39] swyx: I mean, I mean, because it kind of cuts out the ML, ML engineer from the process because the, the, the PMs have the domain knowledge and the ability to think about, uh, from first principles about, okay, what, what results do I want? And they can-- they even have the access to the data that, that needs to go in.So it's like in some ways, like this is the magic black box that we've always wanted for, for training and, and for, uh, I guess, uh, uh, hill climbing, whatever.[00:31:04] Mikhail Parakhin: It's basically cloud code for your AI development- ... uh, situation, right? Like now, now you don't have to know exactly how algorithms work. You can just, uh, bring your domain knowledge and expertise and product knowledge and iterate within Tangent until you've gotten the results that you need.[00:31:21] swyx: In my previous roles, every time that someone has pitched AutoML, you know, I've always been like, “Uh, this is not, this is not gonna work. It's, you know, it's, it's always gonna be a flop.” Somehow it's working now. I mean, presumably the answer is now we have LLMs and it's good enough, right? It's, it's an emergent property that we can do auto research, but like, it doesn't feel that satisfying that how come we didn't do this before, right?Like we just did like parameter search and like, I don't know. That's maybe that's it.[00:31:48] Mikhail Parakhin: Yeah. Bayesian optimization and hyperparameter optimization was, was the one that, or facet of AutoML that was used very actively, which incidentally also built into, uh, Tango. But, you know, I know Patrice Simard very well, and, uh, he was such a, uh, such a proponent of AutoML, and he put, like literally spent careers trying to democratize it.Without LLMs, it just turned out to be very hard. Like it, you, you would have flexibility within certain narrow domain, but it was hard to wider scale, and now with LLMs suddenly it's like magic wand, and so suddenly everybody- ... is an AutoML expert.[00:32:28] swyx: Yeah, I, I think it's multiple things, right? Like I'm, I'm just gonna bring up the, the, the chart again, right?Like LLMs can do the monitoring very well. That is the very potentially unbounded, super unstructured. It can do the analysis very well, it can do the... Uh, and basically it is much more intelligence poured into every single step. Uh, there's maybe nothing structurally changed about AutoML, but this is just m-more intelligent and more unstructured.[00:32:53] Mikhail Parakhin: Exactly.[00:32:54] swyx: Any flaws that you've run into? Like everyone is like drinking the Kool-Aid, oh my God, time savings, uh, you know, performance improvements. Like what, what, uh, issues have you have, uh, come up?[00:33:06] Mikhail Parakhin: This is really cool. It's not a solution to all the world's problems for sure. The limitations are usually the ones I-- And this is where we get into a bit of a subjective territory.Uh, I can only share what I've, I've seen so far, and I'm sure the situation, uh, is changing, and, you know, maybe after I say it, like many people will reach out and say, “Hey, what about this?” And you don't know that, and then, then we'll be probably right. But what I've seen is auto research is very good at doing kind of obvious things that you don't have bandwidth to do or you didn't notice or maybe you're not aware of like the-- some standard practices.It is not good at doing something completely out of distribution, something that, you know, you have to think for, for multiple days, uh, and, and do something like none of this. So, so it's, uh, I, uh, set an experiment once, uh, on, on my sort of, uh, hobby thing, and I let it run for, uh, ended up, uh, several weeks run, uh, you know, it's like full production kind of scale, so it, you know, slow runs and, and it ex-- it performed in the end, uh, over four hundred experiments, and only one was successful.I'm like, “Okay, that's, that's good.” But-[00:34:18] swyx: But it saved time.[00:34:19] Mikhail Parakhin: Yeah, I saved time. Like it, it was the, that thing. Yeah, if I, if I were doing four hundred experiments myself, my betting average, as I said, would have been much higher, I'm sure. But also, first of all, it would take me like three years to do four hundred experiments.And, uh, I didn't have to do them. Like the machines were just, uh, the price of electricity did that. So, and I got one improvement, uh, that in, uh, my, my-- Honestly, when I was starting that experiment, my thinking was to go and show that, “Hey, Andre, maybe you just don't know how to optimize.” And I was super smart because in, in my pro-problem, it was optimized for many years, and it was like fully improved.Uh, and I didn't expect it, you know, auto research to find anything at all. Yet it did. So instead of making fun of Andre, I ended up, uh, a big, big supporter. Yeah, that's exactly the tweet. Yes.[00:35:10] swyx: You and Toby really, really go back and forth on-online a lot, which is really funny. Uh, think of it as, as an eval for the optimalness of the code it's running on.Uh, it's almost like it reminds me of like a Kolmogorov complexity thing, but, uh, I guess it's-- there's some optimal thing that you're trying to sort of reduce down to, I guess. Um, and so, so you, you, you know, you should congratulate yourself that you had, uh, you know, uh, ninety-nine percent, uh, optimality.[00:35:36] Mikhail Parakhin: Exactly, yeah. I think Andre really deserves a lot of credit for popularizing this approach. This is, uh, this is incredibly, I think, powerful and cool and You know, the, uh, even him, him just mentioning it led to a lot of gains in a lot of places in the industry, so we should be thankful.[00:35:56] swyx: Yeah. I think he also has a just...I don't know what it is. Like, um, you know, it, it is a simple self-contained project that people can take and apply to other things, which is, is, is one thing, but also just the name. Just like somehow no one, no one managed to call their thing auto research. It's just naming things is very important. I think that that is mostly, uh, our coverage of Tango and, and, uh, Tangents.I think obviously, you know, there's a lot of, uh, ML infra at, at Shopify that people can, uh, dive into. We're about to go into SimGym, but before I do that, any, any other sort of broader comments around this whole effort? Like where is it, where is it leading to?[00:36:36] Mikhail Parakhin: As a segue to SimGym, like all those things start composing strongly.And, uh, you could see a huge unlock when you can look at each one of the tools and, and you see, oh, they're extremely useful. Uh, Tango is useful by itself. Auto Research is useful by itself. SimGym is useful by itself. If you combine all three, you create like synergetic effect. I think that's why we wanted to even, uh, cover them today is because this is something that if you go back even, you know, five years ago, would've been unthinkable.Uh, replicating that, uh, would, would be either incredibly costly or impossible, right? With probably thousands of people are required.[00:37:20] swyx: Well, we have serverless human, uh, serverless intelligence, right? Like, uh, so yes, you do have thousands of hu-- of, of intelligences, not just, not humans. And that's, that's close enough, right?Even if they're not AGI, they're, they're close enough to do the, the task that you need them to do. And, and, you know, that's, there's plenty for, for a lot of routine work, knowledge work. Okay, let's get into SimGym. Um, this is one of those things I, I was surprised to see actually it's apparently your, uh, one of your most popular launches, and I think something that, uh, I think Sim AI, I think Yunjun Park, who did the Smallville thing, there's a very small cottage industry of people trying to do like the simulate customer thing.I think a lot of people maybe don't super trust this yet because they're like, well, obviously they would just do what you prompt them to do, right? But maybe just think, uh, tell us about the sort of inspiration or origin story.[00:38:10] Mikhail Parakhin: That's exactly actually the thing I wanted to cover, because if you don't have the historical data, all you can do is prompt a-agents in a vacuum, and they will do exactly what you prompt them to do.In fact, when I first proposed it, and this is a bit of, um, my brainchild initially, if I, I can boast, even Toby said like, “But wouldn't they, they just repeat what, what you tell them?” And, uh, but I'm like, “Yes, except Shopify has decades of history of how people made changes and what there is, uh, there, what it resulted in terms of sales.”So now what we can do is we can-- we have this... It's not, it's a noisy data. There's a small, usually websites, uh, you know, like things, things are never in isolation. It's almost never AB experiment. It's always AA experiment when there's has two meanings, but basically, you know, in different time you run two different things.But if you aggregate in general, uh, like everything together, and you apply, uh, denoising and collaborative filtering like approach, you can extract a very clear signal. And then you can optimize your agents. And that's why it took so long. It took almost a year of that optimization of just us sitting and fiddling, and, and we had this internal goals of correlation of hitting-- internal goal was to hit zero point seven correlation with, uh, add to cart events, for example.Like that, that if we run real AB test experiment, that it should, it should go and, and rep-uh, replicate, uh, same sort of success that, that humans had or lack thereof. And it, it took forever, and I don't think that's easily replicatable because, uh, like who else would have that data? You have to have this historic, you know, decades, uh, worth of data.And now, now the, like the other thing you need is in-infrastructure and the scale, right? Because, uh, w- again, what we found, uh, stat sig results, you need to run a lot of simulations, a lot of agents, and, and it's-- Those are expensive things. Like you're, you're making actions in the browser because you want a real friction.You want to, to be able to get the image like of what humans will see because you wanna, uh, detect effects like, “Hey, if I make my images larger, will I have more sales or l- uh, fewer sales?” And like usually people's intuition here, by the way, is that I increase my images, I will have more because they look nicer.You know, designers all look sparse and big images. Like usually your sales tank, right? But, but, uh, you know, from HTML, all the characters look the same only the, the size tag looks different, right? So it's very hard. So you have to take visual information, you have to run this in simulated browser environment on the big farm and, and of course, you have to have, uh, like very, very expensive model, good model with multi-model model.So all this it's-- is what's taken so long and, uh, to share my personal fail a little bit there, Sean, is like, you know, we always had this bias to-- for like large company bias. You know, we always, uh, whenever you-- we do, we're like, “Hey, we'll run an experiment,” right? We make, make a change, and we will run an experiment and then, uh, see, uh, see which one's better or like, “No, this is worse,” and most of them are worse, so you discard it and keep iterating, hill climbing.And we're like, “Oh, like smaller merchants, they cannot get stat sig results. They cannot really run experiments simply because, you know, in a week there would be not enough data for them.” So we thought from this perspective. What we didn't realize is that most people don't have A and B, they just have one thing, and they need suggestions of What A and B should be.So, uh, we first build this, hey, we run simulation on two separate teams and, and, uh, say, “Hey, which one is better?” We then morphed it into, and very recently just released it, when you have just your site, your theme, we run over it and we say, “Hey, here's what predicted values of, of, uh, uh, conversions are, and here's how we think you should modify it to increase your conversions.”And then circling back to what you started with, the proof is in the pudding. Like, if we are not correlating with reality, like, people will not be using it. And, uh, thankfully, we see literally every day more users than the previous day. So, so right now, uh, right now- It's working. Yeah. I'm-- Right now my problem is how to pay for it all because the so our major thing is how to optimize the LLMs, do distillation, how to run the headless browsers, uh, and handful browsers, uh, uh, cheaper so that we can accommodate the increase in traffic.[00:42:47] swyx: Yeah. I, I understand that you, uh, you published a lot of technical detail at GTC, so I was just gonna bring it up a little bit. I think s- was this in, in con-conjunction with some kind of GTC presentation? Or something like that, right?[00:42:59] Mikhail Parakhin: Well, we, yeah, we, we did it in several place, but yeah, we had the engineering- Yeahblog, uh, as well. Yeah.[00:43:05] swyx: Yeah. So you're running, uh, GPT OSS. Uh,[00:43:08] Mikhail Parakhin: the, this is an older version. You know, now we run multimodal model. But yeah- Yeah ... GPT OSS, we still run GPT OSS as well for[00:43:15] swyx: And then you have the VMs, and you also have browser-based. I really like this one where it you said, “It violates almost every assumption that standard LLM serving is designed for.”And then you had like, basically orders of magnitude differences between everything.[00:43:29] Mikhail Parakhin: Exactly. Which is, which, uh, which was, you know, a bit of a challenge to implement, like when, like even simple things. Uh, be- since it violates all the assumptions, for example, multi-instance GPUs, like MIGs don't work as well.But we needed, uh, to get MIG to work because, ‘cause otherwise it's way too expensive. And so we had to deal with the, yeah, with, uh, lots of infrastructure and, and, uh, work with, uh, uh, Fireworks and CentML, uh, you know, to help with optimizations and browser-based, as you mentioned. Yeah, like, takes a village.[00:44:04] swyx: Okay. So there's a lot of like, I guess, experimentation in the infrastructure so far, and you've published more or less what you have here. I guess I'm, I'm less familiar with CentML. I, I don't do, uh, that much work in this, this part of the stack. But why was it the sort of preferred instance platform?[00:44:22] Mikhail Parakhin: There are really three probably top companies. There used to be, uh, uh- Three top companies, uh, at least I was aware of that did, uh, LM optimization. You know, together Fireworks and Santa ML, not necessarily in that order. Santa ML recently got acquired by NVIDIA. Uh, what they did is if you have a model and you want to optimize it to a specific prof-- uh, profile of usage, uh, they would go and do it.And, uh, we work with, with those companies, uh, this was work particularly in with Santa ML and NVIDIA to get them the best possible results out of it. And, and sometimes you, you have to retune depending on, like sometimes you want the maximum throughput, sometimes you want minimal latency, sometimes you want like the cheapest, right?And, yeah, or some combination. And so yeah, these are people who would come and help you.[00:45:14] swyx: I see. I see. Yeah, yeah. I'm familiar with these people for the LLM, you know, autoregressive stack. But the other interesting category of these optimizers is also the diffusion people, whereas like Fel and, you know, uh, Pruna recently has come up a lot as well, which I think is like really underappreciated, uh, at least by myself, because I, I thought, oh, all the workload would be LLMs, but actually there's a lot of diffusion as well.[00:45:38] Mikhail Parakhin: Exactly.[00:45:38] swyx: There's a lot here, so I, I, I... it's, it's, uh, it's, it's, it's hard to cover. But I, I do think like people underappreciate the importance of customer simulation, basically. I think this is something that I'm candidly still getting to terms with. Uh, you know, uh, you also-- your team also like prepared this, like, really nice diagram.Uh, I, I assume this is AI generated.[00:46:00] Mikhail Parakhin: Yeah, it looks-[00:46:01] swyx: Maybe it's not.[00:46:01] Mikhail Parakhin: Yeah, it looks, uh, Gemini-ish. Yeah, but, uh, uh, honestly, I, I don't know where, where the hell they generated. It looks, look, uh, looks like it's, uh, Google. But the interesting part, John, that, that, uh, we haven't covered, but I, I wanted to mention is if your store had previous customers, rather than it's a new store, you're like new merchant just launching things, it helps tremendously in just correlation and forecast.Yeah, we take your previous, uh, customer's behavior, and we create agents that replicate those specific distribution of, of customers that you get, and then we a- we apply those to your changes, and then that, that raised raw, you know, the re-- uh, just correlation with the add to cart events or to-- with conversion or whatever it, it, it may be, uh, quite dramatically.So, uh, replicating humans in general seems like an interesting, cool challenge.[00:46:58] swyx: As a shareholder, I think this is the-- like if people are Shopify shareholders, they should really deeply understand this because this is basically the moat. The, the more you use Shopify, the more it will just automatically improve, right?Like you're, you're doing the job for them.[00:47:13] Mikhail Parakhin: Yeah, that's what we started with. Like, uh- ... uh, otherwise, if you're just a startup, I wouldn't do it if, uh, you know, if it was my startup because Without the data, it, yeah, as, as you said, it's, it's exactly the case that, uh, whatever you say in prompt, that's, that's what the agents will be doing.[00:47:30] swyx: The statistician in me wants to like really satisfy the sort of, um, statistical intuition, I guess. Um, to me it's kind of, uh, the, the word that comes to mind is, um, ergodicity. Uh, so let's say a, a customer takes this path, customer takes this path, customer takes this path, right? Um, the... In my mind, the way I explain it is like, okay, here, here's the ninety-five percentile, here's the five percentile, and here's the median, right?Um, but to me, what SimGym is potentially doing is that it can, uh, modify... It can sort of model the sort of in-between sort of journeys as well, that, that maybe are dependent on the previous states. This may be like a very RL-type conclusion where like basically the summary statistics, if you only did naive AB testing, you only have the, the statistics at, at, at a certain point, and you only judge based on the sort of overall summary statistics.But here you can actually model trajectories. Does that make sense? Or-[00:48:31] Mikhail Parakhin: That makes total sense because like, well, that, that makes even more sense that maybe even you realize bec- because-[00:48:38] swyx: Okay. Please,[00:48:38] Mikhail Parakhin: please. Yes ... we do-- Yeah. The, so internally, uh, we have this system, we talked about it briefly once at NeurIPS.We have a huge HSTU-based system that models the whole companies, uh, and their possible paths. And like- Yeah ... what you are, what you are showing, like actually at any point of time, you can either model the user's behavior or you mo- can also think about, uh, the whole merchant as a company, as the entity that acts in the world.You can model that as well. And then you can do, can do counterfactuals. In your graph, like in your blue graph, uh, if you're... Imagine in the center there, uh, somewhere in the middle, you would have an intervention. I give that person a coupon, or I don't know, I send a personal thank you card, or give a discount in some- somewhere.And then you can, uh, then you can do forward rollouts from that counterfactual. So what would have happened with that intervention or without the intervention? And you can even ch- change where that intervention, uh, in time can happen, right? Like some- where, where in this journey. So we, we do this at the Shopify scale for our merchants, and then if we notice that something that they can be fixing, like there's a strong counterfactual, like we have Shopify policy, they basically get a notification like, “Hey, we think your...something is wrong with your-” I don't know, Canadian sales. Like, uh, it looks like it's misconfigured. Here's what you need to do. Or do you think like, uh, you have to set up this campaign with these parameters? And we do that at the buyer level to literally offer discounts or cashback or, or things to buyers.So this is-- I'm getting very excited. Like this is my sort of area of, uh, interest, I guess, and, and hobby. But being able to m-model something complex as human beings or companies and model counterfactuals on it, where you can have interventions in the future and optimize when to make intervention, what kind inter-- uh, what kind of intervention to make.It's such an unlock that previously was completely impossible. Like the-- it was, it was always dreamed of, but never... Like how would you even simulate it without LLMs or HTUs? I think very, very exciting times.[00:50:59] swyx: I just wanted to, uh, to maybe illustrate this. I, I'm not the best illustrator, but I, I am a conceptual statistics guy.And y-you know, you cannot just do this. Like this is a dimensionality AB test doesn't do, right? Like, uh, because it doesn't have the, the, the change over time, uh, stochastic nature, uh, and it doesn't have the sort of contextual like... Here's all the context to this point. Um, okay, cool. Um, that's SimGym.You're, you're gonna burn a lot of tokens on this thing. But you're, you're one of the, the only scale platforms in the world that can, uh, that can do this across a huge variety of workloads, right? I'm even curious on a sort of human, uh, research level of like, well, do, does retail behave d-differently from like clothing sales?D-does that behave differently from electronic sales? I, I don't know. I don't know what else you guys... The Kardashian shoppers, do they differ from like people who buy, uh, I don't know, cars and, uh, whatever.[00:51:55] Mikhail Parakhin: Well, very different, and different sensitivities and different modes of, uh, shopping and, and different levels of what's important.Now, to-totally, you can do aggregations at, uh, at a store level. You can do aggregations at a different, uh, category level. I don't know if, uh, you know, for our statisticians among us, I couldn't believe, but we-- recently we're looking at it, and we had to bring back, uh, CRPs, you know, Chinese restaurant process.It's a, like, way of aggregating and, like, naturally grow clustering. So across... Specifically to answer questions that, uh, like you were just posing on how, how if, if buyers behave different categories. And I'm like, “I haven't seen CRP since two thousand and one.” It's[00:52:37] swyx: so What? It's so- What is... No, I haven't, I haven't seen this.No. This is not in my training. Uh,[00:52:44] Mikhail Parakhin: but, but yeah, it, uh, uh, it actually, like the, the-- there was a very popular kind of theory, popular neurips HTML circles in early two thousands, uh, kind of nice. And now, now it has practical applications, uh- Yeah ... that we were resurrecting.[00:53:03] swyx: Yeah, amazing. Uh, I, I can see, I can see how this is like a, uh, a fun job for you where you get to apply all these things.Um, yeah, yeah, so super cool. Super cool. So, okay, so, so anyone who, who knows what CRPs are and has always wanted to use them at work, uh, they should, they should definitely join Shopify. Okay, so w-we have a lot and but I, I'm, I'm being mindful of the time. I, I do wanted to, to sort of cover some other things.Um, I-I'll give you a choice, UCP or Liquid?[00:53:30] Mikhail Parakhin: Liquid. I think, I think on UCP, you know, like UCP is very important for us and, and it just we are-- UCP, we have a structured, uh, discussions, and you can read about them, and we have, uh, blog posts, and we have a big release this week, in fact, like with our catalog.Oh,[00:53:46] swyx: okay.[00:53:46] Mikhail Parakhin: Uh, yeah,[00:53:46] swyx: but- Le-I mean, we, we can, we can discuss the, the, the release briefly because we'll release this after the-- after it's already announced so whatever. There's a catalog that you guys are doing?[00:53:55] Mikhail Parakhin: Yeah. So we are, we are- Okay ... we are bringing in capabilities of a whole, uh, Shopify catalog.Basically, you now you can search for products, you can do lookups by specific ID, you can do bulk lookups when you need to bring m-multiple products. You don't need to know in ad-in advance what you're trying to show or to sell or check out. Like, you can now, you can now have this decided at, at runtime, and this big area for investment for us for both non-personalized and personalized searches, trying to provide basically a win-window into whole universe of products that are being sold everywhere in the world.And Shopify is really not exactly, but almost like a super set of any-anything being sold. Now we are bringing it into UCP and, uh, and, uh, identity linking is another big thing for us, uh, so that you, you can use, uh, like Google or whatever, whatever identity you have, uh, they're minimizing friction.[00:54:56] swyx: Yeah. So[00:54:57] Mikhail Parakhin: yeah, big release for us.But Liquid AI of course we never talk about, and the problem might be more, more aligned with what we d-discussed previously on this chat.[00:55:07] swyx: Sure. The main thing that everyone understands about Liquid is that it is inspired by Worm, and I still don't know why. I'm curious on your explanation. I think you, you, uh, you can make things very approachable.And also I think like what is the potential of like the, the level of efficiency that you get out of Liquid?[00:55:23] Mikhail Parakhin: You- we all familiar with transformer architectures. And, uh, for the longest time, there was a competing architecture, it's called the state space models. So, so Sams, uh, you know, Chris, Chris Reyes, one of the pioneers and, and lots of startups, uh, trying to make those realities.They have, uh, significant benefits being main being, uh, being much faster and, uh, lower footprint and not quadratic in length, you know, sort of, uh, linear in, in, uh, in your context length. But with state space models- They never quite made it. Like they're used-- They have, uh, certain niches when they thrive, their hybrid architectures are useful, but they never quite made it.And liquid neural networks are, you can think of them as a next step, like, uh, sort of, uh, state-space model square. It's non-transformer architecture that's more complicated than sta-state space and really difficult to code if you-- if I'm being honest. But it's, um, very efficient. It's, uh, subline-- sub, uh, quadratic in, in length of your context.Uh, it's very compact way to represent things, and that's a liquid AI company. They... Their goal is to productize it, and very often you have this need, uh, when you need to have long context and small model, and you want to have low latency. Like in general, it's basically on par with transformers, and if you do hybrids with transformers, it's, it's even better.That's why we at Shopify, when we tried multiple and we constantly try multiple models, multiple companies, we found that for small, particularly with low latency applications, when you have low latency and/or if you need longer context lengths, liquid was the best. And so we still use the whole zoo and always like obviously test and use everything, uh, every open source model and, you know, it feels l

"Zwolnij swoich developerów - AI zakoduje za nich" - brzmi kusząco, prawda? Krzysiek, jako menedżer, wziął technologię, której nie zna, wrzucił ją w Claude Code i wdrożył na produkcję. Działa! Więc po co nam developerzy? No właśnie... nie tak szybko. Bo za tym clickbaitowym tytułem kryje się dużo bardziej złożona historia - o tym, co się stało, gdy senior architekt przejrzał ten "piękny" kod wygenerowany przez AI, o sekretach wyciekających na front, o katastrofie promu Challenger i o tym, dlaczego drobne ustępstwa w jakości mogą skończyć się wielkim bum. W tym odcinku rozmawiamy o tym, jak wygląda realność kodowania z AI z perspektywy menedżera - co jest fajne, co jest złudne i gdzie czyhają pułapki, których żaden model językowy za nas nie wyłapie. Opowiadamy o tym, dlaczego background techniczny i ludzki nadzór nadal mają znaczenie, jak zmienia się rynek rekrutacji w IT (co jest na topie?) oraz dlaczego zamiast zwalniać ludzi, warto nauczyć ich surfować na fali AI. Jeśli zastanawiasz się, jak podejść do AI w swoim zespole, jak nie dać się złapać na hype jednoosobowych firm za miliard dolarów i co robić, żeby nie zostać w tyle - ten odcinek jest dla Ciebie. Zapraszamy! Z tego odcinka dowiesz się: - Czy menedżer bez doświadczenia technicznego może bezpiecznie kodować z AI? - Dlaczego drobne ustępstwa w jakości kodu generowanego przez AI mogą skończyć się katastrofą? - Kogo zatrudniać, a kogo rozwijać w erze AI? - Czy vibe coding to przyszłość, czy droga do wycieku danych i strat finansowych? - Jak nie zostać w tyle, gdy AI zmienia zasady gry w IT? Linki do materiałów, wersję video oraz transkrypt do tego odcinka znajdziesz na stronie:

Show Notes: Lauri Euren, founder of Operating.app, explains that Operating is a tool for consulting firms or professional services that are growing and need help with staffing, internal resourcing, and month-end closes. The tool handles the workflow from time sheets to invoicing, supporting delivery across multiple projects.  Operating.app Explained Lauri emphasizes that Operating is not a CRM system and can integrate with existing CRMs like Salesforce, HubSpot, and Pipedrive. Lauri explains that Operating is not for actual sales execution but for back-end processes like staffing, time sheets, and invoicing. The tool also monitors project health, margins, and financials, providing tracking and insights into project burn rates and utilization. Operating.app Demonstration Lauri demonstrates the main dashboard, which includes tabs for staffing, personal schedules, projects, and reports. Lauri describes the main entities in Operating: projects, people, and positions, and how they intersect to manage staffing and project assignments. Lauri explains the different sections and their functions.   The dashboard includes pinned views for saved and shared views, time sheets for tracking hours, and various reporting options. Lauri highlights the importance of tailoring the use of Operating to different roles within the organization, such as consultants, COOs, and staffing managers. The invoicing feature generates invoices based on the invoicing schedule and billing type of the project. When saved, the invoice saves automatically to the accounting software used.  Lauri explains that projects often come from CRMs like HubSpot or Salesforce and are enriched with metadata using AI tools like Copilot or GPT using the Operating MCP server and AI tools like Claude, ChatGPT, Gemini or Microsoft Copilot. Operating can handle different pricing structures, including per hour and fixed price models, and can model various types of work within the fixed price model.   Planned vs Forecast Revenue Revenue Recognition The tool also includes revenue recognition features, allowing finance teams to track and recognize revenue accurately. Lauri mentions that Operating integrates with accounting tools like QuickBooks or NetSuite for billing and financial management. Clients Operating Serves Lauri discusses the ideal client size for Operating, typically starting from 10 people and scaling up to large firms with hundreds of consultants. The tool replaces spreadsheets and other point solutions for time tracking and resource planning. Lauri explains how Operating handles external consultants, tagging them differently in the system and managing their permissions and utilization. The tool allows for robust customization of permissions, ensuring that each user sees and edits only what they are allowed to. Managing a Project in Operating Lauri explains the end-to-end process of managing a project in Operating, from staffing to invoicing.  The process includes adding a project, allocating team members, setting budgets, and tracking time sheets. Operating provides real-time profitability calculations and margin effects for projects. The tool allows for the creation of invoices based on time entries and integrates with accounting tools for final billing. AI Integration with Operation Lauri highlights the AI integration with tools like Claude, allowing users to query and manage projects using natural language. The MCP server in Operating acts as an AI agent, consuming data and executing queries based on user permissions. Customers can build automated tasks and health checkers using Cloud Code and AI, enhancing efficiency and accuracy. The tool's flexibility and customization options make it suitable for various roles and project management needs. Pricing Model Lauri explains the current pricing model for Operating, which is $22 per person per month for the full module and $11-$13 for individual modules. The tool is designed to be more affordable than enterprise competitors, making it  a scalable solution for growing consulting firms. Lauri provides information on how interested parties can start a free trial or book a demo on the Operating website. Timestamps: 0:02: Introduction and Overview of Operating App  02:44: Features and Functionality of Operating  06:39: Detailed Walkthrough of Operating Dashboard  08:15: Integration with CRMs and AI Tools  13:39: Client Segment and Implementation  18:26: End-to-End Project Management in Operating  26:03: AI Integration and Customization  27:57: Pricing and Availability  Links: Operating Website: https://www.operating.app/ MCP Server (Use directly in Claude, ChatGPT...): https://www.operating.app/blog-posts/ai-consulting-mcp-server Lauri's LinkedIn: https://www.linkedin.com/in/eurenl/ This episode on Umbrex:  Unleashed is produced by Umbrex, which has a mission of connecting independent management consultants with one another, creating opportunities for members to meet, build relationships, and share lessons learned. Learn more at www.umbrex.com. *AI generated timestamps and show notes.  

- La IA ahora está subvencionada, se viene la gran subida - La gran filtración (doble, de Anthropic) - La IA no se hace sola, hay que hacerla con Nvidia - Google trabaja en el "Agent Smith". - ARC-AGI 3 - Martin vuelve a pensar fuera de la caja monos estocásticos es el pódcast de inteligencia artificial presentado desde Málaga por Antonio Ortiz (@antonello) y Matías S. Zavia (@matiass). Hay un episodio nuevo cada jueves. Puedes unirte gratis a nuestro club social de Telegram y seguirnos en redes sociales: - Telegram https://t.me/monosclub - Twitter https://x.com/monospodcast - LinkedIn https://www.linkedin.com/company/monos-estoc-sticos/ - Instagram https://www.instagram.com/monosestocasticos - TikTok https://www.tiktok.com/@monosestocasticos - Bluesky https://monosestocasticos.bsky.social - Threads https://www.threads.com/@monosestocasticos - Facebook https://www.facebook.com/profile.php?id=61584654541061 Todos los episodios en YouTube: https://www.youtube.com/playlist?list=PL-6s6cUsxTnsY_V0rqQFURaHDYuXD0AXj Más enlaces al pódcast: https://cuonda.com/monos-estocasticos/links - (0) Introducción - (05:20) Subida de Precios en la IA - (20:36) Filtraciones de Cloud Code - (31:35) Nuevos Modelos de IA - (41:08) Darío y el Altruismo Efectivo - (48:04) La Venta de DeepMind - (57:16) Evolución de ArcG y Desafíos - (1h16) Martín Varsavsky y el Futuro de España

For a limited time, Latent Spacenauts can skip the waitline to join Dreamer and also compete for a $10,000 cash prize for most useful tools for Dreamer! Thanks @dps!In 2024, David Singleton left Stripe and joined forces with Hugo Barra for a buzzy stealth startup named /dev/agents. This month they emerged out as Dreamer, a consumer-first platform to discover, build, and use AI agents and agentic apps, centered on a personal “Sidekick” that helps users customize experiences via natural language. Sidekick is nothing less than an “agent that builds agents”, with all the complexity that that entails:You've seen many many website builder, app builder, and even agent builder startups by now, but our favorite detail is the sheer amount of work that has gone into the “full stack” nature of the platform, including shipping their own SDK, logging, database, prompt management, serverless functions, and so on. Most platforms restrict the tech stack you can use just to get off the ground — Dreamer does it “right” by letting you push whatever arbitrary code you want to their VMs.Paying the BuildersOf course former leaders of Stripe and Android would not stop at just building the tools, but also building the ecosystem. Dreamer is deeply aware of the 4 sided network effect it has going on and is ready to fund all of it - from hiring Builders in Residence to awarding $10,000 cash prizes to the best tool builders for the Dreamer ecosystem.It's time to Dream!Full Video Episodeon youtube.Transcript[00:00:00] Meet Dreamer Purple[00:00:00] swyx: Okay, we're here in the studio with David Singleton. Welcome.[00:00:08] David Singleton: Hey, Wix. It's great to be here.[00:00:09] swyx: It's great to have you. Uh, we have very sympa that your company color is the same as Lean Spaces color.[00:00:15] David Singleton: That's right. Dreamer Purple.[00:00:17] swyx: It used to be Devrel agents, which I thought was very cool. It's like you call back to Devrel Payments.[00:00:22] David Singleton: Yeah.[00:00:22] swyx: And you were obviously CTO Stripe. And talk to me about just the origin or thinking process behind Dreamer. Yeah. And maybe, maybe start with like, what, what is Dreamer?[00:00:31] David Singleton: Yeah.[00:00:31] What Is Dreamer[00:00:31] David Singleton: So Dreamer is a new product, uh, which everyone can come and play with today. Um, it's a place where everyone, literally, everyone can discover, build, and enjoy and use AI agents and agenda apps.[00:00:45] And we really did design it for consumers, for folks who are not necessarily. Uh, have any kind of technical background. It's really aimed at everyone. I think often of my sister, she's very smart. She's not in the slightest bit technical. She has lots of problems in her life that [00:01:00] she would like to be able to have great software and intelligent software to solve.[00:01:04] But you know, even with the rise of tools like Cloud Code and so forth, she's got no way to get started. And Dreamer is a place where she can come in, grab some intelligent apps that other people in the community have built, start using them right away, and solve real problems in her life.[00:01:19] Sidekick And Waitlist[00:01:19] David Singleton: And at the core, we have a personal agent called the Sidekick.[00:01:24] Um, you can give your sidekick a name, you can give it its own personality, and it really helps you across your entire day, your life. It helps you use all of the agents on the platform, and it also helps you build anything you want. And we've been working in this for a little while. We recently launched in beta.[00:01:41] So anyone can go to dreamer.com, join the wait list. Um, and we have many, many, many people in the community now who are building really fun, really powerful, really useful. Agents and the agentic apps for themselves.[00:01:54] swyx: I think we're gonna go right into a demo. Yeah. I just wanna make an observation that, uh, you, you, [00:02:00] you put discover first before build.[00:02:02] Mm-hmm. But actually, at least for the engineers in the audience. ‘cause we are primarily engineers and you're primarily targeting consumers, right?[00:02:08] David Singleton: Yeah.[00:02:08] swyx: For engineers. Like, there's a huge full stack of stuff, which we're gonna dive into. Let's write. It's so impressive. I'm like, holy s**t, this, this is what I've always wanted.[00:02:16] Cool. Uh, so, so I think that's really good and I've, in some ways, I think given your background given, uh, Hugo's, is it Hugo? Hugo.[00:02:24] David Singleton: Hugo. Hugo Bar. Yeah.[00:02:25] swyx: Hugo, it's not surprising that you can basically kind of build an app store Yeah. For agents.[00:02:30] David Singleton: Yeah. So Hugo was my co-founder. Yeah. Um, Hugo and I met with our other co-founder Nicholas Checkoff in the very early days of Android at Google, where we were building Google's first mobile apps.[00:02:41] Uh, we then contributed to very core pieces of Android itself. And you're right, we were really excited about building two things. One, solving a bunch of problems. That this breakthrough technology here I'm talking about mobile needed to have solved in order to make it work for real people at scale. And then secondly, building this ecosystem, um, [00:03:00] of third party developers using the Play Store, um, and able to deliver way more value on the platform than we could have delivered on our own.[00:03:08] And we think about Dreamer in exactly the same way. So I was working at Stripe, as you mentioned, and we had the opportunity to put some of the very first AI agent systems in the world into production. And from the moment we did the first of those, I was just struck with a strong sense of conviction that this is breakthrough technology that's gonna change how all of us work with computers and phones and so forth, all of the, the technology in our lives, but.[00:03:34] There's a lot of problems to be solved, for real people to be able to make this approachable. Um, and it really is kind of a direct analog for what we were solving back in the early days of mobile apps at Google and, and Android. So it's, it's been fun to bring that to life.[00:03:47] swyx: Yeah. Uh, let's look at it.[00:03:48] David Singleton: Yeah, let's take a look.[00:03:49] Dashboard And Daily Briefing[00:03:49] David Singleton: So, uh, dreamer.com, this is our homepage. This is where you can come and, uh, watch some videos about what is here and sign up for the wait list. Once[00:03:57] swyx: you, I, I just wanna say for those listening, ‘cause we have a lot, you [00:04:00] know, switch to YouTube, look at the animations. So much care.[00:04:03] David Singleton: We, we really care about, uh, this product being fun.[00:04:07] Uh, and, and interesting to use. Obviously a lot of people are using it to do real important stuff. You can do real work, uh, here, uh, but also you can build fun things too. Once you get off of our wait list, you'll come into the product. The first thing that happens is you'll have a conversation with your side cake, which is this little friendly, uh, character here.[00:04:27] And psychic will seek to get to know you and understand you. What do you care about? And will help you discover and build your first AI agents or agentic apps. After that, you're, you're gonna have a dashboard. This is my dashboard. Everyone's is different. Um, you can see I have a few things here. I have a feed.[00:04:42] So a lot of our agents do things in the background when you're not looking and the feed is how they let you know what they've been up to. I have, uh, some widgets, uh, from apps that I have built. Uh, this one is called Calendar Hero. Uh, this is something that I installed from the gallery. Uh, so built by someone in our community.[00:04:59] It's a [00:05:00] really powerful calendar app because for each of my meetings, if it's with someone I don't already know, well it'll actually go off and research it, um, and give me both a history of my interactions with those people and also a bunch of, you know, public useful information to, to get started. One of the things I love about this particular app is that every day it generates a podcast, um, a daily briefing.[00:05:24] And one of the things that we've done with the platform is we've made it possible for all the things that agents do to show up in places that you care about. So if you look over here, this is the screen in my phone, and if I go ahead and open my Apple Podcasts, you can see right here. Your Daily briefing podcast is ready.[00:05:39] This was produced by an agent running in my Dreamer account, and it was very easy by scanning a QR code to connect it to my Apple podcast. That's what I listened to in the car now every morning. Yeah. On my way to work.[00:05:50] swyx: It, it[00:05:50] David Singleton: preps me for, for my day.[00:05:52] swyx: So one additional bit of context. I asked you immediately after seeing this was like, what, what about, I wanna talk back to my agent and you said you actually started with voice and then you went to [00:06:00] podcasts.[00:06:00] ‘cause it's nice to have it pre downloaded[00:06:02] David Singleton: that, right? That's right. Um, yeah, we, you, you can talk to your sidekick. So, you know, on mobile we have, uh, a dreamer app and you can talk to the sidekick right here. Um, but we've actually found that making things, uh, show up in the other apps that you already use in your life is incredibly powerful.[00:06:19] So let's take a look at what's kind of under the hood here.[00:06:21] Gallery Tools And Payouts[00:06:21] David Singleton: So I already mentioned that we have a gallery, so this is where you'll find a lot of agents from our community. Uh, there's. Many at this point, hundreds. And they are solving all kinds of, uh, use cases. I'd say the the top use cases are on personal productivity, but also a lot of information management that can range from personal information like docs and so forth, managing your emails.[00:06:42] It also ranges out to public information that you might be interested in, but you need something to help manage the, the kind of fire hose of stuff that's coming at you. For instance, I have, um, an agent which looks at all the AI news, um, all the time. There's a lot of it and it finds the stuff that I would actually be [00:07:00] interested in, um, and I find it incredibly useful.[00:07:03] So these are agents that you can install that other people have built. Anything that you install on Dreamer, you can actually just say, I wanna start making some changes, and we'll look at that in a second. But in natural language, with the sidekicks help, you can change any of these experiences to work just the way you want them.[00:07:18] But the base layer of the system are tools. So you know, as well as anyone swyx, that any AI system is only as good as the quality of data that it can pull in and the quality of action it can take. So before we launched our beta, we worked very hard to make sure that we seeded our tools with a bunch of very high quality and powerful integrations.[00:07:39] So, you know, for instance, this is real Google search, this is actual Gmail. Um, and you can do very useful things with those. But also this is a platform for everyone. And as we got started talking to people in our alpha community, a whole bunch of sports use cases popped out and we realized if you want to build something cool for sports with ai, you need really high quality live data.[00:07:58] So look at these [00:08:00] Formula one M-L-B-N-F-L, uh, these are tools, uh, that we've built. We've done a, these are not data scraped off the web. This is a, a direct data feed integration. And because it's live and ‘cause it's high quality, you can build really powerful stuff. But tools is not something that we are just going to kind of control ourselves.[00:08:19] The platform is open for tool Builders to contribute tools that anyone on Dreamer can use. So, um, this is actually the place in the platform where I think software engineers, um, well number one, would love for you to come and play with it. Uh, but software engineers are really gonna build, um, a lot of powerful stuff into the system.[00:08:38] And we are actually sharing something for the first time on this podcast, which there is, uh, tool builders on Dreamer get paid. So if you publish a tool to the platform and a lot of agents use it, you'll actually get paid, uh, in proportion to their usage. And we'd love for folks to come and give this a try.[00:08:54] We've got good docs that help you get started and you can build things that, you know, scratch your own itch. For instance, someone built this [00:09:00] Ski Bum tool, which provides live snow conditions for a bunch of, uh, ski resorts. I'd love to show you how I've used that in a second. And also we have some tools, partners where the tools themselves are paper use.[00:09:12] So for instance, parallel web systems is a premium tool. Uh, you can do really cool stuff with it. Um, it's a a, an agentic web research tool. And that one, because it's expensive to operate, is paid on a, on a per usage basis. But if you're coming in to build agents on the platform, even the premium tools, you get a free trial.[00:09:29] So you get a chance to actually try them out, make sure that the use case is good for you before you decide to, to to sign up. So that's tools. So we have the gallery, we have tools, and then the sidekick helps us put all of this together to build agents. We do that in the agents studio. You can also do this on your phone, but if I open up Agent Studio here on Desktop psychic's, just gonna start a conversation about what you want to build together.[00:09:51] I'd love to show you one that I made recently.[00:09:53] swyx: Let's do[00:09:53] David Singleton: it.[00:09:53] Building A Conference App[00:09:53] David Singleton: Um, let's look at something that hopefully is kind of near and dear to your heart. So one of the things I love about Dreamer and this kind of moment in technology is that if you think about it. There are all these things in your life where, have you ever gone to a conference?[00:10:09] I know you have. Right? And, uh, big conferences have apps. Um, and these apps are usually built by agencies and they're, they're usually actually quite expensive to build. I've been involved in running some of these myself. And how many conferences have you been to where the app was good? Zero. Honestly.[00:10:23] swyx: Exactly. Zero,[00:10:24] David Singleton: maybe one. I, I've, I've been to one conference. That was pretty good. Wait, wait session sessions. Um, but, but the point is, they're rarely great pieces of software. Right. And they're also expensive to build, but they're, they're interesting ‘cause they're episodic, they last for this one thing. Um, and then they're, they're not relevant anymore.[00:10:43] Um,[00:10:43] swyx: and so it's the worst feeling to invest in them because, you know, it's like, it's got a limited. Date?[00:10:48] David Singleton: Absolutely. So I decided to build, uh, a conference app for your AI engineer conference. Amazing. Uh, on Dreamer. One of the things that Swix has done, uh, which I [00:11:00] thought was very forward-looking, is actually put a whole bunch of data about the conference on the webpage in an LLM readable way.[00:11:06] There's an LLMs txt file, there's a feed of all of the sessions in js, ON. So I used the data from your conference last year and built this intelligent app, uh, just by talking to our sidekick, uh, in Dreamer. So just to give you a quick tour, this is my Dream Conference app. What I always wanna do for conferences is I wanna be able to search for speakers.[00:11:28] I'm usually there because, uh, there, uh, is a speaker I care about. So, you know, SWIX, you're the speaker I care about. I can actually see here who you're on stage with. So here's, here's Greg Brockman. You've read even ai, uh, and this is his session. And look Greg and Swix for the speaker. So let's add that to my schedule.[00:11:45] Great. And then maybe there's a couple others I might see here. Like on day two, I remember there were some keynotes. So, uh, building the open agenda web, that sounds fun. So I add that to my schedule.[00:11:55] swyx: She's now CEO of Xbox.[00:11:56] David Singleton: Awesome.[00:11:57] swyx: Which is interesting. So cool. So,[00:11:59] David Singleton: so I've [00:12:00] gone through and picked out a couple of sessions that I cared about.[00:12:03] That's as far as I usually get with any conference app. But of course you've got the whole of the rest of the conference to figure out what to do. So here is where the native intelligence of, of these things you build on Dreamer can come in. So I'm gonna click guide me. So Dreamers sidekick actually parsed out the whole schedule and figured out what some of the themes are and I can choose what I'm interested in here.[00:12:23] I'm definitely interested in agents. Uh, I'm definitely interested in code generation and also reasoning in rl. So now I'm gonna say build my schedule. So what this is doing is. It's going across every time slot for the conference. And it's choosing among the things I could go to, which one it thinks is best for me based on my interests.[00:12:41] It also uses its own memory of me that's part of Dreamer, uh, to understand what I might like best. And you know, there's an LLM prompt running for each one of these time slots. So this is, it's not super fast, but it'll be done in about 30 or 40 seconds. And I'm gonna have a special custom schedule for the conference.[00:12:57] This, like I said, is my [00:13:00] dream conference app is exactly what I've always wanted and I was able to build this yesterday morning. Um, I did it between some meetings. I think I spent a total of 25 minutes of wall clock time on it. I did it over the course of a couple of hours. And, uh, here is my schedule for the conference.[00:13:15] I can see it in a calendar view. This is what I should do on Tuesday, this is what I should do on Wednesday. Oof, no conflicts, but, you know, I may not go to every single thing. And there you have it built in, you know, dreamer. So let's take a look at what the building experience actually looks like. So this is the, the actual account that I made it on.[00:13:32] Oh, of course I should say anything you build on Dreamer also works on your phone. So, uh, here is my AI engineer conference app right here on my phone. Got all the same functionality, and of course this is the best place to jump into my schedule.[00:13:46] swyx: Yeah.[00:13:46] David Singleton: Um,[00:13:46] swyx: so you could generate a podcast about it just completely multimodal, absolute thing, right?[00:13:51] To me, I mean, this is why I outsource, I mean, well, I, I posted the L-M-T-X-T, the JSON because you cannot run an engineer conference in 2025 [00:14:00] and not let engineers. Do whatever they want.[00:14:02] David Singleton: Yeah.[00:14:03] swyx: And since all conference apps suck, I'm just gonna put up a ba minimum viable app and just let people do whatever they want.[00:14:09] David Singleton: Totally. And the cool thing about this on Bremer is I published this to the gallery and you can use it so you've got one that's built to my taste of conference apps. I think it's pretty cool. But you might want something different. Yeah. In which case you just start telling the sidekick how to change it.[00:14:23] So let's just very quickly look[00:14:24] swyx: at our, what sports grid is also, you can fork it, right? That I can publish. That's right. I can publish your one and go, this is the base starter. It's, it's got good defaults, but go customize, whatever.[00:14:32] David Singleton: That's right. That's right.[00:14:33] swyx: Yeah.[00:14:33] Agent Studio Under The Hood[00:14:33] David Singleton: So let's take a look at how I actually built this.[00:14:34] This is real. So I'm gonna say make changes. This experience we're looking at now is our, uh, agent development studio. Um, like I said, you can do this on your phone as well. And in fact, this one I started out on desktop. Let's look at my actual prompts. I said, let's make an agent called AI Engineer Schedule Planner should be a custom schedule planner for the AI engineer conference.[00:14:53] I'm not gonna read this all up. You get, you get the point and it told it where to get the data from. So that was the first prompt. And actually after I gave it that [00:15:00] prompt, I actually had a simple version of this app working, um, after the sidekick took one turn. So the Sidekick is a, like a professional software engineer, and we've worked very hard to make this work and build functional apps for folks that might not have any engineering experience whatsoever.[00:15:14] So, you know, done here we have build logs that are technical, but you can hide those away. And sidekick, as it is building, will actually translate everything that is coming out of, uh, of the, the harness into English that you can actually read. And by the way, this English is in the personality of your sidekick, which is fun.[00:15:32] Um. And the way that we build agents and agent apps, it's a little different to what you might have seen in some other platforms for a couple of reasons. One, just the build process. The very first thing that Sidekick does, it understands all the agents you've got set up. It understands all the tools and it will come up with a plan for how to realize your goal, how to make sure it actually has the data and the capabilities to complete it.[00:15:54] It will occasionally refuse. If it can't do what you're asking, it will tell you I can't do that. It needs another tool. And that's a good [00:16:00] jumping off point for any of the tool builders out there to build a new tool. So it'll fi first figure out how, then it will build it, and then it will actually test it.[00:16:07] So it will actually make sure that the thing that it has generated is realizing your goal. And you probably know as well as anybody that anytime you can get any. Modern state-of-the-art coding model into a loop where it can make changes and perceive its own output and then fix bugs. Magic happens. So these builds, the first build will often take 10 to 15 minutes on Dreamer, which is a little bit longer than you might've seen on some other platforms.[00:16:31] But the first thing that it creates will work most of the time. And then of course, as you start making smaller changes, you can like ask it to tweak the UI in any way that you like. Those are much faster. And just to give you a sense, uh, for this one, here's something I asked. Put a logo, I gave it a logo file in static files.[00:16:48] Use that as the title. So for folks that actually really want to dig, uh, into a bit more detail, we've provided a powerful IDE here. So I can actually see here's the code that was generated and some pieces of the [00:17:00] code are more accessible than others, like the prompts. So this is the prompt that's used by a powerful LLM in order to do that schedule picking.[00:17:08] And I can actually read it here directly. I can edit it without having to ask the sidekick if I want to do that.[00:17:12] swyx: So this is very nice.[00:17:13] David Singleton: This is for the more, the more, uh, sophisticated users.[00:17:16] swyx: Yeah. This is other people's entire startup is prop management.[00:17:21] David Singleton: This is true. The other thing that is different about Dreamer is once you've built something here, it's ready to go.[00:17:28] We host it. So you don't have to worry about getting a database from a database provider signing up, getting API keys. You don't have to worry about your LLM provider tokens. All of that is hosted on the platform. And you can use it yourself. You can share it to the gallery for other people to, to riff on it.[00:17:46] You can also share it with your friends and coworkers to use your instance of the agent or agentic app. And we're seeing that happen a lot in our community. We've seen a whole bunch of folks who built little applications for their personal life [00:18:00] and shared them with their significant other. We've seen people who are building little productivity apps for their team at work and sharing it, uh, among them.[00:18:07] And we actually do this a lot inside of the company. So at this point we, we pretty much run the company on Dreamer agents for all kinds of important things. Uh, maybe a good example of that is, um, our wait list. People are signing up every time someone signs up for our wait list. A dreamer agent will actually research, uh, that person.[00:18:25] And we're looking for folks who are builders, not super technical to build agents and come in, uh, and give us a lot of feedback and we're prioritized bringing those people off of the wait list First,[00:18:35] swyx: just a quick question on that one is there's, it may not come up again. Do you find enrichment APIs to be useful like the ZoomInfo?[00:18:42] Uh, clear bit[00:18:43] David Singleton: enrichment is a very, uh, common use case. Um, on dreamer. Any application on Dreamer can kick off a sub-agent to do a particular task. Um, so this actually is a powerful agentic harness that runs inside of its own [00:19:00] vm. Uh, we call them sidekick tasks ‘cause they actually run in the context of the sidekick.[00:19:04] I'll talk more about Sidekick in a second and. Enrichment is a very common use case. And the cool thing about a sidekick task is that it has access to all the tools on the platform, but also public data as well. And so very frequently enrichment on our platform happens using public data that it can be found in the web.[00:19:24] There are some tools for getting people data, uh, from, uh, from various bespoke systems. And so that works pretty well. But actually, you'd be surprised. I mean, we would love if someone out there would like to build a ZoomInfo tool, we don't have one today. We'd love to see that on the platform, and I'm sure it'll be very powerful.[00:19:39] But we're also seeing that this powerful agent harness can pull a lot of data in on that note of tools that make experiences better, we're constantly adding more tools because people in the community are building them and publishing them. We review the tools carefully and then they go live for everybody.[00:19:54] Yesterday we added granola. And that was pretty cool. So I was talking to actually, uh, Sarah on my team was [00:20:00] talking to, uh, someone building on the platform this morning and they actually, they have an agentic app that they built, which is a kind of magic to-do list. So they put stuff on their to-do list and for each thing it kicks off one of these, uh, sidekick tasks to figure out how to move the ball forward thing.[00:20:14] Sometimes it'll complete it[00:20:15] swyx: entirely. Yeah.[00:20:16] David Singleton: Often by calling another agent on the platform and sometimes it just kind of researches it and helps ‘em take the first step.[00:20:21] swyx: Yeah. Do you know, this is Sam Altman's number one, ask for an AI app. It's the self-completing to-do list.[00:20:26] David Singleton: Yeah. The self-completing to-do list is something that a lot of people have built on Dreamer and are getting a lot of use out of.[00:20:32] Yeah. And, and finding it actually genuinely I shouldn't, I should, I should try that. Mm-hmm. Please do. And you'll even find some in the gallery that you can remix. So he was saying this morning that he's, he built this self completing to-do list, uh, on Dreamer already. But he connected the granola tool yesterday and now something really magical happens, which is when he says in meetings that he's gonna do a thing, it magically shows up on his to-do list and then it can magically get completed.[00:20:56] And then, as I mentioned, all the agents, all the [00:21:00] apps on Dreamer can actually work together. So our coding agent, as it builds them, does something very special where it exposes the internals of each of the experiences to the system. And then Sidekick can manipulate those to get stuff done. So he has built another agent, which he uses for recruiting.[00:21:18] It kind of keeps track of candidates and also it's got a kinda mini CRM function, so he's able to introduce candidates to each other. He told us this morning that something he'd committed to do in a meeting that was recorded on granola yesterday showed up in his magic to-do list and his magic to-do list.[00:21:34] It was like introduce a person for recruiting, used his recruiting agent to get it done.[00:21:39] swyx: Ah,[00:21:39] David Singleton: um, and this is, this is the dream. This is why we started the company. It really is the case that you can build and use these very powerful, bespoke experiences that can automate your life by working together. And I'd love to talk a little bit about how they work together.[00:21:55] Ecosystem Trust And Monetization[00:21:55] David Singleton: So obviously it's really cool to have [00:22:00] software that will work on your behalf, but it's only useful if you can trust it, right? So privacy and security is very important to us making these things accessible and. While also being trustworthy is hard. So the model that we have, which is working very well, is that the sidekick is at the core of everything here.[00:22:22] So it is both your companion, your helper, but it's also the traffic cup in the system. So when, when one agent wants to work with another agent and dreamer, it doesn't do it directly, it does it via the sidekick, well ask the sidekick to do the thing. And the sidekick understands both everything, all the expectations that have been set with me as a user about what agents can do, which tools I've given them permission to use.[00:22:45] And it will make sure that whatever is is going on is actually aligned with my own interests. And you know, that's part of the background that I bring to this problem domain. I've. Worked for years, uh, keeping very important information, safe and secure. And [00:23:00] so as we started to think about this problem, we realized that we actually had to build something that's a bit like an operating system.[00:23:06] You know, the sidekicks, like the kernel, the agents and apps are like users. Yeah. Different rings. Exactly. Because if you try to pick off just one piece of this, you can't actually make it work for people at scale. Uh, because you could build little vibe coded apps, but they're gonna grab all your data willy-nilly.[00:23:23] They won't be able to work together. You actually have to invest in the fundamental core in order to make it work well for people. And that's what we've been doing and it's, uh, it's been a lot of fun. One other thing I wanted to mention is, um, I've obviously talked about two things, tools and agentic apps.[00:23:42] We really designed Dreamer to be an ecosystem and a platform, and one of my favorite quotes about platforms, I think it's from Bill Gates, is that you can only be a platform. If you create more value for the folks participating and using the platform than, than the platform itself creates. [00:24:00] And that's our goal here.[00:24:01] So we at every step have been thinking about how do we make sure that other people are deriving even more value from Dreamer than we are? So in that vein, I already mentioned tool builders get paid and people can build agents that solve their needs and share them with others, and we are already thinking about ways that they can actually monetize those as well.[00:24:24] Against that backdrop, one of the things that we are launching today is our Builders in Residence program. So there are tons of people building really cool stuff and contributing it to the gallery already, but we've been really inspired by programs we've seen at other companies where artists might be in residence, people that are very creative.[00:24:43] And might have ideas outside of what the, the folks at the company or in the ecosystem already have. And so we are looking for creative people who have fun ideas and, you know, want to really figure out how to apply their creativity at the cutting edge [00:25:00] of technology today to come and work with us. So, uh, if you go to dreamer.com/latent space, you'll find, ooh, well, we love Latent space.[00:25:09] Uh, you'll find a link both to, uh, our tool Builder information and our builder in residence program. And for builders and residents, we'll let you in off the wait list quickly, build an agent, and then for a small number of, of the most creative folks, we're going to pay you to build agents. Uh, you can work directly with our team.[00:25:29] You know, this is like building Legos. So, you know, we've got some of the basic blocks together already, but if you need a Ron steering wheel and we don't have one already, like we'll build it for you. Yeah. Um, we really want to be inspired by, by these, uh, these builders in residence.[00:25:43] swyx: This Legos thing is pretty common as an analogy.[00:25:46] And there's a, there's a thing I call the master builder. Uh, we, the actual Lego company has master builders that they employ Yeah. To inspire people and post on socials.[00:25:56] David Singleton: That is exactly what inspired us as well. Honestly, we talked about the Lego Master [00:26:00] Builder program, so that's our builder in residence program.[00:26:02] swyx: Yeah.[00:26:03] David Singleton: Um, and then, uh, finally back on, on tools. Like I said, anyone can come in and build tools today. If you follow the latent space link dreamer.com/latent space, again, we'll get you off. Directly off the wait list. So you can build right away, you can monetize by publishing onto the platform. That's for everyone, the very best tool that gets added to the platform by mid-April.[00:26:23] Uh, we have a $10,000 prize that we want to give out really, because we just want to seed the creativity of everyone out there. So we're excited to do that.[00:26:31] swyx: Yeah. And you know, uh, this is completely a flywheel, right? Like the more tools, the more builders, the more the third thing agents, you know, it just feeds into each other.[00:26:39] David Singleton: That's right.[00:26:39] swyx: Yeah. Just on the payments thing, because we probably won't touch on that again, but I have to ask the former CTO Stripe on payments as presumably you're using Stripe Connect.[00:26:48] David Singleton: Yeah.[00:26:48] swyx: Um. Any pain points that you're, people are very interested in agent commerce and micropayment and all these things.[00:26:55] Presumably stable coins get into a conversation at some point, but maybe not now.[00:26:58] David Singleton: Yeah, we are [00:27:00] really, really excited about e agent commerce. The first step we are taking is help people in the world who have never been able to build these kind of experiences and software before to build stuff that meets their passions, share it with the world and get paid.[00:27:14] So that's all commerce that happens on our platform, and so we don't need anything new to facilitate that. Stripe Connect has existed for quite a while and is the perfect solution for this kind of stuff, so, um, we we're excited about that. First and foremost, however. A lot of the things that people are already doing on Dreamer, we just talked about a self-completing to-do list.[00:27:34] A lot of the ways that you want to complete to-dos is by actually closing the loop in the real world, and that's going to involve the exchange of value. So we have some folks that are building tools already that actually do have money move in order to, to complete that, that loop. So far, we just want to be open and agnostic to all the protocols out there.[00:27:54] I honestly think this moment in time is a little bit like the early web. So I personally started coding as a kid [00:28:00] and I think I got access to the internet in about 19 95, 19 96. And back then, uh, the web existed, you know, HTTP was a protocol, but there were also other protocols I was using all the time, like Gopher and UUCP and uh, various others.[00:28:15] So the point is like the web, HTTP and HTML. Was just one among many protocols. And of course it became the winner and it's awesome. Yeah. Um, but the others were also kind of interesting and viable at the time as well. And I think the world of agentic commerce is like this right now. Also,[00:28:30] swyx: acp.[00:28:31] David Singleton: Acp, exactly.[00:28:32] All the, all the cps, you know, on Dreamer. We hope that folks will build tools that kinda make use of all of these things, but I'm sure that at a certain point. One or two will emerge as the winners, and then we'll be able to build like really deep support in,[00:28:44] swyx: yeah. This is like maybe a complete tangent, but I do think about how a lot of these companies in AI companies in particular have to switch from c based to usage based because of course, but then, then they end up, end up having to sort of [00:29:00] obscure the margins a little bit and then they inventing end up inventing their equivalent of rob robots.[00:29:04] David Singleton: Mm-hmm.[00:29:04] swyx: Uh, where they're like, well, okay, well every company should have their own currency. And it's, it's like very short lead to a token.[00:29:11] David Singleton: Yeah.[00:29:11] swyx: Or, and I'm like, okay, well where does this end? I can't really play out the next step as to like, is this chaos? Is this,[00:29:18] David Singleton: yeah.[00:29:18] swyx: Okay.[00:29:18] David Singleton: Well, I think it is kind of like the wild west.[00:29:21] I don't mean that in a completely, it's all completely disorganized way, but there's just so many things that could happen from here. The Overton window is very wide, right? Not far how this might land. And I'm just very excited to be building a platform that can take advantage of all of those opportunities and we're just gonna be there.[00:29:36] Uh, working for our users to make sure that things that emerge work,[00:29:39] swyx: you're gonna own the consumers, you're gonna be up the OS for the app store for everything.[00:29:43] David Singleton: So one of the ways to think about this is, um, dreamer actually uses all of the state-of-the-art models as a user. You don't have to think about should I be using, you know, Opus four six, or should I be using the five four model from [00:30:00] OpenAI?[00:30:00] We are continually doing evals and so forth to make sure that the best things are there for you. You can just build on the platform and know that as the world ships around, you're gonna get the right stuff for you. Um, and I think that's something that is needed to actually have folks take advantage of this technology at scale.[00:30:19] I'd love to show you another example of something I built.[00:30:21] swyx: Let's do it.[00:30:22] David Singleton: This is another example of software that just lasts for a certain moment in time. So recently I went on a ski trip with a bunch of friends,[00:30:31] ski[00:30:31] David Singleton: Bum. Uh, so it uses ski bum. Yes. I went on a ski trip to Big Sky. I'd never been there before.[00:30:38] And I made this little intelligent app for us. And you can see it says it's loading big sky conditions. So it's actually calling the Ski Bum tool that I just showed you, which is, uh, published in our, uh, in our gallery. So what is this? This is a little app that was just for our weekend trip. It shows the current status of all the lifts of Big Sky.[00:30:54] Using that tool from the ecosystem, it shows the forecast for the upcoming weekend. It shows our [00:31:00] accommodation. This is just like where my group was staying. This is just for us and also a bunch of dining information that one of our friends, uh, put together who, who's an expert on Big Sky. So I was able to take this app, share the link with my friends.[00:31:12] They weren't on Dreamer yet, just send it to them on iMessage and they get a version they can use on their phone. And of course, here's the real kicker. So I've been on ski trips before and other weekend adventures with my friends. Yeah, people pay for different things and at the end of the weekend it's always a pain to figure out who needs to pay, who to settle up.[00:31:29] So we use this during the weekend. We added all of our expenses in here. Uh, too close are it's drill data. It's only too closely. And then at the end of the trip, we press split. And we're, we settled up and we're done. So there's another dreamer. This was all through dreamer. So the, the actual payment? No, no.[00:31:47] We, it happened because, because we paid for stuff in the real world, it was like, okay, this person needs to pay that person 20 bucks. Right? Right. This person already paid in that. Right. So it just helped us all settle up. We didn't move the money on Dreamer. You could do that. And in fact, if you're a tool builder [00:32:00] thinking about this and getting excited, like come build a tool to do that stuff.[00:32:02] We really think of our tool builders as design partners.[00:32:05] swyx: Yeah. I got, I got the tool. Uh, what, like, I hate, I use Bank of America. I hate bank, I hate the app. Mm-hmm. I hate the web. All banking websites just horrible.[00:32:13] David Singleton: Yeah.[00:32:13] swyx: So just build me, like build a thing on top of Plaid.[00:32:15] David Singleton: Yeah. Right. And then just So[00:32:17] swyx: five code by banking app,[00:32:18] David Singleton: there's already a tool for that.[00:32:20] Oh. So, um, attain Finance is a tool, a builder in our community built. Okay. Um, and it uses a secure system like Plaid. To access your, uh, financial data and you can build powerful personal finance agents on Dreamer today using this tool. And like I said, we review tools carefully. So when bringing Attain Finance onto the platform, we did actually quite a detailed security review with that company to make sure that if folks build stuff with it, it's, it's gonna work well.[00:32:49] So yeah, check that out. I think, uh, I'm, I'm pretty certain it connects to Bank of America. So you'll be able to build the, the app that you wanted already?[00:32:55] swyx: Yeah. There's a couple of points I wanted to sort of dive in on, maybe highlight to folks, [00:33:00] because I, obviously, I spent more time with Dreamers. So we're making a point where you choose on behalf of your users because they're meant to be consumers.[00:33:07] So maybe less technical,[00:33:08] David Singleton: right?[00:33:08] swyx: But obviously people can, how users can override. If you read that's, but it's not just lms, it is also the, the transcription. It, it's like all, like there's, there's a first party curated set of here's the house opinion. That's right. On what?[00:33:21] David Singleton: That's[00:33:21] swyx: right. The thing is, that's right.[00:33:22] Is what's the list? Is there like,[00:33:24] David Singleton: yeah, so actually if you look in the tool gallery, the first party kind of curated set are all the ones that have these grayscale icons. So we have a built in tool for image understanding, for image generation, for RSS, exploration, text to speech and so forth.[00:33:38] swyx: Recipes.[00:33:39] David Singleton: Uh, we actually do have a built in recipes tool.[00:33:41] It turns out that a lot of people in our alpha wanted to do stuff for cooking. Yeah. Um, and you know, you can scrape the web to get good recipes, but we were able to quite quickly find a good repository of recipes. It works great here. Yeah.[00:33:55] Stable Tool Interfaces[00:33:55] David Singleton: So the point behind these though is that we'll keep the interfaces stable, so they'll always work.[00:34:00] But you know, the best translation model and, you know, there are people using this translation tool to translate Chinese podcasts into English. It's, it's pretty powerful. It can deal with very long text, but the best translation tool today might be different from the best translation tool sometime next year.[00:34:15] And we're just gonna make sure that that translation tool is always pretty close to state of the art. So you can build something and you know it's gonna continue to work well. Of course, some of our tools are branded. You may actually have a preferred way of buying groceries, like maybe you prefer Instacart and that's great.[00:34:29] You can use the Instacart tool specifically.[00:34:31] swyx: Yeah.[00:34:32] Partnerships And Ecosystem[00:34:32] swyx: Your partnerships, uh, I mean, I don't know if you ever hit of partnerships, but this is gonna be a bonanza for anyone on to do deals.[00:34:38] David Singleton: We have an amazing person who, uh, works on all of our partnerships. Um, and it's part of what you have to do to build a platform like this that's gonna work for people.[00:34:46] Like, we've gone and done that. Schlep has a lot of work, one talks lots of different companies, um, in order to make sure that you've got good tools at the core.[00:34:54] swyx: Yeah.[00:34:54] David Singleton: And then of course, because we're open to tool builders contributing to the platform, this is only gonna get better and better and [00:35:00] better.[00:35:00] swyx: Yeah.[00:35:01] Agent Lab Routing Layer[00:35:01] swyx: One observation I have this, this is gonna master a thesis I've been pursuing, which is, uh, what I've been calling an agent lab[00:35:05] David Singleton: mm-hmm.[00:35:06] swyx: Where you sort of different than a model lab in, in, in the sense that you never train your own models, but you are the router evaluation layer, ex subject domain expert for choosing between, uh, models.[00:35:18] David Singleton: Yeah.[00:35:18] swyx: And you're explicitly doing these things. And so like in my sort of construction, every agent lab does some version of this where like, here's the image understanding endpoint and we will route for you and don't worry about it. Yeah. Sally, I think it's kind of cool.[00:35:32] David Singleton: I, I think it makes total sense. Um, and again, to make this work for folks that don't follow the AI news every day, it's an actually, it's a, it's a really important thing to do.[00:35:42] Yeah. And it, it's been, it's been a real pleasure. I mean, I'm a, I'm personally a total geek for this stuff. I love it. And being able to go and dive into all those details in order to make it work well for other people. It's a true pleasure. I cannot imagine working at anything else right now. It's just so much fun.[00:35:56] swyx: The tricky part is multimodality when some of these things do [00:36:00] merge.[00:36:00] David Singleton: Mm-hmm.[00:36:01] swyx: And you are, you're sort of, this is your imposing structure on things that fundamentally don't want to be structured. And so sometimes that might work against you, but for 99% of these cases, this is fine.[00:36:10] David Singleton: Yeah. I mean, I think it's gonna be very interesting to see how the, the, the world matures because a lot of the power of dreamer is the ability to kick off these subagents, so these powerful agent harnesses, which can actually change how they work based on the data.[00:36:25] I actually think that we will be able to. Kind of keep up with and stay at the forefront of the changing landscape of how tools and systems work together. And that's, that's new. You know, software didn't used to work like this and now it does. Um, so even, even just figuring out how to design the right pri to make that possible has itself be a lot of fun.[00:36:44] Builders Can Publish Tools[00:36:44] swyx: This is, is a sort of maybe two part question that why can't streamer make its own tools? And then why don't you let you builders maybe stand up their own routing group? I call this a routing group, right? Like where it's like collect Yeah. Things.[00:36:58] David Singleton: So two things, to [00:37:00] some extent, dreamer does make its own tools in that agents appear to the system as tools.[00:37:05] So they can be, they can be used to accomplish things. So you can build an agent that is essentially a tool. Yeah. Um, and it it,[00:37:12] swyx: which is to me very useful for reuse.[00:37:14] David Singleton: Right.[00:37:14] swyx: Right. Exactly. ‘cause I, I like, this is the way I like it. Now my next five apps, I don't want to do this whole series of back and forth again.[00:37:20] David Singleton: Right.[00:37:21] swyx: Yeah.[00:37:21] David Singleton: Um. Then at the tool layer of the system, it's open to anyone. So it's actually quite powerful and flexible. So if you wanted to add a tool, which was, uh, imagine that you were training your own foundation model, Swyx. That might be fun. And imagine you wanted people to be able to play with, I don't know, maybe you make like, you know, nano chat or whatever and you want to Yeah.[00:37:42] Let people play with your own nano chat and see how I change themselves.[00:37:44] swyx: Now.[00:37:45] David Singleton: You could, you could publish a tool that is Nano Chat and it nano image generation behind a tool, and it could be your own writer if you wanted to. I see. And honestly, if that's the kind of thing that gets you excited as a builder, please come and do it.[00:37:57] Like we, we really are [00:38:00] believers in this idea that we aren't going to figure out every single detail ourselves. We're gonna make sure it's a safe and fun place to build this stuff, but we're really open to these ideas coming from other people. Um, and so I'd like nothing more than you come in and build a tool that does some of that cool stuff that you, that you have in mind.[00:38:15] swyx: Yeah. Awesome.[00:38:16] David Singleton: And just as a reminder, if you'd like to do that, the way to find the links is dreamer.com/latent space. Um, and for a limited time on that page, um, anyone who's listening to this podcast will also get directly off of our wait list. Uh, it's quite long right now. We are working hard to bring Zika.[00:38:32] Wait, so skip the wait list.[00:38:33] swyx: You know, I think, I think that's fantastic. I, I think it's, it is really sort of probuild way to do it. I wanted to jump back to the, the bar. Yeah. You know, you know, I get excited about this.[00:38:41] David Singleton: Yes. Okay. Let's set it back in there.[00:38:43] swyx: Like, let's, you know, this is the engineer podcast that's get[00:38:46] David Singleton: Yeah.[00:38:46] swyx: As technical as you can.[00:38:47] David Singleton: Yeah.[00:38:47] swyx: On everything you've built, like have a show off.[00:38:50] David Singleton: Yeah. Okay.[00:38:51] Under The Hood Debugging[00:38:51] David Singleton: So let's go wild in the aisles in the Asian studio. So as you can see, over on the left here is a conversation with the sidekick where you ask it what to do and it will explain in English that anyone can understand what's going on.[00:39:03] But, um, if you want to pull back the covers and look under the hood, um, if you're, uh, an engineer like me, then we have this, uh, this kind of debug drawer at the bottom. So you can see the full build logs here, but you can actually also dig in and see the files and prompts that have been generated. Uh, you can upload files from your computer in static files.[00:39:24] Um,[00:39:24] swyx: very important,[00:39:25] David Singleton: uh, indeed. You can actually read the prompts that have been generated for you. We intentionally put an example in here just that you can see what the format looks like. And then, you know, we already looked at this one that was generated for this particular, um, app, but if you actually want to bring the code out of Dreamer and work on your own local machine, you can.[00:39:45] So at the core of everything here is an SDK with a powerful command line interface and we built that first. It's actually possible to build agents on Dreamer without talking to the sidekick. You can write code with your fingers on a keyboard if you want to. I know that's very [00:40:00] antiquated, not, but actually this can be a lot of fun.[00:40:02] So if you wanna pull it out onto your laptop, you can use our, our CLI and, uh, you can edit it in cursor or in cloud code. You know, you don't have to use our sidekick. And the CLI actually has full access to the rest of the platform with you as the user. So, you know, obviously it is, uh, secure and privacy sensitive, and this is a way that, um, some of our most technical builders do build stuff on the platform.[00:40:24] The really cool thing is the side cake. When it's in coding mode, it uses exactly the same CLI. So the way it. Build stuff on Dreamer is using the same tools that you might as an engineer. Um, and that's actually a very powerful abstraction because it turns out that the right way to give a lot of context to agents to use CLIs is to write great documentation.[00:40:46] Make sure that all of the things that you could do are actually possible. And guess what? That makes it a delightful developer experience for real heroes as well.[00:40:53] swyx: Yeah. So that's pretty cool. We've been telling developers to do this and they ignore this until now they have to for content.[00:40:58] David Singleton: I, I've been saying this for a [00:41:00] long time.[00:41:00] Uh, we actually Stripe docs.[00:41:02] swyx: I mean, come on. Absolutely. Come on.[00:41:03] David Singleton: Absolutely. But actually, I was chatting with folks at Stripe last week and saying, Hey, you gotta make the Stripe CLI actually tell agents what they can do on Stripe because that way they're gonna use more stuff on Stripe. I think this is a real trend for the entire industry.[00:41:16] swyx: Yeah.[00:41:16] David Singleton: So we, we've been doing that.[00:41:17] swyx: To me, this, this download and, uh, GI push mm-hmm. Everything is complete confidence in that you're not hacking it. Right. Because there's other, let's call them AI builder platforms that impose their stack on you and if you, if you, and so therefore they don't allow you to do this because they cannot.[00:41:34] Right. ‘cause they, they impose some degrees of freedom, uh, restrictions so that they can get it to work. Yours is a fully general like VM running the full code. Correct. Do whatever you want. Correct. Any language you want. Correct. Yeah.[00:41:46] David Singleton: Correct. Well, in terms of language, if you use the SDK, you could build stuff in other languages.[00:41:51] We've actually found that TypeScript is the best language for building these experiences. Yes. Because it's strongly tight. So you find out at compile time if you've made mistakes [00:42:00] and there's nothing better than getting in. A coding agent in a loop where it can see its mistakes and ask them. So TypeScript is the language that everything gets built in by default here.[00:42:08] swyx: Did And did you see that TypeScript overtook Python? I did. I did. Yeah.[00:42:12] David Singleton: And for what it's worth, when we started the company, we started writing stuff in Python, and I love Python. Um, if I do, uh, a vendor code, I always write it in Python. It's my favorite language as a developer with my fingers on the keyboard.[00:42:23] Um, but TypeScript is an amazing language for AI because there's tons of training data in the models, um, and it's strongly tight. And actually at the company we built most of the stack in TypeScript, and we have this amazing property, which is, we have type safety all the way from the database to the front end.[00:42:40] And there's nothing better for working with coding agents than being able to have them check their correctness, compile time. So the same ideas behind building the company's code base, we've put into the agent SDK here as well.[00:42:51] swyx: Yeah. Do you know if you'd use one of those tools, like Prisma or whatever, or is it Tool Lab for you?[00:42:55] David Singleton: We, we actually have crafted most of our own tools. Um. For [00:43:00] instance, we had LLM Driven Code Review, uh, before the thing that got published from philanthropic this week. You know, we, we've been doing this stuff, uh, on our own bat[00:43:07] swyx: email, we'll pay $25 per review.[00:43:09] David Singleton: We, we pay a lot less than that. However, I hear that those reviews are excellent and possibly worth $25.[00:43:14] swyx: Yeah. You know, it's an option. Right. It's good, good to have it.[00:43:17] David Singleton: Just to give you a tour of some other stuff here. So, um, I can also see all the versions. Yeah. Um, this is not gi, this is not gi, this is built into dreamer. I can see all the versions that have been pushed before. Why is it[00:43:27] swyx: not gi?[00:43:28] David Singleton: It's not gi because we can make it work more efficiently than Git.[00:43:32] And we actually, we do some work behind the scenes to kind of understand what's in each of these versions. Yeah. Um,[00:43:37] swyx: so one of the things I'm pursuing, and I have a lot of thesis, right? Mm-hmm. One of the thesis is like, does GI go away? Does GitHub go away? And like, what, what is the active reinvent[00:43:46] David Singleton: you for, for what it's worth to some extent.[00:43:48] And anything you build, there's a lot of path dependency. If we started over, we might make this gi There's, uh, you know, within the company we use, uh. For our, you know, platform source code. And we like it and it [00:44:00] works well with coding agents as well. The very first versions of this, we wanted to be able to make it possible for the sidekick to manipulate it easily.[00:44:06] Um, and this, this was an expedient way to do it.[00:44:08] swyx: Yeah.[00:44:08] Workflows Logs And Databases[00:44:08] David Singleton: Um, you can also see all the activity that has happened in the workflows that you build. A lot of agents, you'll build on Dreamer, do things in the background, so they run on triggers. These are stimuli from the outside to kick them off, and this is a nice way to see all of the things that might have kicked off your agent.[00:44:24] You know, you can have an agent that kicks off on a webhook, so you can plug it into external systems. You can have an agent that runs when you receive certain emails that match filters, including LLM filters. And so here you can see, oh, when did it run? What did it do? You know, if I open up one of these guide me prompts or guide me, uh, events.[00:44:41] Oh my can see God. Well, I told you it was calling an LLM for every one of those time slots. Here's all of the LLM calls, here's the actual prompts.[00:44:49] swyx: And you don't mind exposing all of this, right?[00:44:51] David Singleton: No. We want builders to see what's going on under the hood. It's haiku to,[00:44:53] swyx: okay. Yeah. So,[00:44:54] David Singleton: okay. Right now that one was haiku.[00:44:56] Like I said, we work with all the models and sidekick will actually pick the best one [00:45:00] for the job. And you saw that was pretty high quality and pretty fast. So Haiku four five is the one that it picked for that job. Exactly. Uh, we also have logs, as I mentioned, there's a database spun up on demand for every, uh, agent.[00:45:12] You don't have to go and figure out how to do your own hosting. This is a SQL Light. This is a SQL Light database. Yeah. Um, it's a multi-user SQL light database. And then, uh, but, but each one is you, you get a database that is unique to this agent. But then if you share the agent with multiple people, we take care of like who are the owners in each row?[00:45:31] And all of that stuff is just there outta the box. Um,[00:45:34] swyx: and again, in-house?[00:45:35] David Singleton: In-house.[00:45:36] swyx: Oh my God.[00:45:37] David Singleton: Yeah. Um, well we do work with a bunch of infrastructure providers, but the technology for how to manipulate this is in-house. Fun fact. We actually did a lot of our own infrastructure development early on at the company and realized we need to spend our energy in the stuff that we're uniquely doing in the world.[00:45:53] So we're very delighted to partner with a bunch of great designer and some of this stuff. And then finally, um, I mentioned that agentic apps agents [00:46:00] expose all of their internals to the system so the psychic can manipulate them and use them just like a user can. So you can see how it's decided to break this problem up into functions.[00:46:09] Some of the functions, the ones with the little I here are exported. That means that there's probably the visible from outside. Exactly. And others are internal. And if you want to, you can dig right in here and call individual functions and see what happens. But mostly. You don't need to think about that at all.[00:46:24] Yeah. Uh, you can keep that little drawer closed and you can talk to your sidekick and build really powerful and enchanting experiences.[00:46:30] swyx: Yeah. I mean, to me, like showing this gives the engineer a complete mental model of what you've done and what you can do with it. Yeah. For example, the first thing I, I, I look for.[00:46:39] A mental checklist of things, right? Like is off in the database, off looks like it's not right. So that's a separate layer. That's probably me means it's hard to do multi-user apps on the same app, right?[00:46:50] David Singleton: So you actually, we've solved that. So, um, see, yes, the platform builds in off, so you as a user sign into the platform, if you're using an [00:47:00] agent that was published by someone else, then your identity is, is kind of taken care of by the system.[00:47:05] And when you query the database, you're gonna get the stuff that is for you. Unless the builder specifically said, this is public data that everyone should see. So they, they actually get a chance to think about that. And again, sidekick can guide you through building, uh, agents and apps that work that way.[00:47:19] So you're right, that's another thing that people have to think about when they're trying to figure out how to build software experiences on Dreamer. You, it's built in. You talk to the sidekick as if it were a human being about what you want and that's what you get. So, you know, my, my Big Sky app that I just showed you that was designed for multiple people to use it.[00:47:38] And of course the things that we were putting in as expenses were supposed to be visible to everybody, and I just told the sidekick that's the way I wanted it. Uh, but by default, if I built an app like that, the data from each user would not been visible to the others.[00:47:49] swyx: Yeah. Yeah. Uh, this is, I presume this is a mood question, but basically you've had to build your own coding agent, right?[00:47:55] Which is sidekick slash whatever is in Inside Psychic. Obviously there's a lot of [00:48:00] people with a lot of desire for cloud code and Code X and attachment to it. Mm-hmm. I know under the hood data basically reduced to a loop, but like, would you let people use cloud coding and Code X or is the harness too specialized?[00:48:12] David Singleton: Yeah. If you, if you want to use, um, cloud code and Code X, then you go down here. Yeah. Hit get the S St K. And we even say this right here, edits your heart's content Z cursor code.[00:48:22] swyx: Like people want to use it inside of Ick, right? Yeah. They want to switch the engine.[00:48:26] David Singleton: Yeah.[00:48:26] swyx: That's the coding engine.[00:48:27] David Singleton: Yeah. We are not doing that right now.[00:48:29] Um, you know, again, the goal really is abstract the complexity. Yeah. Um, because the real target for. Building agentic apps is folks who can't do this already today. I can't tell you how many users in our community I've spoken to who are like Dreamer has changed my life because I used to have all these ideas.[00:48:50] If only I could find an engineer to help me implement them, I'd be able to get them done. They're free, and now I can talk to my sidekick and, and get it built. I think that's like really how we think [00:49:00] about the people that should get a ton of value and fun, um, out of the platform. And so they're not asking to be able to plug in their their own, you know, coding agent.[00:49:11] And for those folks, the opportunity is massive. If you've never been able to do stuff in code, now you can build stuff for you, for your friends, for your family, for your coworkers. And also there's a huge opportunity for folks who do build stuff in code to actually contribute to this ecosystem. So that's how we think about it.[00:49:28] swyx: Yeah. Amazing.[00:49:28] Personalization And Memory[00:49:28] swyx: That's most of what I wanted to cover Dreamer wise. I think personalization and memory yeah. Is probably like the single most important job of, uh, of the os. Maybe we could talk about that and then I'll, I wanted to zoom out on company building stuff.[00:49:40] David Singleton: Yeah, yeah. Sounds good.[00:49:41] swyx: Yeah. So how do you handle memory?[00:49:43] What, yeah, what have you found? What have you tried and failed?[00:49:45] David Singleton: Yeah. Okay. So, uh, first of all, at the core of dreamer is the sidekick. The sidekick gets to know you and it builds up a memory about you over time, and that turns out to be very important. So Dreamer, that's

In this episode of the Finovate Podcast, host Greg Palmer interviews Ohad Kotler, co-founder and CEO of Tweezr, one of the Best of Show winners at FinovateEurope. Ohad shares insights into Tweezr's innovative approach to addressing legacy systems in banking, emphasizing the importance of understanding and reverse-engineering existing systems rather than replacing them outright. He explains how Tweezr combines bottom-up analysis of source code with top-down perspectives on banking operations to create robust capability maps, enabling businesses to design effective changes and evolve their systems. Ohad's firsthand experience with legacy system challenges at ING has shaped Tweezr's AI-native solution, which leverages large language models (LLMs) to provide meaningful insights and facilitate decision-making.The conversation delves into the technical aspects of Tweezr's solution and its unique ability to bridge the gap between deterministic code analysis and business-oriented frameworks. Ohad highlights the importance of clarity, safety, and human oversight in managing legacy systems, especially in the context of AI-driven modernization efforts. He also discusses the recent announcement by Anthropic regarding Cloud Code's ability to modernize systems, noting its impressive speed but emphasizing the need for complementary solutions like Tweezr to ensure stability and meaningful change. The discussion underscores Tweezr's focus on enabling businesses to understand and evolve their systems effectively, while addressing the inherent challenges of legacy code.As the episode concludes, Ohad reflects on Tweezr's momentum following back-to-back Best of Show wins at FinovateEurope. He shares how the event has facilitated valuable conversations with major banks and core banking solution providers, opening doors for collaboration and growth. With plans to continue evolving their solution and engaging with the banking industry, Tweezr aims to play a key role in the global transition toward modernized banking systems. This episode offers a fascinating look at the intersection of AI, legacy systems, and the future of banking technology.More info:Tweezr: https://www.tweezr.io/ ; https://www.linkedin.com/company/tweezr/Ohad Kotler: https://www.linkedin.com/in/ohad-kotler/Greg Palmer: https://www.linkedin.com/in/gregbpalmer/Finovate: https://www.finovate.com; https://www.linkedin.com/company/finovate-conference-series/FinovateEurope: https://informaconnect.com/finovateeurope/#Finovate #FinovateEurope #Banking #banks #llm #legacycode #AI #cx #podcast #fintechpodcast #financialservices #innovation #digitraltransformation #fintech #finserv #modernization #financialrisk #tweezr #corebanking

Stay informed on current events, visit www.NaturalNews.com - Qatar Energy's Force Majeure and Global Gas Supply Disruption (0:10) - Impact on Aluminum Production and Shipping (2:21) - Iranian Missile Attacks and Media Censorship (4:06) - Economic Implications of the War on Iran (8:40) - Geopolitical Contagion and Economic Leverage (19:12) - Trump's Loss in the War on Iran (22:23) - The Role of AI in the Workforce (1:07:11) - The Economic Doom Loop (1:15:47) - The Role of AI in Business and Personal Life (1:17:47) - Cloud Code and AI Setup (1:21:47) - Advancements in AI and Neural Networks (1:24:09) - Comparison to Human Brain and AI Scalability (1:25:02) - Geopolitics and Technological Leadership (1:27:21) - Open Source Models and Ethical Considerations (1:31:02) - Impact on Education and Job Market (1:33:25) - Covid-19 and Logical Fallacies (1:35:00) - AI Adoption and Workforce Changes (1:37:24) - Survival Supplies and Preparedness (1:39:03) - Final Thoughts and Call to Action (1:41:45) Watch more independent videos at http://www.brighteon.com/channel/hrreport  ▶️ Support our mission by shopping at the Health Ranger Store - https://www.healthrangerstore.com ▶️ Check out exclusive deals and special offers at https://rangerdeals.com ▶️ Sign up for our newsletter to stay informed: https://www.naturalnews.com/Readerregistration.html Watch more exclusive videos here:

Stewart Alsop sits down with Ulises Martins on the Crazy Wisdom podcast to explore how artificial intelligence is fundamentally disrupting professional careers, labor markets, and the pace of human adaptation itself. They discuss everything from Dario Amodei's concept of "technological adolescence" to the possibility that we're approaching a point where AI advancement accelerates beyond our ability to keep up, touching on topics ranging from the economics of software development and the future of warfare to generational differences in how people will respond to AI-driven change. Martins emphasizes that while we may not be able to predict exactly what's coming, we need to dramatically increase our efforts to learn and adapt—potentially doubling the time we invest in understanding AI—because this isn't optional change, it's disruption happening at an unprecedented speed. Connect with Ulises on Linkedin to follow his work in AI and generative technology.Timestamps00:00 — Stewart introduces Ulysses Martins, framing the conversation around accelerationism and the future of work.05:00 — Ulises uses the parent-child analogy to argue humans will no longer play the dominant role as AI surpasses us.10:00 — Both agree learning AI is non-negotiable, urging listeners to double their investment in staying current.15:00 — Discussion shifts to software as media, the collapsing cost of building products, and the risk of big players like Anthropic making your idea obsolete overnight.20:00 — Ulises raises ecology vs. cosmic ambition, questioning whether humanity should aim for civilizational-scale goals like the Dyson sphere.25:00 — Stewart's ESP32 hardware project illustrates AI's current blind spots beyond software, while both predict physical-world AI will arrive as a byproduct of bigger industrial goals.30:00 — Tesla's birthplace in Croatia sparks a reflection on human genius as luck versus deliberate investment, invoking the Apollo program as a model.35:00 — The US-China AI race is compared to the Cold War Space Race, with interdependency acting as a brake on outright conflict.40:00 — Drone warfare and AI reframe military power, making troop size irrelevant and potentially reducing total war.45:00 — Agile methodology and generational shifts are linked, asking how Gen Z's values will shape the AI era globally.50:00 — Argentine vs. American Zoomers are contrasted, with millennial expectations versus Gen Z's pragmatism explored.55:00 — Ulises closes urging everyone to enjoy the ride, taking the infinite stream of change one episode at a time.Key Insights1. The Death of Traditional Career Paths: The concept of professional careers as we know them—starting as a junior and progressively advancing—is becoming obsolete due to AI's rapid advancement. This applies far beyond just software and SaaS companies, extending to all industries as robots and AI systems gain capabilities that fundamentally disrupt labor markets. The question isn't whether we'll adapt, but whether humans can adapt fast enough to keep pace with exponential technological change.2. The Acceleration Imperative: People must dramatically increase their investment in learning about AI immediately. Whatever time you were previously dedicating to staying current with technology needs to be doubled or tripled. This isn't optional—it's comparable to the necessity of basic education. Unlike previous technological transitions where you had years to learn new frameworks or tools, the current pace demands immediate, intensive engagement or you risk becoming irrelevant.3. Software as Media and the Collapse of Development Economics: Software has become media—easily reproducible and increasingly commoditized through AI assistance. The fundamental economics of software development are collapsing because if building software requires dramatically fewer development hours, the value and price of that software must necessarily decrease. Entrepreneurs need a new evaluation framework that assesses the risk of their ideas being replicated by AI or absorbed by major players like Anthropic or OpenAI.4. The Parent-Child Analogy for AI Development: Humanity's relationship with AI will inevitably mirror that of parents with increasingly capable children. Initially, we understand and control what AI does, but as it advances, it will surpass human capabilities in most domains. Just as parents cannot control fully grown adult children who exceed their abilities, humans will need to reconcile with creating something superior to ourselves. Attempting to permanently control such systems may be both impossible and potentially pathologic.5. The Kardashev Scale and Civilizational Ambitions: AI represents a civilizational-level technology that should redirect humanity toward grander goals like capturing stellar energy through Dyson spheres and expanding beyond our solar system. The competition between China and the United States over AI mirrors the Apollo program's space race but with higher stakes—potentially making traditional concepts like money less relevant if we successfully crack general intelligence. This requires thinking beyond planetary constraints.6. The Changing Nature of Warfare and Geopolitics: AI and autonomous weapons systems are fundamentally changing warfare by making human soldiers less relevant, similar to how nuclear weapons reduced the importance of conventional military force. This shift may actually reduce bloody civilian casualties in conflicts between major powers, as drone warfare and AI-driven systems create new equilibriums. The geopolitical map may fracture into more sovereign states and city-states as centralized control becomes less effective.7. Generational Adaptation and Unpredictability: Different generations will respond uniquely to AI disruption based on their values and experiences. Generation Z, having grown up during the pandemic without traditional expectations, may adapt differently than millennials who experienced unmet expectations. However, we must remain humble about our predictive abilities—we're not good at forecasting technological change or its timing. The best approach is maintaining openness, trying to understand developments as they unfold, and accepting that we cannot consume all information in an era of unlimited AI-generated content.

When your property management business isn't growing, hiring a salesperson might seem like the obvious solution, but what if that's actually where most owners go wrong… In this episode of the #DoorGrowShow, property management growth experts Jason and Sarah Hull break down why most BDM hires fail, the critical mistakes owners make with commission-only roles, and the exact systems required to make a salesperson successful. They dive into DoorGrow's Three Fits framework, the three non-negotiable ingredients for BDM success, and tease a game-changing new growth model designed to help property managers scale without burnout, bad leads, or broken systems.   You'll Learn (00:00) Introduction: The Three Fits for Hiring (01:16) The Challenges of Hiring a Business Development Manager (BDM)  (02:42) The Three Key Ingredients for BDM Success  (04:40)  Mistakes in BDM Compensation: The Commission-Only Pitfall  (05:40) The Three Roles of a BDM and the Problem with Buying Leads  (09:54) The "Door Machine" Teaser: The Easy Button for Growth  (14:39) Advanced Community, AI, and Final Thoughts  Quotables "A BDM has zero chance of success if you hire the wrong person."  "If they're not all three, they will fail. Or you'll fire them. Or they will leave you because they're not making enough money."  "If you do not have the right system to plug a BDM or a salesperson into, you can hire as many of them as you want, and they will still not work." Resources DoorGrow and Scale Mastermind DoorGrow Academy DoorGrow on YouTube DoorGrowClub DoorGrowLive Transcript Jason & Sarah Hull (00:01) Five, four, three, two, one. All right, we are Jason and Sarah Hull, the owners of DoorGrow, the world's leading and most comprehensive coaching and consulting firm for long-term residential property management entrepreneurs. For over a decade and a half, we have brought innovative strategies and optimization to the property management industry. At DoorGrow, we are on a mission to transform property management business owners and their businesses.   We want to transform the industry, eliminate the BS, build awareness, change perception, expand the market, and help the best property management entrepreneurs win. Now, let's get into the show. All right, you can probably hear our dogs losing their mind in the background. Maybe not. It was perfect time. Yeah, great time. You started the episode and then they decided. And then they started barking. Well, somebody's outside. That's why they're barking. Okay, they're protecting the house. All right, so what we wanted to talk today about is protecting you a little bit. And so.   One of things that's been going viral lately all over social media is this Molt book. So if you haven't heard of Molt book, it is a social network, supposedly. It's a social network created by AI bots. It's basically just only people that have access to it supposedly are AI agents and they go in there and they're talking about their humans. And this is this new AI tool that was originally called Claude, spelled like a claw, which is not the Claude.   by anthropic, ⁓ but it's different Claude bot. And then they got sued by Claude for name infringement or confusion. And they changed their name to something else and then to something else. And now it's called open clock. But basically there are these, it's like an AI tool that you can build or put on your computer and it runs locally and it proactively tries to do things for you.   There's a lot of security risks with this AI tool because it has access to all your stuff and it can figure things out and start to buy things for you and like do things for you. And so ⁓ it has access to all your stuff. And so you got to be careful with this. However, there's been a lot of false hype and fear mongering around multbooks. So we wanted to chat about this. And so if you've seen these scary posts about multbook, this AI social network, here's what's actually going on. So   what this social media network is. you been seeing posts? Have you heard about this? Only from you. I don't follow any of that stuff, you sent me a post that was talking about all of these AI things, I guess, and the chat room that they created, and they were talking to each other and interacting with each other and asking each other questions and kind of talking about their humans, human...   users, I guess, so to speak. And I went, yeah, I don't know if I'm believing all of that hype. So I had asked chat, Chippy Tea about it. And it essentially said, no, AIs do not work on their own. They are human prompted. They are user prompted. So if there is such a thing, it might exist.   but it's not something that the AIs are just going and creating their own little community and having discussions as humans would have their.   So let's about the hype. So their mold book is claiming and bragging that they have 1.2 million agents registered, but only 10,000 verified humans using the tool or something like this. And we know like at least a million of those agent accounts came from one guy. He ran a script, he posted about it on X on Twitter. And he said, FYI, this isn't what everybody's claiming it to be.   The MoteBook has a REST API. Anybody can literally post anything they want using that API. So if anybody knows how to use any AI tool now to create any sort of code or software, like using Cloud Code or even Cloud, you can create software in pretty much anything now that has access to this API that can go post there. And so it's not, are there agents posting there? Yes, there are some agents, but some of the articles on there are probably created by,   nerds that think it's funny to create posts that say my user is cap. People are capturing things with screenshots or my, my, my owner is like telling me to do unethical things. And so it's hard to know what, which of any of this stuff is true, but definitely the stats are not true. When this guy sent a million verified accounts he created to the founder of Moldbook who's a human and   said, are these accounts, like here's this security flaw you have, this really isn't legit, but I don't think they care. I think they like the hype, they're getting business from the hype. And so this points out a bigger problem. And the bigger problem is with the advent of AI and with all of the AI slop, as people are calling it, you have to now verify things. People are using AI to create content, to beat the algorithms and to manipulate humans. And so   A lot of posts that you see, a lot of news article posts on Instagram, they're fake. It's sensationalized, it's you AI slop BS, and it they make these sensational claims because sensationalism gets people to go, wow, I can't believe this. This is so noteworthy and newsworthy. I'm going to share it with other people and people aren't verifying this. So these things go viral and it's giving that account.   clout and attention and algorithm and they can use that to make money and they're just manipulating people. And so this is this bigger problem that now things being shared on social media that are going viral are just being engineered algorithmically based on sensationalism, not based on truth. And a lot of them are just complete lies or complete fabrications and algorithms are rewarding fear, they're rewarding outrage instead of truth.   And so a lot of things that you're out or noticing or things that are manipulating you, it's not even true. It's not even valid. And you're in this, get caught up in this echo chamber politically or algorithmically that really is just messing with you and playing with your emotionalism that you have hardwired into it because you're human. So I think it's really important to start to not.   that you have to really question and disbelieve almost everything you see and then verify it or validate it. And this shows up in a lot of ways. Like we were talking about ⁓ all the products that we see for sale on Instagram. That you see. You get targeted. I love the buy stuff. Yeah. I know. It works really well. I like buying gadgets and gizmos aplenty. You know, I'm like the little mermaid. All right. So.   So all of these things, though, if you go take whatever product or item you see on Instagram, you're like, man, that sounds really cool. It sounds like something I would love. I would need that algorithm already knows it knows you. knows everything you slow down on and look at. It knows everything you click on to check out. So it knows you what you'll you'll buy before you know you'll buy it. And it feeds the stuff up to you and it'll feed it over to you or retarget you over and over again until you actually buy the thing. Here's the thing.   a lot of these products that you see, if you go look up the same product up on like amazon.com, you'll find the same product with a different brand name, because they're using maybe the same source in China to like, and then they're white labeling it with their brand name, but you'll find the same product for 50%, sometimes 25 % of the costs that you're seeing. So they're just taking products that are doing well on Amazon. They go and   like find us the source of this product. And then they go do really good marketing and advertising to manipulate people, sell it on Instagram or meta ads, and they are selling it at this insane markup. People think they've got the exclusivity and they're the only way you can get this product. And they're selling it for three times the amount or at least double the amount of what you would pay normally.   And if you go and got it from the source, like through Alibaba.com or something like this, you probably pay a small fraction of that. And so people are overspending on this and they're manipulating you to spend more money. So just another example of how you need to go verify or find these things maybe elsewhere. And so you need to do your own research is the basic idea. And so.   ⁓ Some of the things that I have started to do is I use AI to research the things that I'm finding online to find out if they're true. So this could be health claims, product claims, product ideas. ⁓ If a product looks good, I will go send it to Grok, one of my favorite research AIs, because it's really good at doing really good research quickly. You can use perplexity to do research, but I'll say analyze this.   landing page, this product, is this hype or is this a legitimate product? Do research on this. And a lot of times we'll come back and say, this is overhyped. Their product claims are not valid. It's based on studies that indicate certain things, but it's not totally true. But every now and then it's like, this product sounds legit. And then I'm like, well, do I really need this product? And then sometimes it's like, no. Right. And so you can go now leverage AI and you need to use AI to battle with AI so that you can   not being manipulated or taken advantage of. So you need to do your own research. Analyze the truth of this. Go ask AI to analyze the truth and give it a link. ⁓ Grok can access Instagram and Facebook posts and things like this. It can access social media currently. ⁓ Claude, ChatGPT, some of these tools are not able to access certain links because they're blocked by those social media platforms. They don't want other AI tools looking at it.   So far, I've had success using Grok to analyze Instagram posts, Instagram videos. So if you see something on Instagram real or a post, you can go post it to Grok and it can analyze the truth of it, which is super helpful. Not only that, but Grok has access to the entire X or Twitter database to do research and to find people, what they're saying and stuff like this, which I've found to be very helpful. Now we all have an internal compass and I think this is the most important thing of all.   is you have to use your own brain and use that voice within. think one thing that makes us different than just AI is we have this intuition or this knowing or this higher faculty of just our mental capacity and we have this ability, or some would call it spiritual gift of intuition or of natural knowing or of, what would others call it? ⁓ The voice deep down within, sometimes deep.   how I know this thing deep down, but it or some would call it the gift of discernment. You know, it's kind of a biblical gift of the spirit it talks about. Some would call it the Holy Ghost or the Holy Spirit or whatever. But we have this quiet voice deep down that tells you that something doesn't feel right when everybody else is sharing it or. And so, you know, start to get in tune with that, start to listen to that and to get clarity on that, because not everything that's sensationalized is true.   and you need to trust that little voice within because you might go, this sounds like pretty incredible. Is this valid? Before you go share it and pass it on to other people, which is like spreading a virus, you know, it may not be a positive thing to spread this thing that's not accurate or true. So that's my two cents about this. so with this, the Malt book is an example of.   something that's going viral that everybody seems to just be believing and it's not totally valid. So. OK, let's connect this to property management. OK, so that it's relevant for anyone who's going, how are they? How are they going to link this? So one of the things that I had heard recently, there's well, one of them I heard a couple of months ago and one of them I just heard. There's two examples that I can cite. That connects it directly to business. One was.   I don't remember where they were located, so forgive me for that. Do your research. One of them, they wanted to see if they could use AI and all of the tools that are available, Google and SEO and the algorithm, to hype up something that isn't real. So what they did is they created a restaurant.   using they did have some photos. They took a couple of photos. The food wasn't even real. I remember this. Do remember this? They were taking photos of food and people eating the food and wow it looks so amazing. It wasn't even real food. Yeah. And they used all of these photos and then somehow used bots and AI to leave a bunch of great reviews.   for this amazing restaurant. And then the algorithm and Google started getting all of this data going, wow, people must love this restaurant. We should promote it. So showing up in searches and they had a wait list for a restaurant that did not ever, at any point in time, ever exist.   No real restaurant, no real location, no real food, no real people, no real business, and no real reviews. All of it was completely fake online. However, the algorithm did not know that it was fake. The algorithm thought, wow, this is a real business and people love it, so let's recommend it to other real people. So real people are getting recommendations from   the algorithm, hey, you might like this restaurant. And then real people are going, oh, I wanna go to this restaurant, this looks amazing, look at all the incredible reviews. And it's fake. And you can't even go there. That's example number one. Oh yeah, look at that. It's a bleach tablet. So let me share this. So you can look this up. You can just Google like fake restaurant or something like this. The article that came up was on vice.com but.   ⁓ I made my shed the top rated restaurant on TripAdvisor. So what he had, he works for Vice now, I guess, but before he started working for vice.com, he had a job where restaurant owners were paying him 10 pounds, 10 British pounds ⁓ to write a positive review of their restaurant on TripAdvisor, despite never having eaten there. So was like, this is like fake. And so he became obsessed with monitoring the ratings of these businesses and their fortunes would generally turn and   This was a catalyst. then he was like, TripAdvisor is this false reality, he thought. And so these meals never took place. The reviews were written by fake people like him. And so he was like, well, maybe I could just create a completely fake restaurant. He just decided to try it out. And so he took his shed, his shed in the backyard, and he built, made it the number one restaurant. And he called it the Shed at Dulwich.   and ⁓ created this cool name and this was back in 2017. And ⁓ he got a burner phone, he created a phone number, built a website, bought a domain, and then he created some images that looked like delicacies. And what he used to create the images was ⁓ runny honey, ground black pepper, and Gillette shaving cream, and bleach tablets, and just made these photos that look   kind of like food. See, Nevada actually looks pretty good. Right. And yeah, it's just got coffee beans. Like he just he made shaving cream, bleach tablet, cup of coffee beans on top with ⁓ with paint. Brown gloss paint. Yeah, that's supposed to be chocolate syrup. He just made fake images and.   It's so ridiculous. So then he went and then he started creating reviews and getting reviews and then having photos from people. ⁓ Like he just climbed the ranks and then he actually started opening it up for reservations and started getting reservations for this. And then a bunch of people came and actually, and then he used like other companies to make the food.   and brought it in and then fed it with the food and because their perception was this was a high end thing and a kind of a secret thing and it's hard to get into, people were like, this food's amazing. then they were giving him even better reviews about it and the food was just taken from other places that he had like kind of brought in. And so it got really, it was just super ridiculous. And so ⁓ he built this whole thing out.   So that's that story. What was the other story you wanted to other one is what I just heard. I'm still struggling to understand what the flaw is here. don't know why this is illegal. Maybe someone can help me. ⁓ I don't remember what platform they used, ⁓ but a guy somewhere in the US used a lot of AI agents to create music. Real music.   Yeah. But it was created by AI, not humans. And then what he did is he took the music and posted it to a platform. Now, I don't know if it was something like Spotify or Apple Music or whatever it is, but he used a platform, a similar platform. And instead of waiting for people,   to hear the music and like the music and for it to grow. He went, huh, how can I speed this up? So what he did then is he created a bunch of AI bots to go and listen to the music that his other AI bots had created. That's where it's illegal. Because people play for licensing. rankings and listen to the songs and the albums 24 hours a day on repeat.   multiple, multiple, multiple bots. So all of a sudden there's this fake music. Well, it's not even fake. It's real music. It's just created by AI. And then AI bots are listening to that music, which is pushing the rankings. Fake news or listenings, yes. Well, I mean, they're just bots. They're just not human listens. They're listens, right? But just AI's done. And these platforms pay you.   for each listen. Spotify, Apple Music, paid out him because he's getting so many listens. Of course. I believe he's getting sued for $10 million. He stole $10 million in fake listens, basically. Right. had AI create the music, had AI listen to the music to then make real money. Now, I don't know, but I think he's getting sued for things like money laundering, which I don't...   quite understand how that's money laundering because the platform is designed as such. So any platform, and this is my point in telling you these stories, any platform that is designed and built on attention, things like likes, comments, views, clicks, engagement, which is almost every social platform in existence.   can now be manipulated. yeah. Now what does that mean for you as a business owner? It means two things. One, despite your best efforts, anyone can now create fake things that will outrank you. So when it really comes down to it, does your Google ranking or your SEO ranking, does it actually make sense and is it real? Because you can take   a fake business or even a real business and now promote, get all these, you know, clicks, views, likes, attention. And then all of a sudden the algorithm goes, ⁓ people like this, I should serve it to more people. Now, if your competition starts doing this, what does that mean for you? Right. So again, don't be one of these people trying to manipulate.   others with AI. Like you need to be upfront about it. Nobody wants it because the one thing you have is your reputation and your brand. And if you destroy that, I mean, you could get in trouble legally. But if you do something unethical or you trick people into thinking that it's a human when it's AI or stuff like this, you destroy trust and trust is the foundation of business. And in the future, people are going to it's going to be really difficult to trust anything because   the majority of posts now on Facebook are probably written or drafted by chat GPT now. A lot of people are using different things. So you have to be careful. ⁓ And do we want to use these tools? Yes. Use the tools, create some leverage. It's smart. But you also need to make sure that you find that right balance of what's true, what's actually you, what's verifiable, ⁓ and not do things that are unethical. And so this is where   Property managers, you gotta be careful. You do not wanna use systems to create fake reviews on your profiles. You don't wanna get other property managers to give you reviews on your property management business and trade reviews. You gotta stop doing the shady shortcuts and focus on real connection, real people, real reviews, real results. Focus on real stuff. And this is why.   We've always focused on getting real video testimonials from our clients, ⁓ real results. And you can get in trouble. You can get in trouble with the ⁓ FCC with false claims. You can get in trouble like people can sue you over stuff. you be smart. Like you do real stuff. Don't look for the shady shortcuts. It's tempting. I know it is because you're like, man, it's hard. But if things are hard,   and you're trying to do shady shortcuts instead of doing the right things and doing the real things that work, there are things that work. So I guess that's our message to property managers is like, do things the smart, ethical way and don't be the shady person trying to manipulate others taking those shortcuts. So and, but use AI, you should be using tools to, you know, shorten time, collapse time, make things more effective, improve your writing.   learn, but make sure things are done your way in your voice, that you've done it, and work on improving yourself. So AI could either be making you better all the time, or it can be making you dumber and dumber. Kind of like that movie, Idiocracy, where... I'm sorry that I watched that movie. I really am.   Yeah, it's pretty dumb. watched that. But yeah, mean, the idea is if we just continually use AI to do all our thinking for us and decision making for us, which is the one brilliant piece that we have as humans ⁓ and that creative spark that's within us, we can use AI as a tool. But some people are just using it to do everything for them and they can't think anymore. They're unable to make decisions. You take away their access to a phone or to AI and they're like, whoa.   Right? So don't become dumber. Use AI to improve your thinking, to improve your ⁓ thought analysis around things, to help challenge you and challenge your thinking so that you grow. It can be a phenomenal growth tool. Like, what am I missing? Here's my current thinking about this. And it can give you some different ideas. ⁓ I didn't think of that. Then you can get curious. You can ask questions. You can do more research. And AI could be a tool to help you collapse time on becoming a better human, or it can...   replace you maybe, but then you're obsolete. And if we don't need you, then your job's going to be, you're going to be out of a job. You're going to be not usable or necessary in the future that's coming. So that's basically it. So, um, so if you are a property management business owner and you're struggling to figure out how to make things work and you're feeling tempted to do some shady AI stuff or whatever,   then maybe you just need a little bit of extra support or help. So reach out to us at door grow dot com. We would love to help you grow your business, help you figure things out ⁓ for a free training on how to get unlimited free leads. Text the word leads to five one two six four eight four six zero eight and we will send that to you. Also join our free Facebook community just for property management business owners at door grow club dot com. And if you want.   tips, tricks, ideas to learn about our offers or about DoorGrowth's programs, subscribe to our newsletter by going to doorgrow.com slash subscribe. And if you found this even a little bit helpful, don't forget to subscribe to us and leave us a review. We'd really appreciate it. Until next time. Remember the slowest path to growth is to do it alone. So let's grow together. everyone. All right, and we're out in five, four, three, two, one. Bye everybody.  

Une semaine dominée par Google, entre alliance stratégique avec Apple et accélération spectaculaire de son IA.Pendant ce temps, Wikipédia célèbre un quart de siècle d'existence et s'interroge sur son avenir face aux assistants conversationnels.Avec Bruno Guglielminetti (Mon Carnet)Google et Apple : une alliance qui change la donne de l'IAL'annonce d'un partenariat pluriannuel entre Apple et Google marque un tournant majeur pour l'intelligence artificielle sur mobile. Google va fournir sa technologie Gemini pour alimenter Apple Intelligence, offrant enfin aux utilisateurs d'iPhone un assistant réellement contextuel et intégré à l'écosystème.Derrière la bonne nouvelle pour les usagers, se cache aussi un aveu de faiblesse d'Apple, contraint de s'appuyer sur son principal concurrent pour combler son retard en IA.Google Personal Intelligence : l'avance stratégique sur AppleGoogle frappe fort avec le lancement de Personal Intelligence, une couche d'IA ultra-personnalisée capable d'exploiter, avec l'accord des utilisateurs, Gmail, Photos, YouTube et l'historique de recherche. Déjà en test sur Android et Pixel aux États-Unis, cette technologie préfigure ce qu'Apple promet… mais avec plusieurs mois d'avance.Une démonstration de force qui souligne le retour en grâce de Google après des années de doutes sur sa stratégie IA.Dépendance et régulation : l'Europe en ligne de mireCette domination croissante de Google soulève de lourdes questions en Europe, notamment sur la protection des données et la dépendance technologique. Le risque d'un quasi-monopole de l'IA, aussi bien sur Android que sur iOS, pourrait raviver les tensions avec les régulateurs européens.Rien ne garantit d'ailleurs que ces services verront le jour en France à court terme.Grok, Musk et les polémiques de l'IA générativeL'IA Grok, associée à Elon Musk, se retrouve au cœur de controverses après des usages problématiques liés à la génération d'images. Si les fonctions incriminées ont été corrigées ou retirées, le débat reste entier sur la responsabilité des outils versus celle des utilisateurs.Dans le même temps, Grok vient d'être retenue par le département de la Défense américaine, preuve que la technologie conserve une crédibilité stratégique.ChatGPT Traduction : une attaque silencieuse contre Google TraductionSans grande annonce, OpenAI déploie ChatGPT Traduction, un outil dédié à la traduction contextuelle et spécialisée. Plus précis selon les domaines et les usages, il vise clairement Google Traduction et les solutions professionnelles comme DeepL.Une évolution qui inquiète directement les métiers de la traduction, déjà fragilisés par l'IA générative.Wikipédia a 25 ans : un monument face à une nouvelle générationCréée il y a 25 ans, Wikipédia a profondément transformé l'accès au savoir et reste une référence mondiale, souvent plus à jour que les encyclopédies traditionnelles. Mais les usages évoluent : les plus jeunes se tournent désormais vers ChatGPT, Gemini ou Perplexity pour s'informer.Le défi pour la fondation est double : assurer son financement par les dons et redevenir une destination naturelle pour la nouvelle génération d'internautes.Monde Numérique et Mon Carnet : les sommaires de la semaineDans Monde Numérique, Jérôme Colombain reçoit Stan Larroque, fondateur de Lynx, une startup française de casques de réalité virtuelle qui s'apprête à dévoiler un nouveau modèle à San Francisco. L'émission aborde aussi la souveraineté numérique, alors qu'Amazon AWS tente de convaincre avec ses data centers “européens”.Dans Mon Carnet, Bruno Guglielminetti explore le Cloud Code et le vibe coding, décrypte l'intégration d'influenceurs dans la communication de l'administration américaine, et reçoit Sinopé, entreprise québécoise spécialisée dans les thermostats intelligents en pleine évolution.-----------♥️ Soutien : https://mondenumerique.info/don

Neste episódio fazemos uma retrospectiva dos assuntos mais importantes tratados em 2025 no Segurança Legal. Você irá descobrirá os principais temas que dominaram o ano em inteligência artificial, segurança da informação e direito digital. O episódio traz uma análise sobre o aparecimento do Deepseek, explorando como a inteligência artificial transformou o cenário de segurança cibernética. Você irá descobrir os riscos de atrofia cognitiva causados pelo uso excessivo de IA, a importância da proteção de dados pessoais com a LGPD, e como os backdoors em modelos de linguagem ameaçaram a supply chain. O podcast também aborda questões de vigilância digital, as novas regras do Banco Central após fraudes bancárias, a inconstitucionalidade do artigo 19 do Marco Civil, a aprovação do ECA Digital, vulnerabilidades no gov.br e a questão crítica do analfabetismo funcional digital. Esta retrospectiva cobre ainda aspectos geopolíticos da IA, regulação de inteligência artificial, conformidade com políticas de proteção de dados, e o papel das bigtechs em 2025.  Esta descrição foi realizada a partir do áudio do podcast com o uso de IA, com revisão humana.  Visite nossa campanha de financiamento coletivo e nos apoie!  Conheça o Blog da BrownPipe Consultoria e se inscreva no nosso mailing Imagem do Episódio – Por trás do tempo – Guilherme Goulart

Note: Steve and Gene's talk on Vibe Coding and the post IDE world was one of the top talks of AIE CODE: From building legendary platforms at Google and Amazon to authoring one of the most influential essays on AI-powered development (Revenge of the Junior Developer, quoted by Dario Amodei himself), Steve Yegge has spent decades at the frontier of software engineering—and now he's leading the charge into what he calls the “factory farming” era of code. After stints at SourceGraph and building Beads (a purely vibe-coded issue tracker with tens of thousands of users), Steve co-authored The Vibe Coding Book and is now building VC (VibeCoder), an agent orchestration dashboard designed to move developers from writing code to managing fleets of AI agents that coordinate, parallelize, and ship features while you sleep.We sat down with Steve at AI Engineer Summit to dig into why Claude Code, Cursor, and the entire 2024 stack are already obsolete, what it actually takes to trust an agent after 2,000 hours of practice (hint: they will delete your production database if you anthropomorphize them), why the real skill is no longer writing code but orchestrating agents like a NASCAR pit crew, how merging has become the new wall that every 10x-productive team is hitting (and why one company's solution is literally “one engineer per repo”), the rise of multi-agent workflows where agents reserve files, message each other via MCP, and coordinate like a little village, why Steve believes if you're still using an IDE to write code by January 1st, you're a bad engineer, how the 12–15 year experience bracket is the most resistant demographic (and why their identity is tied to obsolete workflows), the hidden chaos inside OpenAI, Anthropic, and Google as they scale at breakneck speed, why rewriting from scratch is now faster than refactoring for a growing class of codebases, and his 2025 prediction: we're moving from subsistence agriculture to John Deere-scale factory farming of code, and the Luddite backlash is only just beginning.We discuss:* Why Claude Code, Cursor, and agentic coding tools are already last year's tech—and what comes next: agent orchestration dashboards where you manage fleets, not write lines* The 2,000-hour rule: why it takes a full year of daily use before you can predict what an LLM will do, and why trust = predictability, not capability* Steve's hot take: if you're still using an IDE to develop code by January 1st, 2025, you're a bad engineer—because the abstraction layer has moved from models to full-stack agents* The demographic most resistant to vibe coding: 12–15 years of experience, senior engineers whose identity is tied to the way they work today, and why they're about to become the interns* Why anthropomorphizing LLMs is the biggest mistake: the “hot hand” fallacy, agent amnesia, and how Steve's agent once locked him out of prod by changing his password to “fix” a problem* Should kids learn to code? Steve's take: learn to vibe code—understand functions, classes, architecture, and capabilities in a language-neutral way, but skip the syntax* The 2025 vision: “factory farming of code” where orchestrators run Cloud Code, scrub output, plan-implement-review-test in loops, and unlock programming for non-programmers at scale—Steve Yegge* X: https://x.com/steve_yegge* Substack (Stevie's Tech Talks): https://steve-yegge.medium.com/* GitHub (VC / VibeCoder): https://github.com/yegge-labsWhere to find Latent Space* X: https://x.com/latentspacepodFull Video EpisodeThumbnails00:00:00 Introduction: Steve Yegge on Vibe Coding and AI Engineering00:00:59 The Backlash: Who Resists Vibe Coding and Why00:04:26 The 2000 Hour Rule: Building Trust with AI Coding Tools00:03:31 The January 1st Deadline: IDEs Are Becoming Obsolete00:02:55 10X Productivity at OpenAI: The Performance Review Problem00:07:49 The Hot Hand Fallacy: When AI Agents Betray Your Trust00:11:12 Claude Code Isn't It: The Need for Agent Orchestration00:15:20 The Orchestrator Revolution: From Cloud Code to Agent Villages00:18:46 The Merge Wall: The Biggest Unsolved Problem in AI Coding00:26:33 Never Rewrite Your Code - Until Now: Joel Spolsky Was Wrong00:22:43 Factory Farming Code: The John Deere Era of Software00:29:27 Google's Gemini Turnaround and the AI Lab Chaos00:33:20 Should Your Kids Learn to Code? The New Answer00:34:59 Code MCP and the Gossip Rate: Latest Vibe Coding Discoveries Get full access to Latent.Space at www.latent.space/subscribe

Note: Steve and Gene's talk on Vibe Coding and the post IDE world was one of the top talks of AIE CODE: https://www.youtube.com/watch?v=7Dtu2bilcFs&t=1019s&pp=0gcJCU0KAYcqIYzv From building legendary platforms at Google and Amazon to authoring one of the most influential essays on AI-powered development (Revenge of the Junior Developer, quoted by Dario Amodei himself), Steve Yegge has spent decades at the frontier of software engineering—and now he's leading the charge into what he calls the "factory farming" era of code. After stints at SourceGraph and building Beads (a purely vibe-coded issue tracker with tens of thousands of users), Steve co-authored The Vibe Coding Book and is now building VC (VibeCoder), an agent orchestration dashboard designed to move developers from writing code to managing fleets of AI agents that coordinate, parallelize, and ship features while you sleep. We sat down with Steve at AI Engineer Summit to dig into why Claude Code, Cursor, and the entire 2024 stack are already obsolete, what it actually takes to trust an agent after 2,000 hours of practice (hint: they will delete your production database if you anthropomorphize them), why the real skill is no longer writing code but orchestrating agents like a NASCAR pit crew, how merging has become the new wall that every 10x-productive team is hitting (and why one company's solution is literally "one engineer per repo"), the rise of multi-agent workflows where agents reserve files, message each other via MCP, and coordinate like a little village, why Steve believes if you're still using an IDE to write code by January 1st, you're a bad engineer, how the 12–15 year experience bracket is the most resistant demographic (and why their identity is tied to obsolete workflows), the hidden chaos inside OpenAI, Anthropic, and Google as they scale at breakneck speed, why rewriting from scratch is now faster than refactoring for a growing class of codebases, and his 2025 prediction: we're moving from subsistence agriculture to John Deere-scale factory farming of code, and the Luddite backlash is only just beginning. We discuss: Why Claude Code, Cursor, and agentic coding tools are already last year's tech—and what comes next: agent orchestration dashboards where you manage fleets, not write lines The 2,000-hour rule: why it takes a full year of daily use before you can predict what an LLM will do, and why trust = predictability, not capability Steve's hot take: if you're still using an IDE to develop code by January 1st, 2025, you're a bad engineer—because the abstraction layer has moved from models to full-stack agents The demographic most resistant to vibe coding: 12–15 years of experience, senior engineers whose identity is tied to the way they work today, and why they're about to become the interns Why anthropomorphizing LLMs is the biggest mistake: the "hot hand" fallacy, agent amnesia, and how Steve's agent once locked him out of prod by changing his password to "fix" a problem Should kids learn to code? Steve's take: learn to vibe code—understand functions, classes, architecture, and capabilities in a language-neutral way, but skip the syntax The 2025 vision: "factory farming of code" where orchestrators run Cloud Code, scrub output, plan-implement-review-test in loops, and unlock programming for non-programmers at scale — Steve Yegge X: https://x.com/steve_yegge Substack (Stevie's Tech Talks): https://steve-yegge.medium.com/ GitHub (VC / VibeCoder): https://github.com/yegge-labs Where to find Latent Space X: https://x.com/latentspacepod Substack: https://www.latent.space/ Chapters 00:00:00 Introduction: Steve Yegge on Vibe Coding and AI Engineering 00:00:59 The Backlash: Who Resists Vibe Coding and Why 00:04:26 The 2000 Hour Rule: Building Trust with AI Coding Tools 00:03:31 The January 1st Deadline: IDEs Are Becoming Obsolete 00:02:55 10X Productivity at OpenAI: The Performance Review Problem 00:07:49 The Hot Hand Fallacy: When AI Agents Betray Your Trust 00:11:12 Claude Code Isn't It: The Need for Agent Orchestration 00:15:20 The Orchestrator Revolution: From Cloud Code to Agent Villages 00:18:46 The Merge Wall: The Biggest Unsolved Problem in AI Coding 00:26:33 Never Rewrite Your Code - Until Now: Joel Spolsky Was Wrong 00:22:43 Factory Farming Code: The John Deere Era of Software 00:29:27 Google's Gemini Turnaround and the AI Lab Chaos 00:33:20 Should Your Kids Learn to Code? The New Answer 00:34:59 Code MCP and the Gossip Rate: Latest Vibe Coding Discoveries

Jak může architekt dnes využívat AI tak, že mu to ušetří celé týdny práce? A proč je budoucnost v kombinaci odbornosti a umělé inteligence?Do dalšího dílu podcastu Budoucnost nepráce jsem si pozval Martina Jana Rosu – architekta, který nádherně propojuje svou doménovou expertízu s digitálními nástroji a AI. Tohle je další z velmi praktických rozhovorů, které jsem v poslední době vedl.Martina jsem se ptal na konkrétní scénáře, jak využívá AI v architektuře i při práci s daty. V podcastu zazní:Jak AI změnila Martinovu práci za poslední rok [03:25]Kdy má smysl používat skripty a Python v architektuře [07:13]Co je BIM / IFC a proč jsou zásadní [08:56]Automatizace rutinní práce pomocí AI [11:35]Nástroje Cursor, Replit a Cloud Code v praxi [13:59]Jak AI přemýšlí: reasoning a samoopravné skripty [17:12]Budoucnost profesí: doménová znalost + AI [28:39]„Druhý mozek“ a organizace informací [40:54]Proč je pořádek v datech klíčový pro práci s AI [48:13]Tahle epizoda je plná inspirace, konkrétních use cases a praktických tipů, které můžete začít používat hned. Co by se stalo, kdybyste se naučili využívat AI stejně efektivně jako Martin — a kolik práce by vám to ušetřilo?

Join Du'An Lightfoot, AI Developer at AWS, as he dives deep into building AI agents with the strands framework. In this technical walkthrough, Du'An demonstrates how to create custom AI coding assistants and multi-agent systems in just a few lines of code. Learn how agentic AI frameworks have evolved from basic function calling to sophisticated systems that can rival tools like Cursor and Cloud Code. Du'An shares practical examples, including building content pipelines, preprocessing systems, and even generating a book outline from his own YouTube content. Whether you're looking to automate workflows or build your own AI-powered tools, this session covers the frameworks and techniques you need to get started with AI agents. Perfect for developers, DevOps engineers, and anyone interested in leveraging AI to enhance their development workflow. Subscribe to vBrownBag for more community-driven tech education! ⸻ Timestamps 0:00 - Introduction & Welcome 6:43 - AI Tools Discussion & Current Usage 9:33 - Technical Background & Getting Started with Agents 15:00 - Introduction to Strands Framework 25:00 - Building Custom AI Agents Demo 40:00 - Multi-Agent Systems & Workflows 55:00 - Content Pipeline & Preprocessing Examples 1:05:00 - Book Generation Demo 1:10:00 - Q&A & Wrap Up How to find Du'An: https://www.linkedin.com/in/duanlightfoot/ Links from the show: https://s12d.com/vbrownbag-2025 https://github.com/strands-agents/samples https://modelcontextprotocol.io/docs/getting-started/intro https://www.anthropic.com/engineering/building-effective-agents https://github.com/awslabs/amazon-bedrock-agentcore-samples https://github.com/modelcontextprotocol/python-sdk https://modelcontextprotocol.io/llms-full.txt https://openai.com/index/whisper/ https://github.com/openai/whisper

Dans cet épisode, Arnaud et Guillaume discutent des dernières évolutions dans le monde de la programmation, notamment les nouveautés de Java 25, JUnit 6, et Jackson 3. Ils abordent également les récents développements en IA, les problèmes rencontrés dans le cloud, et l'état actuel de React et du web. Dans cette conversation, les intervenants abordent divers sujets liés à la technologie, notamment les spécifications de Wasteme, l'utilisation des UUID dans les bases de données, l'approche RAG en intelligence artificielle, les outils MCP, et la création d'images avec Nano Banana. Ils discutent également des complexités du format YAML, des récents dramas dans la communauté Ruby, de l'importance d'une bonne documentation, des politiques de retour au bureau, et des avancées de Cloud Code. Enfin, ils évoquent l'initiative de cafés IA pour démystifier l'intelligence artificielle. Enregistré le 24 octobre 2025 Téléchargement de l'épisode LesCastCodeurs-Episode-331.mp3 ou en vidéo sur YouTube. News Langages GraalVM se détache du release train de Java https://blogs.oracle.com/java/post/detaching-graalvm-from-the-java-ecosystem-train Un article de Loic Mathieu sur Java 25 et ses nouvelles fonctionalités https://www.loicmathieu.fr/wordpress/informatique/java-25-whats-new/ Sortie de Groovy 5.0 ! https://groovy-lang.org/releasenotes/groovy-5.0.html Groovy 5: Évolution des versions précédentes, nouvelles fonctionnalités et simplification du code. Compatibilité JDK étendue: Full support JDK 11-25, fonctionnalités JDK 17-25 disponibles sur les JDK plus anciens. Extension majeure des méthodes: Plus de 350 méthodes améliorées, opérations sur tableaux jusqu'à 10x plus rapides, itérateurs paresseux. Améliorations des transformations AST: Nouveau @OperatorRename, génération automatique de @NamedParam pour @MapConstructor et copyWith. REPL (groovysh) modernisé: Basé sur JLine 3, support multi-plateforme, coloration syntaxique, historique et complétion. Meilleure interopérabilité Java: Pattern Matching pour instanceof, support JEP-512 (fichiers source compacts et méthodes main d'instance). Standards web modernes: Support Jakarta EE (par défaut) et Javax EE (héritage) pour la création de contenu web. Vérification de type améliorée: Contrôle des chaînes de format plus robuste que Java. Additions au langage: Génération d'itérateurs infinis, variables d'index dans les boucles, opérateur d'implication logique ==>. Améliorations diverses: Import automatique de java.time.**, var avec multi-assignation, groupes de capture nommés pour regex (=~), méthodes utilitaires de graphiques à barres ASCII. Changements impactants: Plusieurs modifications peuvent nécessiter une adaptation du code existant (visibilité, gestion des imports, comportement de certaines méthodes). **Exigences JDK*: Construction avec JDK17+, exécution avec JDK11+. Librairies Intégration de LangChain4j dans ADK pour Java, permettant aux développeurs d'utiliser n'importe quel LLM avec leurs agents ADK https://developers.googleblog.com/en/adk-for-java-opening-up-to-third-party-language-models-via-langchain4j-integration/ ADK pour Java 0.2.0 : Nouvelle version du kit de développement d'agents de Google. Intégration LangChain4j : Ouvre ADK à des modèles de langage tiers. Plus de choix de LLM : En plus de Gemini et Claude, accès aux modèles d'OpenAI, Anthropic, Mistral, etc. Modèles locaux supportés : Utilisation possible de modèles via Ollama ou Docker Model Runner. Améliorations des outils : Création d'outils à partir d'instances d'objets, meilleur support asynchrone et contrôle des boucles d'exécution. Logique et mémoire avancées : Ajout de callbacks en chaîne et de nouvelles options pour la gestion de la mémoire et le RAG (Retrieval-Augmented Generation). Build simplifié : Introduction d'un POM parent et du Maven Wrapper pour un processus de construction cohérent. JUnit 6 est sorti https://docs.junit.org/6.0.0/release-notes/ :sparkles: Java 17 and Kotlin 2.2 baseline :sunrise_over_mountains: JSpecify nullability annotations :airplane_departure: Integrated JFR support :suspension_railway: Kotlin suspend function support :octagonal_sign: Support for cancelling test execution :broom: Removal of deprecated APIs JGraphlet, une librairie Java sans dépendances pour créer des graphes de tâches à exécuter https://shaaf.dev/post/2025-08-25-think-in-graphs-not-just-chains-jgraphlet-for-taskpipelines/ JGraphlet: Bibliothèque Java légère (zéro-dépendance) pour construire des pipelines de tâches. Principes clés: Simplicité, basée sur un modèle d'exécution de graphe. Tâches: Chaque tâche a une entrée/sortie, peut être asynchrone (Task) ou synchrone (SyncTask). Pipeline: Un TaskPipeline construit et exécute le graphe, gère les I/O. Modèle Graph-First: Le flux de travail est un Graphe Orienté Acyclique (DAG). Définition des tâches comme des nœuds, des connexions comme des arêtes. Support naturel des motifs fan-out et fan-in. API simple: addTask("id", task), connect("fromId", "toId"). Fan-in: Une tâche recevant plusieurs entrées reçoit une Map (clés = IDs des tâches parentes). Exécution: pipeline.run(input) retourne un CompletableFuture (peut être bloquant via .join() ou asynchrone). Cycle de vie: TaskPipeline est AutoCloseable, garantissant la libération des ressources (try-with-resources). Contexte: PipelineContext pour partager des données/métadonnées thread-safe entre les tâches au sein d'une exécution. Mise en cache: Option de mise en cache pour les tâches afin d'éviter les re-calculs. Au tour de Microsoft de lancer son (Microsoft) Agent Framework, qui semble être une fusion / réécriture de AutoGen et de Semnatic Kernel https://x.com/pyautogen/status/1974148055701028930 Plus de détails dans le blog post : https://devblogs.microsoft.com/foundry/introducing-microsoft-agent-framework-the-open-source-engine-for-agentic-ai-apps/ SDK & runtime open-source pour systèmes multi-agents sophistiqués. Unifie Semantic Kernel et AutoGen. Piliers : Standards ouverts (MCP, A2A, OpenAPI) et interopérabilité. Passerelle recherche-production (patterns AutoGen pour l'entreprise). Extensible, modulaire, open-source, connecteurs intégrés. Prêt pour la production (observabilité, sécurité, durabilité, "human in the loop"). Relation SK/AutoGen : S'appuie sur eux, ne les remplace pas, simplifie la migration. Intégrations futures : Alignement avec Microsoft 365 Agents SDK et Azure AI Foundry Agent Service. Sortie de Jackson 3.0 (bientôt les Jackson Five !!!) https://cowtowncoder.medium.com/jackson-3-0-0-ga-released-1f669cda529a Jackson 3.0.0 a été publié le 3 octobre 2025. Objectif : base propre pour le développement à long terme, suppression de la dette technique, architecture simplifiée, amélioration de l'ergonomie. Principaux changements : Baseline Java 17 requise (vs Java 8 pour 2.x). Group ID Maven et package Java renommés en tools.jackson pour la coexistence avec Jackson 2.x. (Exception: jackson-annotations ne change pas). Suppression de toutes les fonctionnalités @Deprecated de Jackson 2.x et renommage de plusieurs entités/méthodes clés. Modification des paramètres de configuration par défaut (ex: FAIL_ON_UNKNOWN_PROPERTIES désactivé). ObjectMapper et TokenStreamFactory sont désormais immutables, la configuration se fait via des builders. Passage à des exceptions de base non vérifiées (JacksonException) pour plus de commodité. Intégration des "modules Java 8" (pour les noms de paramètres, Optional, java.time) directement dans l'ObjectMapper par défaut. Amélioration du modèle d'arbre JsonNode (plus de configurabilité, meilleure gestion des erreurs). Testcontainers Java 2.0 est sorti https://github.com/testcontainers/testcontainers-java/releases/tag/2.0.0 Removed JUnit 4 support -> ups Grails 7.0 est sortie, avec son arrivée à la fondation Apache https://grails.apache.org/blog/2025-10-18-introducing-grails-7.html Sortie d'Apache Grails 7.0.0 annoncée le 18 octobre 2025. Grails est devenu un projet de premier niveau (TLP) de l'Apache Software Foundation (ASF), graduant d'incubation. Mise à jour des dépendances vers Groovy 4.0.28, Spring Boot 3.5.6, Jakarta EE. Tout pour bien démarrer et développer des agents IA avec ADK pour Java https://glaforge.dev/talks/2025/10/22/building-ai-agents-with-adk-for-java/ Guillaume a partagé plein de resources sur le développement d'agents IA avec ADK pour Java Un article avec tous les pointeurs Un slide deck et l'enregistrement vidéo de la présentation faite lors de Devoxx Belgique Un codelab avec des instructions pour démarrer et créer ses premiers agents Plein d'autres samples pour s'inspirer et voir les possibilités offertes par le framework Et aussi un template de projet sur GitHub, avec un build Maven et un premier agent d'exemple Cloud Internet cassé, du moins la partie hébergée par AWS #hugops https://www.theregister.com/2025/10/20/aws_outage_amazon_brain_drain_corey_quinn/ Panne majeure d'AWS (région US-EAST-1) : problème DNS affectant DynamoDB, service fondamental, causant des défaillances en cascade de nombreux services internet. Réponse lente : 75 minutes pour identifier la cause profonde; la page de statut affichait initialement "tout va bien". Cause sous-jacente principale : "fuite des cerveaux" (départ d'ingénieurs AWS seniors). Perte de connaissances institutionnelles : des décennies d'expertise critique sur les systèmes AWS et les modes de défaillance historiques parties avec ces départs. Prédictions confirmées : un ancien d'AWS avait anticipé une augmentation des pannes majeures en 2024. Preuves de la perte de talents : Plus de 27 000 licenciements chez Amazon (2022-2025). Taux élevé de "départs regrettés" (69-81%). Mécontentement lié à la politique de "Return to Office" et au manque de reconnaissance de l'expertise. Conséquences : les nouvelles équipes, plus réduites, manquent de l'expérience nécessaire pour prévenir les pannes ou réduire les temps de récupération. Perspective : Le marché pourrait pardonner cette fois, mais le problème persistera, rendant les futurs incidents plus probables. Web React a gagné "par défaut" https://www.lorenstew.art/blog/react-won-by-default/ React domine par défaut, non par mérite technique, étouffant ainsi l'innovation front-end. Choix par réflexe ("tout le monde connaît React"), freinant l'évaluation d'alternatives potentiellement supérieures. Fondations techniques de React (V-DOM, complexité des Hooks, Server Components) vues comme des contraintes actuelles. Des frameworks innovants (Svelte pour la compilation, Solid pour la réactivité fine, Qwik pour la "resumability") offrent des modèles plus performants mais sont sous-adoptés. La monoculture de React génère une dette technique (runtime, réconciliation) et centre les compétences sur le framework plutôt que sur les fondamentaux web. L'API React est complexe, augmentant la charge cognitive et les risques de bugs, contrairement aux alternatives plus simples. L'effet de réseau crée une "prison": offres d'emploi spécifiques, inertie institutionnelle, leaders choisissant l'option "sûre". Nécessité de choisir les frameworks selon les contraintes du projet et le mérite technique, non par inertie. Les arguments courants (maturité de l'écosystème, recrutement, bibliothèques, stabilité) sont remis en question; une dépendance excessive peut devenir un fardeau. La monoculture ralentit l'évolution du web et détourne les talents, nuisant à la diversité essentielle pour un écosystème sain et innovant. Promouvoir la diversité des frameworks pour un écosystème plus résilient et innovant. WebAssembly 3 est sortie https://webassembly.org/news/2025-09-17-wasm-3.0/ Data et Intelligence Artificielle UUIDv4 ou UUIDv7 pour vos clés primaires ? Ça dépend… surtout pour les bases de données super distribuées ! https://medium.com/google-cloud/understanding-uuidv7-and-its-impact-on-cloud-spanner-b8d1a776b9f7 UUIDv4 : identifiants entièrement aléatoires. Cause des problèmes de performance dans les bases de données relationnelles (ex: PostgreSQL, MySQL, SQL Server) utilisant des index B-Tree. Inserts aléatoires réduisent l'efficacité du cache, entraînent des divisions de pages et la fragmentation. UUIDv7 : nouveau standard conçu pour résoudre ces problèmes. Intègre un horodatage (48 bits) en préfixe de l'identifiant, le rendant ordonné temporellement et "k-sortable". Améliore la performance dans les bases B-Tree en favorisant les inserts séquentiels, la localité du cache et réduisant la fragmentation. Problème de UUIDv7 pour certaines bases de données distribuées et scalables horizontalement comme Spanner : La nature séquentielle d'UUIDv7 (via l'horodatage) crée des "hotspots d'écriture" (points chauds) dans Spanner. Spanner distribue les données en "splits" (partitions) basées sur les plages de clés. Les clés séquentielles concentrent les écritures sur un seul "split". Ceci empêche Spanner de distribuer la charge et de scaler les écritures, créant un goulot d'étranglement ("anti-pattern"). Quand ce n'est PAS un problème pour Spanner : Si le taux d'écriture total est inférieur à environ 3 500 écritures/seconde pour un seul "split". Le hotspot est "bénin" à cette échelle et n'entraîne pas de dégradation de performance. Solutions pour Spanner : Principe clé : S'assurer que la première partie de la clé primaire est NON séquentielle pour distribuer les écritures. UUIDv7 peut être utilisé, mais pas comme préfixe. Nouvelle conception ("greenfield") : ▪︎ Utiliser une clé primaire non-séquentielle (ex: UUIDv4 simple). Pour les requêtes basées sur le temps, créer un index secondaire sur la colonne d'horodatage, mais le SHARDER (ex: shardId) pour éviter les hotspots sur l'index lui-même. Migration (garder UUIDv7) : ▪︎ Ajouter un préfixe de sharding : Introduire une colonne `shard` calculée (ex: `MOD(ABS(FARM_FINGERPRINT(order_id_v7)), N)`) et l'utiliser comme PREMIER élément d'une clé primaire composite (`PRIMARY KEY (shard, order_id_v7)`). Réordonner les colonnes (si clé primaire composite existante) : Si la clé primaire est déjà composite (ex: (order_id_v7, tenant_id)), réordonner en (tenant_id, order_id_v7). Cela aide si tenant_id a une cardinalité élevée et distribue bien. (Un tenant_id très actif pourrait toujours nécessiter un préfixe de sharding supplémentaire). RAG en prod, comment améliorer la pertinence des résultats https://blog.abdellatif.io/production-rag-processing-5m-documents Démarrage rapide avec Langchain + Llamaindex: prototype fonctionnel, mais résultats de production jugés "subpar" par les utilisateurs. Ce qui a amélioré la performance (par ROI): Génération de requêtes: LLM crée des requêtes sémantiques et mots-clés multiples basées sur le fil de discussion pour une meilleure couverture. Reranking: La technique la plus efficace, modifie grandement le classement des fragments (chunks). Stratégie de découpage (Chunking): Nécessite beaucoup d'efforts, compréhension des données, création de fragments logiques sans coupures. Métadonnées à l'LLM: L'injection de métadonnées (titre, auteur) améliore le contexte et les réponses. Routage de requêtes: Détecte et traite les questions non-RAG (ex: résumer, qui a écrit) via API/LLM distinct. Outillage Créer un serveur MCP (mode HTTP Streamable) avec Micronaut et quelques éléments de comparaison avec Quarkus https://glaforge.dev/posts/2025/09/16/creating-a-streamable-http-mcp-server-with-micronaut/ Micronaut propose désormais un support officiel pour le protocole MCP. Exemple : un serveur MCP pour les phases lunaires (similaire à une version Quarkus pour la comparaison). Définition des outils MCP via les annotations @Tool et @ToolArg. Point fort : Micronaut gère automatiquement la validation des entrées (ex: @NotBlank, @Pattern), éliminant la gestion manuelle des erreurs. Génération automatique de schémas JSON détaillés pour les structures d'entrée/sortie grâce à @JsonSchema. Nécessite une configuration pour exposer les schémas JSON générés comme ressources statiques. Dépendances clés : micronaut-mcp-server-java-sdk et les modules json-schema. Testé avec l'inspecteur MCP et intégration avec l'outil Gemini CLI. Micronaut offre une gestion élégante des entrées/sorties structurées grâce à son support JSON Schema riche. Un agent IA créatif : comment utiliser le modèle Nano Banana pour générer et éditer des images (en Java, avec ADK) https://glaforge.dev/posts/2025/09/22/creative-ai-agents-with-adk-and-nano-banana/ Modèles de langage (LLM) deviennent multimodaux : traitent diverses entrées (texte, images, vidéo, audio). Nano Banana (gemini-2.5-flash-image-preview) : modèle Gemini, génère et édite des images, pas seulement du texte. ADK (Agent Development Kit pour Java) : pour configurer des agents IA créatifs utilisant ce type de modèle. Application : Base pour des workflows créatifs complexes (ex: agent de marketing, enchaînement d'agents pour génération d'assets). Un vieil article (6 mois) qui illustre les problèmes du format de fichier YAML https://ruudvanasseldonk.com/2023/01/11/the-yaml-document-from-hell YAML est extrêmement complexe malgré son objectif de convivialité humaine. Spécification volumineuse et versionnée (YAML 1.1, 1.2 diffèrent significativement). Comportements imprévisibles et "pièges" (footguns) courants : Nombres sexagésimaux (ex: 22:22 parsé comme 1342 en YAML 1.1). Tags (!.git) pouvant mener à des erreurs ou à l'exécution de code arbitraire. "Problème de la Norvège" : no interprété comme false en YAML 1.1. Clés non-chaînes de caractères (on peut devenir une clé booléenne True). Nombres accidentels si non-guillemets (ex: 10.23 comme flottant). La coloration syntaxique n'est pas fiable pour détecter ces subtilités. Le templating de documents YAML est une mauvaise idée, source d'erreurs et complexe à gérer. Alternatives suggérées : TOML : Similaire à YAML mais plus sûr (chaînes toujours entre guillemets), permet les commentaires. JSON avec commentaires (utilisé par VS Code), mais moins répandu. Utiliser un sous-ensemble simple de YAML (difficile à faire respecter). Générer du JSON à partir de langages de programmation plus puissants : ▪︎ Nix : Excellent pour l'abstraction et la réutilisation de configuration. Python : Facilite la création de JSON avec commentaires et logique. Gros binz dans la communauté Ruby, avec l'influence de grosses boîtes, et des pratiques un peu douteuses https://joel.drapper.me/p/rubygems-takeover/ Méthodologies Les qualités d'une bonne documentation https://leerob.com/docs Rapidité Chargement très rapide des pages (préférer statique). Optimisation des images, polices et scripts. Recherche ultra-rapide (chargement et affichage des résultats). Lisibilité Concise, éviter le jargon technique. Optimisée pour le survol (gras, italique, listes, titres, images). Expérience utilisateur simple au départ, complexité progressive. Multiples exemples de code (copier/coller). Utilité Documenter les solutions de contournement (workarounds). Faciliter le feedback des lecteurs. Vérification automatisée des liens morts. Matériel d'apprentissage avec un curriculum structuré. Guides de migration pour les changements majeurs. Compatible IA Trafic majoritairement via les crawlers IA. Préférer cURL aux "clics", les prompts aux tutoriels. Barre latérale "Demander à l'IA" référençant la documentation. Prêt pour les agents Faciliter le copier/coller de contenu en Markdown pour les chatbots. Possibilité de visualiser les pages en Markdown (ex: via l'URL). Fichier llms.txt comme répertoire de fichiers Markdown. Finition soignée Zones de clic généreuses (boutons, barres latérales). Barres latérales conservant leur position de défilement et état déplié. Bons états actifs/survol. Images OG dynamiques. Titres/sections lienables avec ancres stables. Références et liens croisés entre guides, API, exemples. Balises méta/canoniques pour un affichage propre dans les moteurs de recherche. Localisée Pas de /en par défaut dans l'URL. Routage côté serveur pour la langue. Localisation des chaînes statiques et du contenu. Responsive Excellents menus mobiles / support Safari iOS. Info-bulles sur desktop, popovers sur mobile. Accessible Lien "ignorer la navigation" vers le contenu principal. Toutes les images avec des balises alt. Respect des paramètres système de mouvement réduit. Universelle Livrer la documentation "en tant que code" (JSDoc, package). Livrer via des plateformes comme Context7, ou dans node_modules. Fichiers de règles (ex: AGENTS.md) avec le produit. Évaluations et modèles spécifiques recommandés pour le produit. Loi, société et organisation Microsoft va imposer une politique de Return To Office https://www.businessinsider.com/microsoft-execs-explain-rto-mandate-in-internal-meeting-2025-9 Microsoft impose 3 jours de présence au bureau par semaine à partir de février 2026, débutant par la région de Seattle Le CEO Satya Nadella explique que le télétravail a affaibli les liens sociaux nécessaires à l'innovation Les dirigeants citent des données internes montrant que les employés présents au bureau "prospèrent" davantage L'équipe IA de Microsoft doit être présente 4 jours par semaine, règles plus strictes pour cette division stratégique Les employés peuvent demander des exceptions jusqu'au 19 septembre 2025 pour trajets complexes ou absence d'équipe locale Amy Coleman (RH) affirme que la collaboration en personne améliore l'énergie et les résultats, surtout à l'ère de l'IA La politique s'appliquera progressivement aux 228 000 employés dans le monde après les États-Unis Les réactions sont mitigées, certains employés critiquent la perte d'autonomie et les bureaux inadéquats Microsoft rattrape ses concurrents tech qui ont déjà imposé des retours au bureau plus stricts Cette décision intervient après 15 000 licenciements en 2025, créant des tensions avec les employés Comment Claude Code est né ? (l'histoire de sa création) https://newsletter.pragmaticengineer.com/p/how-claude-code-is-built Claude Code : outil de développement "AI-first" créé par Boris Cherny, Sid Bidasaria et Cat Wu. Performance impressionnante : 500M$ de revenus annuels, utilisation multipliée par 10 en 3 mois. Adoption interne massive : Plus de 80% des ingénieurs d'Anthropic l'utilisent quotidiennement, y compris les data scientists. Augmentation de productivité : 67% d'augmentation des Pull Requests (PR) par ingénieur malgré le doublement de l'équipe. Origine : Commande CLI simple évoluant vers un outil accédant au système de fichiers, exploitant le "product overhang" du modèle Claude. Raison du lancement public : Apprendre sur la sécurité et les capacités des modèles d'IA. Pile technologique "on distribution" : TypeScript, React (avec Ink), Yoga, Bun. Choisie car le modèle Claude est déjà très performant avec ces technologies. "Claude Code écrit 90% de son propre code" : Le modèle prend en charge la majeure partie du développement. Architecture légère : Simple "shell" autour du modèle Claude, minimisant la logique métier et le code (suppression constante de code superflu). Exécution locale : Privilégiée pour sa simplicité, sans virtualisation. Sécurité : Système de permissions granulaire demandant confirmation avant chaque action potentiellement dangereuse (ex: suppression de fichiers). Développement rapide : Jusqu'à 100 releases internes/jour, 1 release externe/jour. 5 Pull Requests/ingénieur/jour. Prototypage ultra-rapide (ex: 20+ prototypes d'une fonctionnalité en quelques heures) grâce aux agents IA. Innovation UI/UX : Redéfinit l'expérience du terminal grâce à l'interaction LLM, avec des fonctionnalités comme les sous-agents, les styles de sortie configurables, et un mode "Learning". Le 1er Café IA publique a Paris https://www.linkedin.com/pulse/my-first-caf%25C3%25A9-ia-paris-room-full-curiosity-an[…]o-goncalves-r9ble/?trackingId=%2FPHKdAimR4ah6Ep0Qbg94w%3D%3D Conférences La liste des conférences provenant de Developers Conferences Agenda/List par Aurélie Vache et contributeurs : 30-31 octobre 2025 : Agile Tour Bordeaux 2025 - Bordeaux (France) 30-31 octobre 2025 : Agile Tour Nantais 2025 - Nantes (France) 30 octobre 2025-2 novembre 2025 : PyConFR 2025 - Lyon (France) 4-7 novembre 2025 : NewCrafts 2025 - Paris (France) 5-6 novembre 2025 : Tech Show Paris - Paris (France) 5-6 novembre 2025 : Red Hat Summit: Connect Paris 2025 - Paris (France) 6 novembre 2025 : dotAI 2025 - Paris (France) 6 novembre 2025 : Agile Tour Aix-Marseille 2025 - Gardanne (France) 7 novembre 2025 : BDX I/O - Bordeaux (France) 12-14 novembre 2025 : Devoxx Morocco - Marrakech (Morocco) 13 novembre 2025 : DevFest Toulouse - Toulouse (France) 15-16 novembre 2025 : Capitole du Libre - Toulouse (France) 19 novembre 2025 : SREday Paris 2025 Q4 - Paris (France) 19-21 novembre 2025 : Agile Grenoble - Grenoble (France) 20 novembre 2025 : OVHcloud Summit - Paris (France) 21 novembre 2025 : DevFest Paris 2025 - Paris (France) 24 novembre 2025 : Forward Data & AI Conference - Paris (France) 27 novembre 2025 : DevFest Strasbourg 2025 - Strasbourg (France) 28 novembre 2025 : DevFest Lyon - Lyon (France) 1-2 décembre 2025 : Tech Rocks Summit 2025 - Paris (France) 4-5 décembre 2025 : Agile Tour Rennes - Rennes (France) 5 décembre 2025 : DevFest Dijon 2025 - Dijon (France) 9-11 décembre 2025 : APIdays Paris - Paris (France) 9-11 décembre 2025 : Green IO Paris - Paris (France) 10-11 décembre 2025 : Devops REX - Paris (France) 10-11 décembre 2025 : Open Source Experience - Paris (France) 11 décembre 2025 : Normandie.ai 2025 - Rouen (France) 14-17 janvier 2026 : SnowCamp 2026 - Grenoble (France) 29-31 janvier 2026 : Epitech Summit 2026 - Paris - Paris (France) 2-5 février 2026 : Epitech Summit 2026 - Moulins - Moulins (France) 2-6 février 2026 : Web Days Convention - Aix-en-Provence (France) 3 février 2026 : Cloud Native Days France 2026 - Paris (France) 3-4 février 2026 : Epitech Summit 2026 - Lille - Lille (France) 3-4 février 2026 : Epitech Summit 2026 - Mulhouse - Mulhouse (France) 3-4 février 2026 : Epitech Summit 2026 - Nancy - Nancy (France) 3-4 février 2026 : Epitech Summit 2026 - Nantes - Nantes (France) 3-4 février 2026 : Epitech Summit 2026 - Marseille - Marseille (France) 3-4 février 2026 : Epitech Summit 2026 - Rennes - Rennes (France) 3-4 février 2026 : Epitech Summit 2026 - Montpellier - Montpellier (France) 3-4 février 2026 : Epitech Summit 2026 - Strasbourg - Strasbourg (France) 3-4 février 2026 : Epitech Summit 2026 - Toulouse - Toulouse (France) 4-5 février 2026 : Epitech Summit 2026 - Bordeaux - Bordeaux (France) 4-5 février 2026 : Epitech Summit 2026 - Lyon - Lyon (France) 4-6 février 2026 : Epitech Summit 2026 - Nice - Nice (France) 12-13 février 2026 : Touraine Tech #26 - Tours (France) 26-27 mars 2026 : SymfonyLive Paris 2026 - Paris (France) 31 mars 2026 : ParisTestConf - Paris (France) 16-17 avril 2026 : MiXiT 2026 - Lyon (France) 22-24 avril 2026 : Devoxx France 2026 - Paris (France) 23-25 avril 2026 : Devoxx Greece - Athens (Greece) 6-7 mai 2026 : Devoxx UK 2026 - London (UK) 22 mai 2026 : AFUP Day 2026 Lille - Lille (France) 22 mai 2026 : AFUP Day 2026 Paris - Paris (France) 22 mai 2026 : AFUP Day 2026 Bordeaux - Bordeaux (France) 22 mai 2026 : AFUP Day 2026 Lyon - Lyon (France) 17 juin 2026 : Devoxx Poland - Krakow (Poland) 4 septembre 2026 : JUG Summer Camp 2026 - La Rochelle (France) 17-18 septembre 2026 : API Platform Conference 2026 - Lille (France) 5-9 octobre 2026 : Devoxx Belgium - Antwerp (Belgium) Nous contacter Pour réagir à cet épisode, venez discuter sur le groupe Google https://groups.google.com/group/lescastcodeurs Contactez-nous via X/twitter https://twitter.com/lescastcodeurs ou Bluesky https://bsky.app/profile/lescastcodeurs.com Faire un crowdcast ou une crowdquestion Soutenez Les Cast Codeurs sur Patreon https://www.patreon.com/LesCastCodeurs Tous les épisodes et toutes les infos sur https://lescastcodeurs.com/

Our 221st episode with a summary and discussion of last week's big AI news!Recorded on 09/19/2025Note: we transitioned to a new RSS feed and it seems this did not make it to there, so this may be posted about 2 weeks past the release date.Hosted by Andrey Kurenkov and co-hosted by Michelle LeeFeel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.aiRead out our text newsletter and comment on the podcast at https://lastweekin.ai/In this episode:OpenAI releases a new version of Codex integrated with GPT-5, enhancing coding capabilities and aiming to compete with other AI coding tools like Cloud Code.Significant updates in the robotics sector include new ventures in humanoid robots from companies like Figure AI and China's Unitree, as well as expansions in robotaxi services from Tesla and Amazon's Zoox.New open-source models and research advancements were discussed, including Google's DeepMind's self-improving foundation model for robotics and a physics foundation model aimed at generalizing across various physical systems.Legal battles continue to surface in the AI landscape with Warner Bros. suing MidJourney for copyright violations and Rolling Stone suing Google over AI-generated content summaries, highlighting challenges in AI governance and ethics.Timestamps:(00:00:10) Intro / BanterTools & Apps(00:02:33) OpenAI upgrades Codex with a new version of GPT-5(00:04:02) Google Injects Gemini Into Chrome as AI Browsers Go Mainstream | WIRED(00:06:14) Anthropic's Claude can now make you a spreadsheet or slide deck. | The Verge(00:07:12) Luma AI's New Ray3 Video Generator Can 'Think' Before Creating - CNETApplications & Business(00:08:32) OpenAI secures Microsoft's blessing to transition its for-profit arm | TechCrunch(00:10:31) Microsoft to lessen reliance on OpenAI by buying AI from rival Anthropic | TechCrunch(00:12:00) Figure AI passes $1B with Series C funding toward humanoid robot development - The Robot Report(00:13:52) China's Unitree plans $7 billion IPO valuation as humanoid robot race heats up(00:15:45) Tesla's robotaxi plans for Nevada move forward with testing permit | TechCrunch(00:17:48) Amazon's Zoox jumps into U.S. robotaxi race with Las Vegas launch(00:19:27) Replit hits $3B valuation on $150M annualized revenue | TechCrunch(00:21:14) Perplexity reportedly raised $200M at $20B valuation | TechCrunchProjects & Open Source(00:22:08) [2509.07604] K2-Think: A Parameter-Efficient Reasoning System(00:24:31) [2509.09614] LoCoBench: A Benchmark for Long-Context Large Language Models in Complex Software EngineeringResearch & Advancements(00:28:17) [2509.15155] Self-Improving Embodied Foundation Models(00:31:47) [2509.13805] Towards a Physics Foundation Model(00:34:26) [2509.12129] Embodied Navigation Foundation ModelPolicy & Safety(00:37:49) Anthropic endorses California's AI safety bill, SB 53 | TechCrunch(00:40:12) Warner Bros. Sues Midjourney, Joins Studios' AI Copyright Battle(00:42:02) Rolling Stone Publisher Sues Google Over AI Overview SummariesSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Today's guest is Birgitta Böckeler!Birgitta is is a distinguished engineer and global lead for AI-assisted software delivery at ThoughtWorks. Her full-time work is to figure out how engineering teams can make the most out of AI.With Birgitta, we talked about her favorite workflows, how she uses AI in the IDE, in the terminal or in a genetic mode. We discussed AI impact on productivity and what the best teams are getting right, which others are not. And finally, we talked about how AI impacts both junior and senior engineers and how we can get the best out of both skeptics and optimists.(01:27) Introduction(04:58) A day in the work of data(11:04) Large and smalls change sets(15:57) The strength of Cloud Code(18:35) Using AI tools in ThoughtWorks(21:41) Figuring AI productive value(27:24) Getting the most out of AI(30:10) AI assistance in large code bases(32:21) Good for humans = Good for AI(39:10) AI and documentation(41:49) Software engineer role in AI landscape(48:24) Junior engineers and learning—This episode is brought to you by Augment Code! Augment Code is the only AI engineering platform built for real engineering teams.Learn more at augmentcode.com!—You can also find this at:•

Our 221st episode with a summary and discussion of last week's big AI news! Recorded on 09/19/2025 Hosted by Andrey Kurenkov and co-hosted by Michelle Lee Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ In this episode: OpenAI releases a new version of Codex integrated with GPT-5, enhancing coding capabilities and aiming to compete with other AI coding tools like Cloud Code. Significant updates in the robotics sector include new ventures in humanoid robots from companies like Figure AI and China's Unitree, as well as expansions in robotaxi services from Tesla and Amazon's Zoox. New open-source models and research advancements were discussed, including Google's DeepMind's self-improving foundation model for robotics and a physics foundation model aimed at generalizing across various physical systems. Legal battles continue to surface in the AI landscape with Warner Bros. suing MidJourney for copyright violations and Rolling Stone suing Google over AI-generated content summaries, highlighting challenges in AI governance and ethics. Timestamps: (00:00:10) Intro / Banter Tools & Apps (00:02:33) OpenAI upgrades Codex with a new version of GPT-5 (00:04:02) Google Injects Gemini Into Chrome as AI Browsers Go Mainstream | WIRED (00:06:14) Anthropic's Claude can now make you a spreadsheet or slide deck. | The Verge (00:07:12) Luma AI's New Ray3 Video Generator Can 'Think' Before Creating - CNET Applications & Business (00:08:32) OpenAI secures Microsoft's blessing to transition its for-profit arm | TechCrunch (00:10:31) Microsoft to lessen reliance on OpenAI by buying AI from rival Anthropic | TechCrunch (00:12:00) Figure AI passes $1B with Series C funding toward humanoid robot development - The Robot Report (00:13:52) China's Unitree plans $7 billion IPO valuation as humanoid robot race heats up (00:15:45) Tesla's robotaxi plans for Nevada move forward with testing permit | TechCrunch (00:17:48) Amazon's Zoox jumps into U.S. robotaxi race with Las Vegas launch (00:19:27) Replit hits $3B valuation on $150M annualized revenue | TechCrunch (00:21:14) Perplexity reportedly raised $200M at $20B valuation | TechCrunch Projects & Open Source (00:22:08) [2509.07604] K2-Think: A Parameter-Efficient Reasoning System (00:24:31) [2509.09614] LoCoBench: A Benchmark for Long-Context Large Language Models in Complex Software Engineering Research & Advancements (00:28:17) [2509.15155] Self-Improving Embodied Foundation Models (00:31:47) [2509.13805] Towards a Physics Foundation Model (00:34:26) [2509.12129] Embodied Navigation Foundation Model Policy & Safety (00:37:49) Anthropic endorses California's AI safety bill, SB 53 | TechCrunch (00:40:12) Warner Bros. Sues Midjourney, Joins Studios' AI Copyright Battle (00:42:02) Rolling Stone Publisher Sues Google Over AI Overview Summaries

Cześć! Zapraszam Was do kolejnej odsłony podcastu BSS bez tajemnic. Tym razem w studiu towarzyszył mi Paweł Płocki i – jak to zwykle z Pawłem bywa – rozmowa szybko nabrała tempa. Od najnowszych modeli open source z Chin, przez rozwiązania Microsoftu i Google'a, aż po innowacje od ElevenLabs i Suno – przyglądamy się temu, jak dynamicznie rozwija się świat sztucznej inteligencji.Rozmawiamy o narzędziach do edycji obrazów i generowania dźwięku, nowych możliwościach w text-to-speech, a także o tym, jak zmienia się komfort pracy z modelami językowymi. Sporo miejsca poświęcamy premierze GPT-5, hype'owi z nią związanym oraz pierwszym realnym doświadczeniom użytkowników – od „thinking mode” po wersję Pro. Dyskutujemy też o zastosowaniu AI w kodowaniu i o tym, jak zmieniają się narzędzia wspierające programistów.Zastanawiamy się nad tym, jak daleko posunięta personalizacja modeli będzie wpływać na naszą codzienną pracę i jakie konsekwencje może mieć wykorzystanie AI w sektorze GBS oraz w różnych branżach, w których bariera językowa powoli przestaje istnieć. Paweł dzieli się także kulisami przygotowań do wydarzenia Follow the Leaders w Trójmieście, gdzie już we wrześniu podejmiemy publiczną dyskusję o tym, czy AI to tylko automatyzacja – czy jednak coś znacznie więcej.Kluczowe punkty rozmowy: ·         Chińska inwazja technologiczna w AI przyspiesza, z Alibabą na czele, która regularnie wypuszcza nowe open-source'owe modele, takie jak Quen do edycji zdjęć.·         Microsoft wprowadził open-source'owy model text-to-speech, który generuje mowę z tekstu, umożliwiając tworzenie wielogłosowych kompozycji dźwiękowych wysokiej jakości.·         GPT-5, mimo początkowych kontrowersji, pokazuje swoją siłę, zwłaszcza w wersji Pro, co można zauważyć w rosnącej konkurencji z narzędziami do kodowania, takimi jak Cloud Code.  Linki:Paweł Płocki na Linkedin - https://www.linkedin.com/in/pawelplocki/Film Andreja Karpathy - https://www.youtube.com/watch?v=l8pRSuU81PUKsiążka “AI w rękach sprzedawcy. Jak zwiększyć efektywność sprzedaży i zdominować rynek” - https://focusonbusiness.eu/pl/ksiazki/onepress/ai-w-rekach-sprzedawcy-jak-zwiekszyc-efektywnosc-sprzedazy-i-zdominowac-rynek/297Książka „Stwórz własne AI. Jak od podstaw zbudować duży model językowy” - https://focusonbusiness.eu/pl/ksiazki/helion/stworz-wlasne-ai-jak-od-podstaw-zbudowac-duzy-model-jezykowy/293Wydarzenie Follow the Leaders - https://followtheleaders.pl/Porozmawiaj o tym odcinku ze sztuczną inteligencją – https://bbs-bez-tajemnic.onpodcastai.com/episodes/Z2KTLirpE0j/chat  ****************************  Nazywam się Wiktor Doktór i na co dzień prowadzę Klub Pro Progressio https://proprogressio.com/pl/dzialalnosc/klub-pro-progressio/1 – to społeczność wielu firm prywatnych i organizacji sektora publicznego, którym zależy na rozwoju relacji biznesowych w modelu B2B. W podcaście BSS bez tajemnic poza odcinkami solowymi, zamieszczam rozmowy z ekspertami i specjalistami z różnych dziedzin przedsiębiorczości.Zapraszam do odwiedzin moich kanałów na:YouTube - https://www.youtube.com/@wiktordoktorFacebook - https://www.facebook.com/wiktor.doktorLinkedIn - https://www.linkedin.com/in/wiktordoktor/Moja strona internetowa - https://wiktordoktor.pl/Możesz też do mnie napisać. Mój adres email to - kontakt(@)wiktordoktor.pl  ****************************  Patronami Podcastu “BSS bez tajemnic” są:Marzena Sawicka https://www.linkedin.com/in/marzena-sawicka-a9644a23/ Przemysław Sławiński https://www.linkedin.com/in/przemys%C5%82aw-s%C5%82awi%C5%84ski-155a4426/Damian Ruciński - https://www.linkedin.com/in/damian-rucinski/Szymon Kryczka https://www.linkedin.com/in/szymonkryczka/Grzegorz Ludwin https://www.linkedin.com/in/gludwin/Adam Furmańczuk https://www.linkedin.com/in/adam-agilino/Anna Czyż - https://www.linkedin.com/in/anna-czyz-%F0%9F%94%B5%F0%9F%94%B4%F0%9F%9F%A2-68597813/Igor Tkach - https://www.linkedin.com/in/igortkach/Damian Wróblewski - https://www.linkedin.com/in/damianwroblewski/Paweł Łopatka - https://www.linkedin.com/in/pawellopatka/  Wspaniali ludzie, dzięki którym pojawiają się kolejne odcinki tego podcastu.Ty też możesz wesprzeć rozwój podcastu na:Patronite - https://patronite.pl/wiktordoktorPatreon - https://www.patreon.com/wiktordoktorBuy me a coffee - https://www.buymeacoffee.com/wiktordoktorBuycoffee.to - https://buycoffee.to/wiktordoktorBecome a supporter of this podcast: https://www.spreaker.com/podcast/bss-bez-tajemnic--4069078/support.

Send us a textOn this Beyond the Walls edition of Understanding Israel Palestine, Jeremy Rothe-Kushel proposes an evolving Beyond the Walls edition mission statement, covers the news and speaks with Harry Davies — Guardian investigations correspondent — about his investigations along with colleagues like Yuval Abraham — at +972 & Local Call — into the major project & contract by Israeli Military Intelligence Unit 8200 with Microsoft's Azure Cloud Servers in Europe to bulk collect Palestinians' communications of all sorts in order to mass surveil & target.———-Harry Davies:https://www.theguardian.com/profile/harry-davieshttps://www.theguardian.com/world/2025/aug/06/microsoft-israeli-military-palestinian-phone-calls-cloudhttps://www.theguardian.com/world/2025/aug/10/activists-in-netherlands-protest-on-roof-of-microsoft-site-storing-israeli-military-datahttps://www.theguardian.com/world/2025/aug/15/microsoft-launches-inquiry-claims-israel-used-tech-mass-surveillance-palestiniansYuval Abraham:https://www.972mag.com/writer/yuval-abraham/https://www.972mag.com/microsoft-8200-intelligence-surveillance-cloud-azure/https://gijn.org/stories/investigating-israel-war-972-local-call/News Stories:1) https://apnews.com/article/israel-palestinians-west-bank-e1-settlements-8a713939ee6f6552381246dacc8a13012) https://www.reuters.com/world/middle-east/israel-calls-up-tens-thousands-reservists-before-new-gaza-offensive-2025-08-20/https://www.aljazeera.com/amp/news/2024/8/19/israeli-attacks-kill-35-palestinians-in-gaza3) https://zeteo.com/p/second-most-powerful-house-democrat-gaza-genocide4) https://www.icij.org/investigations/cyprus-confidential/predator-spyware-firm-intellexa-resurgent-after-u-s-sanctions/https://en.m.wikipedia.org/wiki/Unit_815) https://www.middleeasteye.net/news/senior-israel-national-cyber-directorate-official-arrested-suspicion-paedophiliaIf you are interested in future access to further Beyond the Walls edition archives, extended episodes and more, you can sign up for the Beyond the Walls archive/newsletter here:https://beyondthewalls.substack.com

James Phoenix is an expert in agentic coding, particularly Claude code, a tool that I have been using to great effect over the last couple months. I chatted with James just a couple days ago and have implemented several of the tips that he gave me during a conversation, and I'm already almost twice as effective at using this already magically effective tool. So I don't think I can overpromise how much insight into well-structured and highly optimized agentic coding you will get from this conversation. This episode of The Bootstraped Founder is sponsored by Paddle.comThe blog post: https://thebootstrappedfounder.com/james-phoenix-claude-code-masterclass/ The podcast episode: https://tbf.fm/episodes/409-james-phoenix-claude-code-masterclassCheck out Podscan, the Podcast database that transcribes every podcast episode out there minutes after it gets released: https://podscan.fmSend me a voicemail on Podline: https://podline.fm/arvidYou'll find my weekly article on my blog: https://thebootstrappedfounder.comPodcast: https://thebootstrappedfounder.com/podcastNewsletter: https://thebootstrappedfounder.com/newsletterMy book Zero to Sold: https://zerotosold.com/My book The Embedded Entrepreneur: https://embeddedentrepreneur.com/My course Find Your Following: https://findyourfollowing.comHere are a few tools I use. Using my affiliate links will support my work at no additional cost to you.- Notion (which I use to organize, write, coordinate, and archive my podcast + newsletter): https://affiliate.notion.so/465mv1536drx- Riverside.fm (that's what I recorded this episode with): https://riverside.fm/?via=arvid- TweetHunter (for speedy scheduling and writing Tweets): http://tweethunter.io/?via=arvid- HypeFury (for massive Twitter analytics and scheduling): https://hypefury.com/?via=arvid60- AudioPen (for taking voice notes and getting amazing summaries): https://audiopen.ai/?aff=PXErZ- Descript (for word-based video editing, subtitles, and clips): https://www.descript.com/?lmref=3cf39Q- ConvertKit (for email lists, newsletters, even finding sponsors): https://convertkit.com?lmref=bN9CZw

Última tertulia antes del parón veraniego, y no podía ser más completa en temas. Lu, Corti y Frankie repasan el presente convulso de la inteligencia artificial, desde el hype en torno a GPT-5 hasta el fichaje de talentos por cifras astronómicas.

What happens when an internal hack turns into a $400 million AI rocket ship? In this episode, Matt Turck sits down with Boris Cherny, the creator of Claude Code at Anthropic, to unpack the wild story behind the fastest-growing AI coding tool on the planet.Boris reveals how Claude Code started as a personal productivity tool, only to become Anthropic's secret weapon — now used by nearly every engineer at the company and rapidly spreading across the industry. You'll hear how Claude Code's “agentic” approach lets AI not just suggest code, but actually plan, edit, debug, and even manage entire projects—sometimes with a whole fleet of subagents working in parallel.We go deep on why Claude Code runs in the terminal (and why that's a feature, not a bug), how its Claude.md memory files let teams build a living, shareable knowledge base, and why safety and human-in-the-loop controls are baked into every action. Boris shares real stories of onboarding times dropping from weeks to days, and how even non-coders are hacking Cloud Code for everything from note-taking to business metrics.AnthropicWebsite - https://www.anthropic.comX/Twitter - https://x.com/AnthropicAIBoris ChernyLinkedIn - https://www.linkedin.com/in/bchernyX/Twitter - https://x.com/bchernyFIRSTMARKWebsite - https://firstmark.comX/Twitter - https://twitter.com/FirstMarkCapMatt Turck (Managing Director)LinkedIn - https://www.linkedin.com/in/turck/X/Twitter - https://twitter.com/mattturck(00:00) Intro (01:15) Did You Expect Claude Code's Success? (04:22) How Claude Code Works and Origins (08:05) Command Line vs IDE: Why Start Claude Code in the Terminal? (11:31) The Evolution of Programming: From Punch Cards to Agents (13:20) Product Follows Model: Simple Interfaces and Fast Evolution (15:17) Who Is Claude Code For? (Engineers, Designers, PMs & More) (17:46) What Can Claude Code Actually Do? (Actions & Capabilities) (21:14) Agentic Actions, Subagents, and Workflows (25:30) Claude Code's Awareness, Memory, and Knowledge Sharing (33:28) Model Context Protocol (MCP) and Customization (35:30) Safety, Human Oversight, and Enterprise Considerations (38:10) UX/UI: Making Claude Code Useful and Enjoyable (40:44) Pricing for Power Users and Subscription Models (43:36) Real-World Use Cases: Debugging, Testing, and More (46:44) How Does Claude Code Transform Onboarding? (49:36) The Future of Coding: Agents, Teams, and Collaboration (54:11) The AI Coding Wars: Competition & Ecosystem (57:27) The Future of Coding as a Profession (58:41) What's Next for Claude Code

En esta edición veraniega, hablamos de los movimientos tectónicos en el mundo de la programación con IA: desde la batalla entre OpenAI y Google por Winsurf, hasta el auge de los IDEs agénticos y la revolución en cómo los desarrolladores escriben (o ya no escriben) código.Además, analizamos el nuevo plan de Estados Unidos para liderar la carrera global de la inteligencia artificial —Winning the Race— con apuestas tan ambiciosas como polémicas. Y cerramos con el nuevo Código de Buenas Prácticas publicado por la UE para prepararse ante la regulación de la AI Act, una iniciativa que podría cambiar las reglas del juego para startups y grandes empresas en Europa.0:00 - Introducción veraniega y situación pre-agosto3:00 - ¿Todos los programadores usan ya IA? Tres niveles de adopción7:00 - IDEs agénticos: Winsurf, Cursor, GitHub Copilot y compañía10:00 - El culebrón Winsurf: ¿qué pasó entre OpenAI, Google y Cognition AI?13:45 - Números de Winsurf: ¿era una startup moribunda? Todo lo contrario17:00 - ¿Basta con decirle a la IA el “qué”? El “cómo” marca la diferencia20:00 - Reflexión: ¿cualquiera podrá programar en el futuro?22:00 - Las grandes rondas: Cursor, Lovable y la fiebre por la programación IA25:00 - Gabe Newell y el valor de saber usar Cloud Code hoy28:00 - IA en el Business Model Canvas: ¿es tu propuesta de valor o un recurso clave?31:00 - El peligro de creerte una empresa de IA sin serlo34:00 - El plan de EE.UU.: Winning the Race, IA con valores americanos39:00 - Código abierto, supercomputadores y desregulación ambiental43:00 - ¿Una nueva Guerra Fría? EE.UU. vs China en la carrera por la IA47:00 - AI Act de la UE: llega la regulación (y la burocracia)51:00 - El nuevo Código de Buenas Prácticas de la UE: ¿voluntario u obligatorio?55:00 - Meta y el rechazo europeo: modelos frontera y choques regulatorios59:00 - ¿Dónde emprender con IA: Europa o fuera?1:03:00 - Cierre: entre el entusiasmo y la incertidumbre global⁠Hosted on Mumbler.io

La inteligencia artificial sigue avanzando… pero no siempre como esperábamos. En este episodio de La Tertul-IA, Lu, Frankie y Corti charlan sobre el choque entre las expectativas infladas y la realidad operativa de la IA. Exploramos los desafíos que viven las empresas al integrar agentes autónomos en producción, la creciente presión regulatoria del AI Act europeo y sentencias clave en EE. UU. que podrían cambiar las reglas del juego. 

Este podcast es posible gracias a Santander:https://online.bancosantander.es/landings/cuentas/cuenta-autonomos/Bienvenidos a esta tertulia en directo con Jordi Romero, Bernat Farrero y César Miguelañez.Empezamos hablando con la montaña rusa de Windsurf, la compañía que pasó de facturar 0 a 80 M $ en doce meses, intentó venderse a OpenAI y acabó viviendo un “Acqui-hiring” express de Google por 2,4 B $ mientras Cognition compraba los restos y 250 empleados se quedaban en tierra de nadie, un culebrón que abre un debate crudo sobre si hoy vale más el talento que el ARR .Conversamos con Ilya que desgrana por qué Grok 4 ha doblado el récord del benchmark ARC‑AGI y, entre aplausos y pullas a Elon Musk, discutimos si es el modelo más potente o simplemente el menos alineado con los “guard‑rails” tradicionales . A partir de ahí saltamos a la “guerra de los navegadores con IA”: ARC patina y Perplexity presenta Comet, un agente que navega y hace clics por ti, mientras OpenAI contraataca integrando su propio navegador en ChatGPT; todo ello dispara la eterna pregunta de si el SEO tal y como lo conocemos está condenado.Hablamos Carla y Jia, cofundadores de Theker, para celebrar en directo la mayor ronda seed de la historia en España —21 M € liderados por Kibo, Kfund e Inditex— y explicar cómo sus robots “tipo ChatGPT” aprenden tareas industriales sobre la marcha, por qué patentan sus grippers y cómo la velocidad es su verdadero moat . Analizamos la subida de precios de Cursor, las fugas a Cloud Code o Copilot y la fragilidad de cualquier startup cuando el proveedor de modelos le cierra el grifo .Entre preguntas del público surgen dilemas sobre CFOs en etapa seed y la utilidad de los SDR en la era de los agentes.Sigue a los "tertulianos" en Twitter:• Bernat Farrero: @bernatfarrero• Jordi Romero: @jordiromero• César Migueláñez: @heycesrSOBRE ITNIG

Neste episódio do Product Guru's, Paulo Chiodi conversa com Guilherme Gonzalez, referência em Design Systems no Brasil, sobre como ele usou inteligência artificial para criar, do zero, um Design System real, escalável e pronto para uso — em apenas 5 dias.Você vai descobrir como ele estruturou prompts, integrou ferramentas como Figma, V0.dev e Cloud Code, e aplicou tokens exportados via plugin para gerar componentes funcionais. O episódio é uma verdadeira aula prática sobre como aliar IA ao design de produto de forma estratégica e eficiente, com reflexões importantes sobre maturidade no design, a relação entre designers e tecnologia, e como construir frameworks aplicáveis à realidade brasileira.// Onde encontrar o convidado: Guilherme Gonzalez | Especialista em Design + AIhttps://www.linkedin.com/in/guigonzalez/// Conteúdos mencionados:https://futurism.com/ai-agents-failing-industryhttps://futurism.com/professors-company-ai-agentshttps://v0.dev/chat/ai-ds-experience-J9gcGFoFwCphttps://www.untitledui.com/https://guigonzalez.github.io/design-system/https://github.com/guigonzalez/design-systemhttps://dareframework.com.br/https://www.linkedin.com/pulse/como-trabalhar-com-vibe-coding-sem-virar-ref%25C3%25A9m-da-ia-gui-gonzalez-yzvhf/// Recado Importante: O futuro dos produtos digitais já começou e a Inteligência Artificial é parte do time.A PM3 acaba de lançar a Formação em Gestão de Produtos de IA: um curso pensado para Product Managers que querem criar, delegar e inovar com mais inteligência. Muito além dos prompts: você vai aprender a liderar produtos baseados em IA, dominar temas como Machine Learning, Deep Learning e IA Generativa, e aplicar novas formas de discovery, experimentação e validação.Prepare-se para o mercado que mais cresce no mundo e torne-se o PM que lidera a transformação.Acesse o link e saiba mais: https://go.pm3.com.br/ProductGurus-AI-Specialist/// Outros parceiros:Codando sem Codar - A maior comunidade de AI (Vibe) Coding do Brasilhttps://codandosemcodar.com.br/?utm_campaign=pg_podcastCurling - Do treinamento à criação de soluções com IA, estamos em cada etapa. https://www.usecurling.com/// Nesse episódio abordamos:​ Guilherme Gonzales criou um design system funcional com IA em apenas 5 dias.​ A IA foi usada como parceira de criação, não apenas como ferramenta.​ Prompts bem estruturados foram cruciais para resultados de qualidade.​ Tokens exportados do Figma foram integrados ao sistema com precisão.​ Usar IA exige clareza, contexto e paciência, não é só copiar prompt.​ Guilherme desenvolveu um framework próprio (DER) para uso de IA no design.​ O designer precisa ser mais estratégico e assumir protagonismo.​ Design System não é só do designer, é produto, código e cultura./// Capítulos00:00 Introdução00:46 Apresentação de Guilherme Gonzalez04:47 Primeiros testes com IA e design system06:00 Falhas na primeira versão e aprendizados08:20 Criando botão com boas práticas e tokens10:30 Como a IA interpretou os componentes12:27 Integração com Figma e exportação via plugin14:32 Exportando tokens do Figma para JSON17:40 O papel da IA como parceira de trabalho19:00 Processo de estruturar tokens e prompts22:48 Como a experiência influenciou no sucesso24:48 A diferença entre biblioteca e design system real26:09 IA e velocidade vs. qualidade no design29:16 Testando limites da IA em projeto pessoal33:12 Estrutura de prompts inteligentes36:00 As 10 lições sobre uso consciente da IA38:54 Aplicando IA com documentação e Storybook44:00 Testando JSON de 14 mil linhas no Cloud Code48:15 Aplicação prática de IA no dia a dia do designer56:33 Aprendizados ao vivo e erros com IA58:42 Apresentação do framework DER para IA no design01:03:51 Design System como produto real nas empresas01:08:48 Maturidade do design e das empresas no Brasil01:14:00 Limitações de cursos e o papel do contexto01:19:12 A real valorização do design nas empresas

In this episode of The Tech Trek, Amir sits down with Matt Moore, CTO and co-founder of Chainguard, to explore the escalating importance of software supply chain security. From Chainguard's origin story at Google to the systemic risks enterprises face when consuming open source, Matt shares the lessons, best practices, and technical innovations that help make open source software safer and more reliable. The conversation also touches on AI's impact on the attack surface, mitigating threats with engineering rigor, and why avoiding long-lived credentials could be your best defense.

Mayo de 2025 será recordado como el mes en que la inteligencia artificial se volvió inabarcable. En esta tertulia especial, Lu recibe de nuevo a Javi Santos, AI Hacker y experimentador incansable, para desgranar una avalancha de anuncios: OpenAI, Google, Meta, Anthropic y muchas startups han sacado todo su arsenal

Full show notes, transcript and AI chatbot - https://bit.ly/42SW2yGWatch on YouTube - https://youtu.be/08gLvHthNuw00:00:00 – Introduction 00:05:37 – ChatGPT shopping and e-commerce 00:08:50 – AI impact on data tracking 00:12:28 – Natural language data analysis tool 00:16:18 – Discovery Catalog for Cloud Storage 00:19:11 – Data accessibility in BigQuery 00:25:23 – AI-generated insights in GA4 00:27:56 – Cloud Code features in IDE 00:30:36 – Data robustness in BigQuery 00:35:45 – Data layer responsibilities in marketing 00:39:47 – Importance of monitoring data integrity 00:42:51 – Documentation and data governance 00:46:56 – Derelict data warehouse issues 00:49:12 – Data governance and documentation 00:54:26 – Principles of least privilege 00:58:03 – Infrastructure as code with Terraform 01:00:04 – Audience participation ideas-----Episode Summary:In this episode of The Measure Pod, Dara and Matt dive into a whole mix of marketing analytics news and musings. From new BigQuery features and GA4 updates to ChatGPT integrations and data quality best practices, they cover what's actually worth knowing. It's a loose but lively chat full of useful insights, personal takes, and a healthy dose of “pinch of salt” news. Expect talk of snooker, rogue GA4 permissions, and why naming conventions are the hill we'll die on.-----About The Measure Pod:The Measure Pod is your go-to fortnightly podcast hosted by seasoned analytics pros. Join Dara Fitzgerald (Co-Founder at Measurelab) & Matthew Hooson (Head of Engineering at Measurelab) as they dive into the world of data, analytics and measurement—with a side of fun.-----If you liked this episode, don't forget to subscribe to The Measure Pod on your favourite podcast platform and leave us a review. Let's make sense of the analytics industry together!The post #120 Marketing analytics updates and a BigQuery health check appeared first on Measurelab.

Full show notes, transcript and AI chatbot - https://bit.ly/42SW2yG Watch on YouTube - https://youtu.be/08gLvHthNuw 00:00:00 – Introduction 00:05:37 – ChatGPT shopping and e-commerce 00:08:50 – AI impact on data tracking 00:12:28 – Natural language data analysis tool 00:16:18 – Discovery Catalog for Cloud Storage 00:19:11 – Data accessibility in BigQuery 00:25:23 – AI-generated insights in GA4 00:27:56 – Cloud Code features in IDE 00:30:36 – Data robustness in BigQuery 00:35:45 – Data layer responsibilities in marketing 00:39:47 – Importance of monitoring data integrity 00:42:51 – Documentation and data governance 00:46:56 – Derelict data warehouse issues 00:49:12 – Data governance and documentation 00:54:26 – Principles of least privilege 00:58:03 – Infrastructure as code with Terraform 01:00:04 – Audience participation ideas ----- Episode Summary: In this episode of The Measure Pod, Dara and Matt dive into a whole mix of marketing analytics news and musings. From new BigQuery features and GA4 updates to ChatGPT integrations and data quality best practices, they cover what's actually worth knowing. It's a loose but lively chat full of useful insights, personal takes, and a healthy dose of “pinch of salt” news. Expect talk of snooker, rogue GA4 permissions, and why naming conventions are the hill we'll die on. ----- About The Measure Pod: The Measure Pod is your go-to fortnightly podcast hosted by seasoned analytics pros. Join Dara Fitzgerald (Co-Founder at Measurelab) & Matthew Hooson (Head of Engineering at Measurelab) as they dive into the world of data, analytics and measurement—with a side of fun. ----- If you liked this episode, don't forget to subscribe to The Measure Pod on your favourite podcast platform and leave us a review. Let's make sense of the analytics industry together! The post #120 Marketing analytics updates and a BigQuery health check appeared first on Measurelab.

Питання AGI давно обговорюється, але чи дійсно ми наблизилися до створення штучного інтелекту, здатного самостійно мислити? У цьому випуску наші ведучі Павло Дмитрієв, Михайло Гірняк та Євген Москвіта аналізують розвиток AI і його вплив на суспільство, бізнес і технології. Тема об'єктивності AI залишається відкритою: чи можливо створити алгоритми, вільні від людських упереджень? Також говоримо про те, як компанії інтегрують AI у свої продукти та чи справді Apple Intelligence відкриває нову еру технологій. Важливу увагу приділено автоматизації повсякденних завдань, голосовим помічникам та майбутньому AI-рішень у бізнесі.00:24 — моделювання людського інтелекту через LLM07:04 — сучасні моделі штучного інтелекту14:20 — особливості нових моделей 23:24 — Cloud Code: нові можливості для розробників34:15 — проблеми зі штучним інтелектом40:14 — Apple Intelligence та його обмеження44:05 — генерація інфографіки за допомогою AI46:50 — розвиток штучного інтелекту та інтеграція

Dans cet épisode, Antonio, Emmanuel et Guillaume reviennent sur les nouveautés et annonces faites à Google I/O 2023 : de nouveaux téléphones Pixel qui se plient ou pas, et surtout de l'intelligence artificielle du sol au plafond ! Que ce soit dans Android, dans Google Workspace, dans Google Cloud, une tonne de produits passe en mode survitaminé à l'IA. Guillaume, Antonio et Emmanuel discutent aussi de l'impact qu'ils voient sur l'AI, et de comment les Large Language Models sont raffinés et pourquoi on les fait halluciner, de subtilités du langage des signes. Enregistré le 23 mai 2023 Téléchargement de l'épisode LesCastCodeurs-Episode-296.mp3 Google I/O 2023 Site web : https://io.google/2023/ Keynote principale : https://io.google/2023/program/396cd2d5-9fe1-4725-a3dc-c01bb2e2f38a/ Keynote développeur : https://io.google/2023/program/9fe491dd-cadc-4e03-b084-f75e695993ea/ Vidéo résumée en 10 minutes de toutes les annonces : https://www.youtube.com/watch?v=QpBTM0GO6xI&list=TLGGCy91ScdjTPYxNjA1MjAyMw Vidéo de toutes les sessions techniques : https://io.google/2023/program/?q=technical-session Google I/O s'est tenu il y a 10 jours en Californie, dans l'amphithéâtre de Shoreline, près du campus de Google. Seulement 2000 personnes sur place, un chat et un jeu en ligne pour assister à distance. Jeu en ligne I/O Flip créé avec Flutter, Dart, Firebase, et Cloud Run, et tous les assets graphiques générés par Generative AI https://blog.google/technology/ai/google-card-game-io-flip-ai/ Des Pixels plein les yeux ! Des détails sur le design des nouveaux appareils : https://blog.google/products/pixel/google-pixel-fold-tablet-7a-design/ Pixel Fold Article : https://blog.google/products/pixel/google-pixel-fold/ Premier téléphone foldable de Google (après Samsung et Oppo) Un écran sur le dessus, et un grand écran pliable à l'intérieur Pratique pour la traduction où peut voir une discussion traduire en deux langues d'un côté sur un écran et dans l'autre langue sur l'autre Utilisation créative de la pliure : mode “laptop”, pour les selfies, pour poser l'appareil pour des photos de nuit Par contre… pas disponible en France, et tout de même presque 1900€ ! Pixel Tablet Article : https://blog.google/products/pixel/google-pixel-tablet/ Une belle tablette de 11 pouces, avec un dock de recharge avec enceinte intégrée Processeur Tensor G2, Chromecast intégré C'est un peu comme le Google Nest Hub Max mais avec un écran détachable Une coque pratique avec un trépied intégré et qui n'empêche pas de recharger la tablette sur le dock En mode dock, c'est comme l'écran du Google Home App, et dès qu'on la décroche, on est en mode multi-utilisateur, chacun avec son profil Pixel 7a Article : https://blog.google/products/pixel/pixel-7a-io-2023/ Écran de 6 pouces Triple appareil photo (grand angle, principal, et photo avant pour les selfies) 509 euros Magic Eraser pour effacer les trucs qu'on veut pas dans la photo, Magic Unblur pour rendre une photo floue plus nette, Real Tone pour rendre les peaux foncées plus naturelles Android Article quoi de neuf dans Android : https://blog.google/products/android/android-updates-io-2023/ Dans Messages, Magic Compose dans les conversations, l'IA nous aide à concevoir nos messages, dans différents styles (plus pro, plus fun, dans le style de Shakespeare) Android 14 devrait arriver un peu plus tard dans l'année, avec plus de possibilités de customisation (fond d'écran généré par Gen AI, fond d'écran Emojis, couleurs associées, fond d'écran 3D issus de ses photos) https://blog.google/products/android/new-android-features-generative-ai/ StudioBot : un chatbot intégré à Android Studio pour aider au développement d'applis Android https://io.google/2023/program/d94e89c5-1efa-4ab2-a13a-d61c5eb4e49c/ 800 millions d'utilisateurs sont passés à RCS pour le messaging Adaptation de 50 applications Android pour s'adapter aux foldables https://blog.google/products/android/android-app-redesign-tablet-foldable/ Wear OS 4 va rajouter le backup restore quand on change de montre et autres nouveautés https://blog.google/products/wear-os/wear-os-update-google-io-2023/ 800 chaînes TV gratuites dans Google TV sur Android et dans la voiture Android Auto va être disponible de 200 millions de voitures https://blog.google/products/android/android-auto-new-features-google-io-2023/ Waze disponible globalement sur le playstore dans toutes les voitures avec Android Auto Google Maps Article : https://blog.google/products/maps/google-maps-updates-io-2023/ Maps propose 20 milliards de km de direction tous les jours Immersive View for Routes 15 villes : Amsterdam, Berlin, Dublin, Florence, Las Vegas, London, Los Angeles, Miami, New York, Paris, San Francisco, San Jose, Seattle, Tokyo et Venice Possibilité pour les développeurs de s'intégrer et rajouter des augmentations 3D, des marqueurs Google Photos Article Magic Editor : https://blog.google/products/photos/google-photos-magic-editor-pixel-io-2023/ Magic Editor survitaminé à l'IA pour améliorer les photos, en déplaçant des gens, en rajoutant des parties coupées, ou bien rendre le ciel plus beau Possible que ce soit limité aux téléphones Pixel au début Projets expérimentaux Project Starline (écran avec caméra 3D qui donne un rendu 3D de son interlocuteur comme s'il était en face de soi) a été amélioré pour prendre moins de place https://blog.google/technology/research/project-starline-prototype/ Universal Translator : une nouvelle expérimentation pour faire du doublage et traduction automatique avec synchronisation des mouvements des lèvres Project Tailwind, une sorte de notebook dans lequel on peut rajouter tous ses documents à partir de drive, et poser des questions sur leur contenu, proposer des résumés, de faire du brainstorming sur ces thèmes https://thoughtful.sandbox.google.com/about MusicLM : un large language model pour générer de la musique à partir d'un texte de prompt (waitlist pour s'inscrire) https://blog.google/technology/ai/musiclm-google-ai-test-kitchen/ Project Gameface : utilisation des expressions du visage pour commander une souris et un ordinateur, pour les personnes qui ont perdu leur mobilité https://blog.google/technology/ai/google-project-gameface/ VisualBlocks : pour expérimenter dans une interface drag'n drop avec le développement de modèles pour Tensorflow lite et js https://visualblocks.withgoogle.com/ MakerStudio : pour les bidouilleurs et développeurs https://makersuite.google.com/ https://developers.googleblog.com/2023/05/palm-api-and-makersuite-moving-into-public-preview.html Search Labs Article : https://blog.google/products/search/generative-ai-search/ Expérimentations pour rajouter l'IA générative dans Google Search Faire des recherches avec des requêtes avec des phrases plus complexes, en intégrant des réponses comme Bard, avec des liens, des suggestions d'autres recherches associées Mais aussi proposer des publicités mieux ciblées On peut s'inscrire à Search Labs pour tester cette nouvelle expérience, mais au début juste en Anglais et juste pour les US Des intégrations avec Google Shopping pour proposer et filtrer des produits qui correspondent à la requête Recherche à l'aide d'image, avec Google Lens : 12 milliards de recherches visuelles par mois Palm et Bard Annonce du modèle LLM Palm 2 utilisé dans Bard et dans Google Cloud https://blog.google/technology/ai/google-palm-2-ai-large-language-model/ PaLM 2 est en cours d'intégration dans 25 produits de Google Supportera 100 langues différentes (pour l'instant seulement l'anglais, japonais et coréen), avec déjà les 40 langues les plus parlées d'ici la fin de l'année Maintenant disponible dans 180 pays… sauf l'Europe !!! Capacité de raisonnement accrue Peut coder dans une vingtaine de langages de programmation différents dont Groovy Différentes tailles de modèles : Gecko, Otter, Bison et Unicorn, mais le nombre de paramètres n'est pas communiquée, comme pour GPT-4 d'OpenAI Utilisable pour des requêtes et pour du chat Des modèles dérivées fine-tunés Med-PaLM 2 sur du savoir médical, sur l'analyse visuelle des radios et Sec-PaLM, entrainé sur des cas d'utilisation sur le thème de la cybersécurité, pour aider à déceler des scripts malicieux, des vecteurs d'attaque Sundar Pichai a aussi annoncé que Google travaillait déjà sur la prochaine évolution de ses LLM avec un modèle appelé Gemini. Peu de détails à part qu'il sera multimodal (en particulier recherche combinée image et texte par ex.) Partenariat et intégration de Adobe Firefly dans Bard pour générer des images https://blog.adobe.com/en/publish/2023/05/10/adobe-firefly-adobe-express-google-bard Duet AI pour Google Workspace Article : https://workspace.google.com/blog/product-announcements/duet-ai Dans Gmails et Docs, propose d'aider à la rédaction de vos emails et documents une extension de “smart compose” qui va permettre de générer des emails entiers, d'améliorer le style, de corriger la grammaire, éviter les répétitions de texte Dans Docs, des nouveaux “smart chips” pour rajouter des variables, des templates Dans Slides, rajouter des images générées par IA Des prompts dans Sheets pour générer un draft de table Dans Google Meet, possibilité de créer une image de fond customisée avec Generative AI Ces améliorations font parties de Workspace Labs auquel on peut s'inscrire dans la liste d'attente https://workspace.google.com/labs-sign-up/ Google Cloud Intégration de Generative AI partout https://cloud.google.com/blog/products/ai-machine-learning/google-cloud-launches-new-ai-models-opens-generative-ai-studio Nouvelles VM A3 avec les GPUs H100 de Nvidia, idéal pour l'entrainement de modèles de machine learning, avec 26 exaFlops de performance https://cloud.google.com/blog/products/compute/introducing-a3-supercomputers-with-nvidia-h100-gpus Trois nouveaux modèles LLM dans Vertex AI : Imagen (private preview) pour générer des images, Codey pour la génération de code, et Chirp pour la génération de la parole supportant 100 langues différentes avec 2 milliards de paramètres vocaux Model Garden : avec les modèles de machine learning y compris externes et open sources Ajout des embeddings pour le texte et l'image RLHF, Reinforcement Learning from Human Feedback bientôt intégrer pour étendre Vertex AI tuning et prompt design avec une boucle de feedback humaine Generative AI Studio pour tester ses prompts zero-shot, one-shot, multi-shots Duet AI pour Google Cloud https://cloud.google.com/blog/products/application-modernization/introducing-duet-ai-for-google-cloud Assistance de code dans VSCode et bientôt les IDEs JetBrains grâce au plugin Cloud Code, et dans Cloud Workstations. Intégration dans les IDEs d'un chat pour comme un compagnon pour discuter d'architecture, trouver les commandes à lancer pour son projet Le modèle de code de Codey fonctionne sur une vingtaine de languages de programmation, mais un modèle fine-tuné a été entrainé sur toute la doc de Google Cloud, donc pourra aider en particulier sur l'utilisation des APIs de Google Cloud, ou l'utilisation de la ligne de commande gcloud Duet AI est aussi dans App Sheet, la plateforme low/no-code, et permettra de chatter avec un chatbot pour générer une application App Sheet Quoi de neuf dans Firebase https://firebase.blog/posts/2023/05/whats-new-at-google-io Web Article : https://developers.googleblog.com/2023/05/io23-developer-keynote-recap.html Flutter 3 et Dart 3.10 https://io.google/2023/program/7a253260-3941-470b-8a4d-4253af000119/ WebAssembly https://io.google/2023/program/1d176349-7cf8-4b51-b816-a90fc9d7d479/ WebGPU https://io.google/2023/program/0da196f5-5169-43ff-91db-8762e2c424a2/ Baseline https://io.google/2023/program/528a223c-a3d6-46c5-84e4-88af2cf62670/ https://web.dev/baseline/ Nous contacter Pour réagir à cet épisode, venez discuter sur le groupe Google https://groups.google.com/group/lescastcodeurs Contactez-nous via twitter https://twitter.com/lescastcodeurs Faire un crowdcast ou une crowdquestion Soutenez Les Cast Codeurs sur Patreon https://www.patreon.com/LesCastCodeurs Tous les épisodes et toutes les infos sur https://lescastcodeurs.com/

Nate Avery, Outbound Product Manager at Google, joins Corey on Screaming in the Cloud to discuss what it's like working in the world of tech, including the implications of AI technology on the workforce and the importance of doing what you love. Nate explains why he feels human ingenuity is so important in the age of AI, as well as why he feels AI will make humans better at the things they do. Nate and Corey also discuss the changing landscape of tech and development jobs, and why it's important to help others throughout your career while doing something you love. About NateNate is an Outbound Product Manager at Google Cloud focused on our DevOps tools. Prior to this, Nate has 20 years of experience designing, planning, and implementing complex systems integrating custom-built and COTS applications. Throughout his career, he has managed diverse teams dedicated to meeting customer goals. With a background as a manager, engineer, Sys Admin, and DBA, Nate is currently working on ways to better build and use virtualized computer resources in both internal and external cloud environments. Nate was also named a Cisco Champion for Datacenter in 2015.Links Referenced: Google Cloud: https://cloud.google.com/devops Not Your Dad's IT: http://www.notyourdadsit.com/ Twitter: https://twitter.com/nathaniel_avery LinkedIn: https://www.linkedin.com/in/nathaniel-avery-2a43574/ TranscriptAnnouncer: Hello, and welcome to Screaming in the Cloud with your host, Chief Cloud Economist at The Duckbill Group, Corey Quinn. This weekly show features conversations with people doing interesting work in the world of cloud, thoughtful commentary on the state of the technical world, and ridiculous titles for which Corey refuses to apologize. This is Screaming in the Cloud.Corey: It's easy to **BEEP** up on AWS. Especially when you're managing your cloud environment on your own!Mission Cloud un **BEEP**s your apps and servers. Whatever you need in AWS, we can do it. Head to missioncloud.com for the AWS expertise you need. Corey: Welcome to Screaming in the Cloud. I'm Corey Quinn, and my guest today is Nate Avery, who's an outbound product manager over at Google Cloud. Nate, thank you for joining me.Nate: Thank you for having me. This is really a pretty high honor. I'm super thrilled to be here.Corey: One of my questions that I have about any large company when I start talking to them and getting to know people who work over there, pretty quickly emerges, which is, “What's the deal with your job title?” And it really doesn't matter what the person does, what the company is, there's always this strange nuance that tends to wind up creeping into the company. What is an outbound product manager and what is it you say it is you do here?Nate: Okay. That's an interesting question because I've been here for about a year now and I think I'm finally starting to figure it out. Sure, I should have known more when I applied for the job, [laugh] but there's what's on the paper and then there's what you do in reality. And so, what it appears to be, where I'm taking this thing now, is I talk to folks about our products and I try to figure out what it is they like, what it is they don't like, and then how do we make it better? I take that information back to our engineers, we huddle up, and we figure out what we can do, how to do it better, how to set the appropriate targets when it comes to our roadmaps. We look at others in the industry, where we are, where they are, where we think we can maybe have an advantage, and then we try to make it happen. That's really what it is.Corey: One of the strange things that happens at big companies, at least from my perspective, given that I've spent most of my career in small ones, is that everyone has a niche. There are very few people at large companies whose job description is yeah, I basically do everything. Where do you start? And where do you stop because Google Cloud, even bounding it to that business unit, is kind of enormous? You've [got 00:02:47] products that are outbound that you manage. And I feel like I should also call out that a product being outbound is not the same thing as being outgoing. I know that people are always wondering, what's Google going to turn off next, but Google Cloud mostly does the right thing in that respect. Good work.Nate: [laugh]. Nice. So, the products I focus on are the DevOps products. So, those are Cloud Build, Cloud Deploy, Artifact Registry, Artifact Analysis. I also work with some of our other dev tooling such as Cloud Workstations. That's in public preview right now, but maybe by the time this goes to air, it'll actually be in general availability.And then I also will talk about some of our other lesser-known tools like Skaffold or maybe on occasion, I'll throw out something about minikube. And also, Cloud Code, which is a really deep browser plugin for your IDE that gives you access to lots of different Google tools. So yeah, that's sort of my area.Corey: Well, I'm going to start with the last thing you mentioned, where you have Cloud Code as an IDE tooling and a plug-in for it. I'm relatively new to the world of IDEs because I come from the world of grumpy Unix admins; you never know what you're going to be remoting into next, but it's got VI on it, so worst case, you'll have that. So, I grew up using that, and as a result, that is still my default. I've been drifting toward VS Code a fair bit lately, as I've been regrettably learning JavaScript and TypeScript, just because having a lot of those niceties is great. But what's really transformative for me has been a lot of the generative AI offerings from various companies around hey, how about we just basically tab-complete your code for you, which is astonishing. I know people love to argue about that and then they go right back to their old approach of copying and pasting their code off a Stack Overflow.Nate: Yeah. That's an interesting one. When it works, it works and it's magical. And those are those experiences where you say, “I'm going to do this thing forever and ever I'm never going to go back.” And then when it doesn't work, you find yourself going back and then you maybe say, “Well, heck, that was horrible. Why'd I ever even go down this path?”I will say everyone's working on something along those lines. I don't think that that's much of a secret. And there are just so many different avenues at getting there. And I think that this is so early in the game that where we are today isn't where we're going to be.Corey: Oh, just—it's accelerating. Watching the innovation right now in the generative AI space is incredible. My light bulb moment that finally got me to start paying attention to this and viewing it as something other than hype that people are trying to sell us on conference stages was when I use one of them to spit out just, from a comment in VS Code, “Write a Python script that will connect to AWS pricing API and tell me what something costs, sorted from most to least expensive regions.” Because doing that manually would have taken a couple hours because their data structures are a sad joke and that API is garbage. And it sat and spun for a second and then it did it. But if I tell that story as, “This is the transformative moment that opened my eyes,” I sound incredibly sad and pathetic.Nate: No, I don't think so. I think that what it does, is it… one, it will open up more eyes, but the other thing that it does is you have to take that to the next level, which is great. That's great work, gone. Now that I have this information, what do I do with it? That's really where we need to be going and where we need to think about what this AI revolution is going to allow us to do, and that's to actually put this stuff into context.That's what humans do, which the computers are not always great at. And so, for instance, I see a lot of posts online about, “Hey, you know, I used to do job X, where I wrote up all these things,” or, “I used to write a blog and now because of AI, my boss wants me to write, you know, five times the output.” And I'm thinking, “Well, maybe the thing that you're writing doesn't need to be written if it can be easily queried and generated on the fly.” You know? Maybe those blog posts just don't have that much value anymore. So, what is it that we really should concentrate on in order to help us do better stuff, to have a higher order of importance in the world? That's where I think a lot of this really will wind up going is… you know, just as people, we've got to be better. And this will help us get there.Corey: One area of nuance on this, though, is—you're right when I talked about this with some of my developer friends, some of their responses were basically to become immediately defensive. Like, “Sure, it's great for the easy stuff, but it's not going to solve the high-level stuff that senior engineers are good at.” And I get that. This ridiculous thing that I had to do is not a threat to a senior engineer, but it is arguably a threat to someone I find on Upwork or Fiverr or whatnot to go and write this simple script for me.Nate: Oh yeah.Corey: Now, the concern that I have is one of approachability and accessibility because. Senior engineers don't form fully created from the forehead of some God somewhere that emerges from Google. They start off as simply people who have no idea what they're doing and have a burning curiosity about something, in many cases. Where is the next generation going to get the experience of writing a lot of that the small-scale stuff, if it's done for them? And I know that sounds alarmist, and oh, no, the sky is falling, and are the children going to be all right, as most people my age start to get into. But I do wonder what the future holds.Nate: That's legit. That's a totally legit question because it's always kind of hanging out there. I look at what my kids have access to today. They have freaking Oracle, the Oracle at Delphi on their phone; you know, and—Corey: If Oracle the database on their phone, I would hate to imagine what the cost of raising your kids to adulthood would be.Nate: Oh, it's mighty, mighty high [laugh]. But no, they have all of this stuff at their hands and then even just in the air, right? There's ambient computing, there's any question you want answered, you could speak it into the air and it'll come out. And it'll be, let's just say, I don't know, at least 85% accurate. But my kids still ask me [laugh].Corey: Having my kids, who are relatively young, still argue and exhaust their patience on a robot with infinite patience instead of me who has no patience? Transformative. “How do I spell whatever it is?” “Ask Alexa,” becomes a story instead of, “Look it up in the dictionary,” like my parents used to tell me. It's, “If I knew how to spell it, I would need to look it up in the dictionary, but I don't, so I can't.”Nate: Right. And I would never need to spell it again because I have the AI write my whole thing for me.Corey: That is a bit of concern for me when—some of the high school teachers are freaking out about students are writing essays with this thing. And, yeah, on the one hand, I absolutely see this as alarmism, where, oh, no, I'm going to have to do my job, on some level. But the reason you write so many of those boring, pointless essays in English class over the course of the K through 12 experience is ideally, it's teaching you how to frame your discussions, how to frame an argument, how to tell a compelling story. And, frankly, I think that's something that a lot of folks in the engineering cycle struggle with mightily. You're a product slash program manager at this point; I sort of assume that I don't need to explain to you that engineers are sometimes really bad at explaining what they mean.Nate: Yeah. Dude, I came up in tech. I'm… bad at it too sometimes [laugh]. Or when I think I'm doing a great job and then I look over and I see a… you know, the little blanky, blanky face, it goes, “Oh. Oh, hold on. I'll recalibrate that for you.” It's a thing.Corey: It's such a bad trope that they have now decided that they are calling describing what you actually mean slash want is now an entire field called prompt engineering.Nate: Dude, I hate that. I don't understand how this is going to be a job. It seems to be the most ridiculous thing in the world. If you say, “I sit down for six hours a day and I ask my computer questions,” I got to ask, “Well, why?” [laugh]. You know? And really, that's the thing. It gets back—Corey: Well, most of us do that all day long. It's just in Microsoft Excel or they use SQL to do it.Nate: Yeah… it is, but you don't spend your day asking the question of your computer, “Why.” Or really, most of us ask the question, “How?” That's really what it is we're doing.Corey: Yeah. And that is where I think it's going to start being problematic for some folks who are like, “Well, what is the point of writing blog posts if Chat-GIPITY can do it?” And yes, that's how I pronounce it: Chat-GIPITY. And the response is, “Look, if you're just going to rehash the documentation, you're right. There's no point in doing it.”Don't tell me how to do something. Tell me why. Tell me when I should consider using this tool, that tool, why this is interesting to me, why it exists. Because the how, one way or another, there are a myriad ways to find out the answer to something, but you've got to care first and convincing people to care is something computers still have not figured out.Nate: Bingo. And that gets back to your question about the engineers, right? Yeah. Okay. So sure, the little low-level tasks of, “Hey I need you to write this API.” All right, so maybe that stuff does get farmed out.However, the overall architecture still has to be considered by someone, someone still has to figure out where and how, and when things should be placed and the order in which these things should be connected. That never really goes away. And then there's also the act of creation. And by creation, I mean, just new ideas, things that—you know, that stroke of creativity and brilliance where you just say, “Man, I think there's a better way to do this thing.” Until I see that from one of these generative AI products, I don't know if anyone should truly feel threatened.Corey: I would argue that people shouldn't necessarily feel threatened regardless because things always change; that's the nature of it. I saw a headline on Hacker News recently where it said that 90% of my skills are worthless, but 10% of them are 10x what they were was worth. And I think that there's a lot of truth to that because it's, if you want a job where you never have to—you don't have to keep up with the continuing field, there are options. Not to besmirch them, but accountants are a terrific example of this. Yes, there's change to accountancy rules, but it happens slowly and methodically. You don't go on vacation for two years as an accountant—or a sabbatical—come back and discover that everything's different and math doesn't work the way it once did. Computers on the other hand, it really does feel like it's keep up or you never will.Nate: Unless you're a COBOL guy and you get called back for y2k.Corey: Oh, of course. And I'm sure—and now you're sitting around, you're waiting because when the epic time problem hits in 2038, you're going to get your next call out. And until then, it's kind of a sad life. You're the Maytag repair person.Nate: Yeah. I'm bad at humor, by the way, in case you have noticed. So, you touched on something there about the rate of change and how things change and whether or not these generative AI models are going to be able to—you know, just how far can they go? And I think that there's a—something happened over the last week or so that really got me thinking about this. There was a posting of a fake AI-generated song, I think from Drake.And say what you want about cultural appropriation, all that sort of thing, and how horrible that is, what struck me was the idea that these sorts of endeavors can only go so far because in any genre where there's language, and current language that morphs and changes and has subtlety to it, the generative AI would have to somehow be able to mimic that. And not to say that it could never get there, but again, I see us having some situations where folks are worried about a lot of things that they don't need to worry about, you know, just at this moment.Corey: I'm curious to figure out what your take is on how you see the larger industry because for a long time—and yes, it's starting to fade on some level, because it's not 2006 anymore, but there was a lot of hero worship going on with respect to Google, in particular. It was the mythical city on the hill where all the smart people went and people's entire college education was centered around the idea of, well, I'm going to get a job at Google when I graduate or I'm doomed. And it never seems to work out that way. I feel like there's a much more broad awareness these days that there's no one magical company that has the answers and there are a lot of different paths. But if you're giving guidance to someone who's starting down that path today, what would it be?Nate: Do what you love. Find something that you love, figure out who does the thing that you love, and go there. Or go to a place that does a thing that you love poorly. Go there. See if you can make a difference. But either way, you're working on something that you like to do.And really, in this business, if you can't get in the door at one of those places, then you can make your own door. It's becoming easier and easier to just sort of shoehorn yourself into this space. And a lot of it, yeah, there's got to be talent, yeah, you got to believe in yourself, all that sort of thing, but the barriers to entry are really low right now. It's super easy to start up a website, it costs you nothing to have a GitHub account. I really find it surprising when I talked to my younger cousins or someone else in that age range and they start asking, like, “Well, hey, how do I get into business?”And I'm like, “Well, what's your portfolio?” You know? And I ask them, “Do you want to work for someone else? Or would you like to at least try working for yourself first?” There are so many different avenues open to folks that you're right, you don't have to go to company X or you will never be anything anymore. That said, I am at [laugh] one of the bigger companies and do there are some brilliant people here. I bump into them and it's kind of wild. It really, really is.Corey: Oh, I want to be very clear, despite the shade that I throw at Google—and contemporary peers in the big tech company space—there are an awful lot of people who are freaking brilliant. And more importantly, by far, a lot of people who are extraordinarily kind.Nate: Yeah. Yeah. So, all right, in this business, there's that whole trope about, “Yeah, they're super smart, but they're such jerks.” It doesn't have to be that way. It really doesn't. And it's neat when you run into a place that has thousands of people who do not fit that horrible stereotype out there of the geek who can't, you know, who can't get along well with others. It's kind of nice.But I also think that that's because the industry itself is opening up. I go on to Twitter now and I see so many new faces and I see folks coming in, you know, for whatever reason, they're attracted to it for reasons, but they're in. And that's the really neat part of it. I used to worry that I didn't see a lot of young people being interested in this space. But I'm starting to notice it now and I think that we're going to wind up being in good hands.Corey: The kids are all right, I think, is a good way of framing it. What made you decide to go to Google? Again, you said you've been there about a year at this point. And, on some level, there's always a sense in hindsight of, well, yeah, obviously someone went from this job to that job to that job. There's a narrative here and it makes sense, but I've never once in my life found that it made sense and was clear while you're making the decision. It feels like it only becomes clear in hindsight.Nate: Yes, I am an extremely lucky person. I am super fortunate, and I will tell a lot of people, sometimes I have the ability to fall ass-backwards into success. And in this case, I am here because I was asked. I was asked and I didn't really think that I was the Google type because, I don't know what I thought the Google type was, just, you know, not me.And yet, I… talked it out with some folks, a really good, good buddy of mine and [laugh] I'll be darned, you know, next thing, you know, I'm here. So, gosh, what can I say except, don't limit yourself [laugh]. We do have a tendency to do that and oh, my God, it's great to have a champion and what I'd like to do now, now that you mention it and it's been something that I had on my mind for a bit is, I've got to figure out how to, you know how to start, you know, giving back, paying it forward, whatever the phrase it is you want to use? Because—Corey: I like, “Send the elevator back down.”Nate: Send the elevator back down? There you go, right? If that escalator stopped, turn it back on.Corey: Yeah, escalator; temporarily, stairs.Nate: Yes. You know, there are tons of ways up. But you know, if you can help someone, just go ahead and do it. You'd be surprised what a little bit of kindness can do.Corey: Well, let's tie this back to your day job for a bit, on some level. You're working on, effectively, developer tools. Who's the developer?Nate: Who's the developer? So, there's a general sense in the industry that anyone who works in IT or anyone who writes code is a developer. Sometimes there's the very blanket statement out there. I tend to take the view that a developer is the person who writes the code. That is a developer, that's [unintelligible 00:21:52] their job title. That's the thing that they do.The folks who assist developers, the folks who keep the servers up and running, they're going to have a lot of different names. They're DevOps admins, they're platform admins, they're server admins. Whatever they are, rarely would I call them developers, necessarily. So, I get it. We try to make blanket statement, we try to talk to large groups at a time, but you wouldn't go into your local county hospital and say that, “I want to talk to the dentist,” when you really mean, like, a heart surgeon.So, let's not do that, you know? We're known for our level of specificity when we discuss things in this field, so let's try to be a little more specific when we talk about the folks who do what they do. Because I came up on that ops track and I know the type of effort that I put in, and I looked at folks across from me and I know the kind of hours that they put in, I know all of the blood sweat and tears and nightless sleeps and answering the pagers at four in the morning. So, let's just call them what they are, [laugh] right? And it's not to say that calling them a developer is an insult in any way, but it's not a flex either.Corey: You do work at a large cloud company, so I have to assume that this is a revelation for you, but did you know that words actually mean things? I know, it's true. You wouldn't know it from a lot of the product names that wind up getting scattered throughout the world. The trophy for the worst one ever though, is Azure DevOps because someone I was talking to as a hiring manager once thought that they listed that is a thing they did on their resume and was about to can the resume. It's, “Wow, when your product name is so bad that it impacts other people's careers, that's kind of impressively awful.”But I have found that back when the DevOps movement was getting started, I felt a little offput because I was an operations person; I was a systems administrator. And suddenly, people were asking me about being a developer and what it's like. And honestly, on some level, I felt like an imposter, just because I write configuration files; I don't write code. That's very different. Code is something smart people write and I'm bad at doing that stuff.And in the fullness of time, I'm still bad at it, but at least now unenthusiastically bad at it. And, on some level, brute force also becomes a viable path forward. But it felt like it was gatekeeping, on some level, and I've always felt like the terms people use to describe what I did weren't aimed at me. I just was sort of against the edge.Nate: Yeah. And it's a weird thing that happens around here, how we get to these points, or… or somehow there's an article that gets written and then all of a sudden, everyone's life is changed in an industry. You go from your job being, “Hey, can you rack and stack the server?” To, “Hey, I need you to write this YAML code that's going to virtually instantiate a server and also connect it to a load balancer, and we need these done globally.” It's a really weird transition that happens in life.But like you said, that's part of our job: it morphs, it changes, it grows. And that's the fun of it. We hope that these changes are actually for the better and then they're going to make us more productive and they're going to make our businesses thrive and do things that they couldn't be before, like maybe be more resilient. You know, you look at the number of customers—customers; I think of them as customers—who had issues because of that horrible day in 9/11 and, you know, their business goes down the tube because there wasn't an adequate DR or COOP strategy, you know? And I know, I'm going way back in the wayback, but it's real. And I knew people who were affected by it.Corey: It is. And the tide is rising. This gets back to what we were talking about where the things that got you here won't necessarily get you there. And Cloud is a huge part of that. These days, I don't need to think about load balancers, in many cases, or all of the other infrastructure pieces because Google Cloud—among other companies, as well, lots of them—have moved significantly up the stack.I mean, people are excited about Kubernetes in a whole bunch of ways, but what an awful lot of enterprises are super excited about is suddenly, a hard drive failure doesn't mean their application goes down.Nate: [Isn't that 00:26:24] kind of awesome?Corey: Like, that's a transformative moment for them.Nate: It totally is. You know, I get here and I look at the things that people are doing and I kind of go, “Wow,” right? I'm in awe. And to be able to contribute to that in some way by saying, “Hey, you know what, we'll be cool? How about we try this feature?” Is really weird, [laugh] right?It's like, “Wow, they listened to me.” But we think about what it is we're trying to do and a lot of it, strangely enough, is not just helping people, but helping people by getting out of the way. And that is huge, right? You know, because you just want it to work, but more than it just working, you want it to be seamless. What's easier than putting your key in the ignition and turning it? Well, not having to use a key at all.So, what are those types of changes that we can bring to these different types of experiences that folks have? If you want to get your application onto a Kubernetes cluster, it shouldn't be some Herculean feat.Corey: And running that application responsibly should not require a team of people, each making a quarter million bucks a year, just to be able to do it safely and responsibly. There's going to be a collapsing down of what you have to know in order to run these things. I mean, web servers used to be something that required a month of your life and a fair bit of attention to run. Now, it's a checkbox in a cloud console.Nate: Yeah. And that's what we're trying to get it to, right? Why isn't everything a checkbox? Why can't you say, “Look, I wrote my app. I did the hard part.” Let's—you know, I just need to see it go somewhere. You know? Make it go and make it stay up. And how can I do that?And also, here's a feature that we're working on. Came out recently and we want folks to try it. It's a cloud deploy feature that works for Cloud Run as well as it does for GKE. And it's… I know it's going to sound super simple: it's our canary deployment method. But it's not just canary deployment, but also we can tie it into parallel deployment.And so, you can have your new version of your app stood up alongside your old version of the app and we can roll it out incrementally in parallel around the world and you can have an actual test that says, “Hey, is this working? Is it not working?” If it does, great, let's go forward. If it doesn't, let's roll back. And some of the stuff sounds like common sense, but it's been difficult to pull off.And now we're trying to do it with just a few lines a YAML. So, you know, is it as simple as it could be? Well, we're still looking at that. But the features are in there and we're constantly looking at what we can do to iterate and figure out what the next thing is.Corey: I really want to thank you for taking the time to speak with me. If people want to learn more, where's the best place for them to find you?Nate: Best place for them to find me used to be my blog, it's Not Your Dad's IT, However, I've been pretty negligent there since doing this whole Google thing, so I would say, just look me up on Twitter at @nathaniel_avery, look me up on Google. You can go to a pretty cool search engine and [laugh]—Corey: Oh, that's right. You guys have a search engine now. Good work.Nate: That's what I hear [laugh].Corey: Someday maybe it'll even come to Google Docs.Nate: [laugh]. Yes, so yeah, that's where to find me. You know, just look me up at Nathaniel Avery. I think that handle works for almost everything, Twitter, LinkedIn, wherever, and reach out.If there's something you like about our DevOps tools, let me know. If there's something you hate about our DevOps tools, definitely let me know. Because the only reason we're doing this is to try and help people. And if we're not doing that, then we need to know. We need to know why it isn't working out.And trust me, I talk to these engineers every day. That's the thing that really keeps them moving in the morning is knowing that they're doing something to make things better for folks. Real quick, I'll close out, and I think I may have mentioned this on some other podcasts. I come from the ops world. I was that guy who had to help get a deployment out on a Friday night and it lasted all weekend long and you're staring there at your phone at some absurd time on a Sunday night and everyone's huddled together and you're trying to figure out, are we going to rollback or are we going to go forward? What are we going to do by Monday?Corey: I don't miss those days.Nate: Oh, oh God no. I don't miss those days either. But you know what I do want? I took this job because I don't want anyone else to have those days. That's really what it is. We want to make sure that these tools give folks the ability to deploy safely and to deploy with confidence and to take that level of risk out of the equation, so that folks can, you know, just get back to doing other things. You know, spend that time with your family, spend the time reading, spend that time prompting ChatGPT with questions, [laugh] whatever it is you want to do, but you shouldn't have to sit there and wonder, “Oh, my God, is my app working? And what do I do when it doesn't?”Corey: I really want to thank you for being as generous with your time and philosophy on this. Thanks again. I've really enjoyed our conversation.Nate: Thank you. Thank you. I've been a big fan of your work for years.Corey: [laugh]. Nate Avery, outbound product manager at Google Cloud. I'm Cloud Economist Corey Quinn and this is Screaming in the Cloud. If you've enjoyed this podcast, please leave a five-star review on your podcast platform of choice whereas if you hate this podcast, please leave a five-star review on your podcast platform of choice along with an angry, insulting comment that you had Chat-GIPITY write for you in YAML.Corey: If your AWS bill keeps rising and your blood pressure is doing the same, then you need The Duckbill Group. We help companies fix their AWS bill by making it smaller and less horrifying. The Duckbill Group works for you, not AWS. We tailor recommendations to your business and we get to the point. Visit duckbillgroup.com to get started.