Podcasts about Lambda

Host Jason Blitman sits down with Seán Hewitt (Open, Heaven) to discuss sense memories, queer representation in school growing up, and Seán's aversion to musicals—despite offering a sharp insight into The Sound of Music's film adaptation. Later, Jason is joined by Guest Gay Reader Jeffery Self, who shares what he's currently reading, talks about his book Self Sabotage, and reflects on theatre icons Cathy Rigby, Sally Struthers, and Gary Beach.Seán Hewitt's debut collection of poetry, Tongues of Fire, won the Laurel Prize in 2021, and was shortlisted for The Sunday Times Young Writer of the Year Award, the John Pollard Foundation International Poetry Prize, and a Dalkey Literary Award. In 2020, he was chosen by The Sunday Times (London) as one of their “30 under 30”  artists in Ireland. His memoir, All Down Darkness Wide, is published by Jonathan Cape in the UK and Penguin Press in the United States (2022). It was shortlisted for Biography of the Year at the An Post Irish Book Awards, for the Foyles Book of the Year in nonfiction, for the RSL Ondaatje Prize, and for a LAMBDA award, and won the Rooney Prize for Irish Literature in 2022. Hewitt is assistant professor in literary practice at Trinity College Dublin, and is a fellow of the Royal Society of Literature.Jeffery Self is a writer and actor whose TV credits include Search Party, The Horror of Dolores Roach, Shameless, 30 Rock, Desperate Housewives, as well as co-creating and starring in the cult low-fi series Jeffery & Cole Casserole with Cole Escola. His film credits include Drop, Spoiler Alert, Mack and Rita, and The High Note. He is the author of the young adult novels Drag Teen and A Very, Very Bad Thing. He lives in New York City.SUBSTACK!https://gaysreading.substack.com/ BOOK CLUB!Use code GAYSREADING at checkout to get first book for only $4 + free shipping! Restrictions apply.http://aardvarkbookclub.com WATCH!https://youtube.com/@gaysreading FOLLOW!Instagram: @gaysreading | @jasonblitmanBluesky: @gaysreading | @jasonblitmanCONTACT!hello@gaysreading.com

Aujourd'hui un peu partout autour du globe, les citoyens lambda ont l'impression de vivre dans Matrix ou la Guerre des Étoiles. Ils ont l'impression de se débattre dans et contre un système qui les malaxe, les avale, les digère et les recrache pour mieux recommencer encore et encore. 

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma convida Luciano Holanda e o Igor Gushiken do Quinto Andar para um papo sobre as mudanças que a tecnologia trouxe para esse mercado tão tradicional. Participantes:Fernando Okuma - ⁠@feokuma⁠Luciano Holanda - @lucianohgoIgor Gushiken - @igor-gushiken Pauta:O Mercado Imobiliário Antes da DigitalizaçãoO Impacto da Tecnologia e da IA no Mercado ImobiliárioO Papel do Quinto Andar na Transformação DigitalFeatures da plataforma do Quinto AndarDesafios e Oportunidades para o FuturoReferências:⁠https://www.quintoandar.com.br/QuintoAndar lança experiência inédita de precificação e busca de imóveis por IA generativa Edição:⁠Compasso Coolab⁠

Venus's extra-thick crust may be extra chewy, allowing convection to occur and helping power volcanoes into the current era. New observations of the distant universe, meanwhile, show that dark energy may not have behaved as expected in the standard cosmological model. We'll break it all down for you together with space news and trivia with your friendly neighborhood astroquarks.

SpaceTime Series 28 Episode 42The Astronomy, Space and Science News PodcastUnraveling Antimatter Mysteries, New Techniques to Detect Dark Matter, and Insights into the Spectrum Rocket FailureIn this episode of SpaceTime, we dive into groundbreaking discoveries at the Large Hadron Collider, where physicists have identified a significant difference in the decay behaviors of ordinary matter and antimatter. This finding could provide vital clues to understanding why our universe is dominated by matter despite the Big Bang's creation of equal amounts of both. We explore the implications of these results and how they align with the Standard Model of particle physics.Innovative Approaches to Dark Matter DetectionNext, we discuss an innovative new technique developed by researchers at the University of Queensland to detect dark matter using atomic clocks and cavity-stabilized lasers. This cutting-edge approach aims to uncover the elusive nature of dark matter, which constitutes about 80% of the universe yet remains largely a mystery. We examine how this method could lead to new insights into the distribution and properties of dark matter.Spectrum Rocket Launch Failure InvestigationAdditionally, we analyze the recent failure of the Spectrum rocket during its inaugural launch from Norway. Investigators are looking into the causes of the incident, which involved thrust vectoring oscillations leading to the rocket's loss of control. We discuss potential technical issues and what this means for future European orbital launches.00:00 Space Time Series 28 Episode 42 for broadcast on 7 April 202500:49 Discovery of decay differences between matter and antimatter06:30 Implications for understanding the universe's matter dominance12:15 New techniques for detecting dark matter18:00 Using atomic clocks for dark matter research22:45 Analysis of the Spectrum rocket failure27:00 Summary of recent scientific developments30:15 Science report: Southern Ocean warming impactswww.spacetimewithstuartgary.comwww.bitesz.com

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma convida Bruno Okamoto para um papo sobre o que é os Micro SaaS, seus desafios e como é a forma de pensar nesse modelo de negócios. Participantes:Fernando Okuma - @feokumaBruno Okamoto - @brunomicrosaas Pauta:O que é um Micro-SaaS?Por que criar um Micro-SaaS?Tecnologias e Ferramentas para ConstruçãoDesafios e Armadilhas ComunsMonetização e Modelos de ReceitaEscalando um Micro-SaaS: Crescer ou Manter Pequeno?O Futuro do Micro-SaaSReferências:Comunidade MicroSaaSBruno Okamoto - YouTubewww.indiehackers.com Edição:Compasso Coolab Créditos das músicas usadas neste programa:Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Brandon Liu is an open source developer and creator of the Protomaps basemap project. We talk about how static maps help developers build sites that last, the PMTiles file format, the role of OpenStreetMap, and his experience funding and running an open source project full time. Protomaps Protomaps PMTiles (File format used by Protomaps) Self-hosted slippy maps, for novices (like me) Why Deploy Protomaps on a CDN User examples Flickr Pinball Map Toilet Map Related projects OpenStreetMap (Dataset protomaps is based on) Mapzen (Former company that released details on what to display based on zoom levels) Mapbox GL JS (Mapbox developed source available map rendering library) MapLibre GL JS (Open source fork of Mapbox GL JS) Other links HTTP range requests (MDN) Hilbert curve Transcript You can help correct transcripts on GitHub. Intro [00:00:00] Jeremy: I'm talking to Brandon Liu. He's the creator of Protomaps, which is a way to easily create and host your own maps. Let's get into it. [00:00:09] Brandon: Hey, so thanks for having me on the podcast. So I'm Brandon. I work on an open source project called Protomaps. What it really is, is if you're a front end developer and you ever wanted to put maps on a website or on a mobile app, then Protomaps is sort of an open source solution for doing that that I hope is something that's way easier to use than, um, a lot of other open source projects. Why not just use Google Maps? [00:00:36] Jeremy: A lot of people are gonna be familiar with Google Maps. Why should they worry about whether something's open source? Why shouldn't they just go and use the Google maps API? [00:00:47] Brandon: So Google Maps is like an awesome thing it's an awesome product. Probably one of the best tech products ever right? And just to have a map that tells you what restaurants are open and something that I use like all the time especially like when you're traveling it has all that data. And the most amazing part is that it's free for consumers but it's not necessarily free for developers. Like if you wanted to embed that map onto your website or app, that usually has an API cost which still has a free tier and is affordable. But one motivation, one basic reason to use open source is if you have some project that doesn't really fit into that pricing model. You know like where you have to pay the cost of Google Maps, you have a side project, a nonprofit, that's one reason. But there's lots of other reasons related to flexibility or customization where you might want to use open source instead. Protomaps examples [00:01:49] Jeremy: Can you give some examples where people have used Protomaps and where that made sense for them? [00:01:56] Brandon: I follow a lot of the use cases and I also don't know about a lot of them because I don't have an API where I can track a hundred percent of the users. Some of them use the hosted version, but I would say most of them probably use it on their own infrastructure. One of the cool projects I've been seeing is called Toilet Map. And what toilet map is if you're in the UK and you want find a public restroom then it maps out, sort of crowdsourced all of the public restrooms. And that's important for like a lot of people if they have health issues, they need to find that information. And just a lot of different projects in the same vein. There's another one called Pinball Map which is sort of a hobby project to find all the pinball machines in the world. And they wanted to have a customized map that fit in with their theme of pinball. So these sorts of really cool indie projects are the ones I'm most excited about. Basemaps vs Overlays [00:02:57] Jeremy: And if we talk about, like the pinball map as an example, there's this concept of a basemap and then there's the things that you lay on top of it. What is a basemap and then is the pinball locations is that part of it or is that something separate? [00:03:12] Brandon: It's usually something separate. The example I usually use is if you go to a real estate site, like Zillow, you'll open up the map of Seattle and it has a bunch of pins showing all the houses, and then it has some information beneath it. That information beneath it is like labels telling, this neighborhood is Capitol Hill, or there is a park here. But all that information is common to a lot of use cases and it's not specific to real estate. So I think usually that's the distinction people use in the industry between like a base map versus your overlay. The overlay is like the data for your product or your company while the base map is something you could get from Google or from Protomaps or from Apple or from Mapbox that kind of thing. PMTiles for hosting the basemap and overlays [00:03:58] Jeremy: And so Protomaps in particular is responsible for the base map, and that information includes things like the streets and the locations of landmarks and things like that. Where is all that information coming from? [00:04:12] Brandon: So the base map information comes from a project called OpenStreetMap. And I would also, point out that for Protomaps as sort of an ecosystem. You can also put your overlay data into a format called PMTiles, which is sort of the core of what Protomaps is. So it can really do both. It can transform your data into the PMTiles format which you can host and you can also host the base map. So you kind of have both of those sides of the product in one solution. [00:04:43] Jeremy: And so when you say you have both are you saying that the PMTiles file can have, the base map in one file and then you would have the data you're laying on top in another file? Or what are you describing there? [00:04:57] Brandon: That's usually how I recommend to do it. Oftentimes there'll be sort of like, a really big basemap 'cause it has all of that data about like where the rivers are. Or while, if you want to put your map of toilets or park benches or pickleball courts on top, that's another file. But those are all just like assets you can move around like JSON or CSV files. Statically Hosted [00:05:19] Jeremy: And I think one of the things you mentioned was that your goal was to make Protomaps or the, the use of these PMTiles files easy to use. What does that look like for, for a developer? I wanna host a map. What do I actually need to, to put on my servers? [00:05:38] Brandon: So my usual pitch is that basically if you know how to use S3 or cloud storage, that you know how to deploy a map. And that, I think is the main sort of differentiation from most open source projects. Like a lot of them, they call themselves like, like some sort of self-hosted solution. But I've actually avoided using the term self-hosted because I think in most cases that implies a lot of complexity. Like you have to log into a Linux server or you have to use Kubernetes or some sort of Docker thing. What I really want to emphasize is the idea that, for Protomaps, it's self-hosted in the same way like CSS is self-hosted. So you don't really need a service from Amazon to host the JSON files or CSV files. It's really just a static file. [00:06:32] Jeremy: When you say static file that means you could use any static web host to host your HTML file, your JavaScript that actually renders the map. And then you have your PMTiles files, and you're not running a process or anything, you're just putting your files on a static file host. [00:06:50] Brandon: Right. So I think if you're a developer, you can also argue like a static file server is a server. It's you know, it's the cloud, it's just someone else's computer. It's really just nginx under the hood. But I think static storage is sort of special. If you look at things like static site generators, like Jekyll or Hugo, they're really popular because they're a commodity or like the storage is a commodity. And you can take your blog, make it a Jekyll blog, hosted on S3. One day, Amazon's like, we're charging three times as much so you can move it to a different cloud provider. And that's all vendor neutral. So I think that's really the special thing about static storage as a primitive on the web. Why running servers is a problem for resilience [00:07:36] Jeremy: Was there a prior experience you had? Like you've worked with maps for a very long time. Were there particular difficulties you had where you said I just gotta have something that can be statically hosted? [00:07:50] Brandon: That's sort of exactly why I got into this. I've been working sort of in and around the map space for over a decade, and Protomaps is really like me trying to solve the same problem I've had over and over again in the past, just like once and forever right? Because like once this problem is solved, like I don't need to deal with it again in the future. So I've worked at a couple of different companies before, mostly as a contractor, for like a humanitarian nonprofit for a design company doing things like, web applications to visualize climate change. Or for even like museums, like digital signage for museums. And oftentimes they had some sort of data visualization component, but always sort of the challenge of how to like, store and also distribute like that data was something that there wasn't really great open source solutions. So just for map data, that's really what motivated that design for Protomaps. [00:08:55] Jeremy: And in those, those projects in the past, were those things where you had to run your own server, run your own database, things like that? [00:09:04] Brandon: Yeah. And oftentimes we did, we would spin up an EC2 instance, for maybe one client and then we would have to host this server serving map data forever. Maybe the client goes away, or I guess it's good for business if you can sign some sort of like long-term support for that client saying, Hey, you know, like we're done with a project, but you can pay us to maintain the EC2 server for the next 10 years. And that's attractive. but it's also sort of a pain, because usually what happens is if people are given the choice, like a developer between like either I can manage the server on EC2 or on Rackspace or Hetzner or whatever, or I can go pay a SaaS to do it. In most cases, businesses will choose to pay the SaaS. So that's really like what creates a sort of lock-in is this preference for like, so I have this choice between like running the server or paying the SaaS. Like businesses will almost always go and pay the SaaS. [00:10:05] Jeremy: Yeah. And in this case, you either find some kind of free hosting or low-cost hosting just to host your files and you upload the files and then you're good from there. You don't need to maintain anything. [00:10:18] Brandon: Exactly, and that's really the ideal use case. so I have some users these, climate science consulting agencies, and then they might have like a one-off project where they have to generate the data once, but instead of having to maintain this server for the lifetime of that project, they just have a file on S3 and like, who cares? If that costs a couple dollars a month to run, that's fine, but it's not like S3 is gonna be deprecated, like it's gonna be on an insecure version of Ubuntu or something. So that's really the ideal, set of constraints for using Protomaps. [00:10:58] Jeremy: Yeah. Something this also makes me think about is, is like the resilience of sites like remaining online, because I, interviewed, Kyle Drake, he runs Neocities, which is like a modern version of GeoCities. And if I remember correctly, he was mentioning how a lot of old websites from that time, if they were running a server backend, like they were running PHP or something like that, if you were to try to go to those sites, now they're like pretty much all dead because there needed to be someone dedicated to running a Linux server, making sure things were patched and so on and so forth. But for static sites, like the ones that used to be hosted on GeoCities, you can go to the internet archive or other websites and they were just files, right? You can bring 'em right back up, and if anybody just puts 'em on a web server, then you're good. They're still alive. Case study of news room preferring static hosting [00:11:53] Brandon: Yeah, exactly. One place that's kind of surprising but makes sense where this comes up, is for newspapers actually. Some of the users using Protomaps are the Washington Post. And the reason they use it, is not necessarily because they don't want to pay for a SaaS like Google, but because if they make an interactive story, they have to guarantee that it still works in a couple of years. And that's like a policy decision from like the editorial board, which is like, so you can't write an article if people can't view it in five years. But if your like interactive data story is reliant on a third party, API and that third party API becomes deprecated, or it changes the pricing or it, you know, it gets acquired, then your journalism story is not gonna work anymore. So I have seen really good uptake among local news rooms and even big ones to use things like Protomaps just because it makes sense for the requirements. Working on Protomaps as an open source project for five years [00:12:49] Jeremy: How long have you been working on Protomaps and the parts that it's made up of such as PMTiles? [00:12:58] Brandon: I've been working on it for about five years, maybe a little more than that. It's sort of my pandemic era project. But the PMTiles part, which is really the heart of it only came in about halfway. Why not make a SaaS? [00:13:13] Brandon: So honestly, like when I first started it, I thought it was gonna be another SaaS and then I looked at it and looked at what the environment was around it. And I'm like, uh, so I don't really think I wanna do that. [00:13:24] Jeremy: When, when you say you looked at the environment around it what do you mean? Why did you decide not to make it a SaaS? [00:13:31] Brandon: Because there already is a lot of SaaS out there. And I think the opportunity of making something that is unique in terms of those use cases, like I mentioned like newsrooms, was clear. Like it was clear that there was some other solution, that could be built that would fit these needs better while if it was a SaaS, there are plenty of those out there. And I don't necessarily think that they're well differentiated. A lot of them all use OpenStreetMap data. And it seems like they mainly compete on price. It's like who can build the best three column pricing model. And then once you do that, you need to build like billing and metrics and authentication and like those problems don't really interest me. So I think, although I acknowledge sort of the indie hacker ethos now is to build a SaaS product with a monthly subscription, that's something I very much chose not to do, even though it is for sure like the best way to build a business. [00:14:29] Jeremy: Yeah, I mean, I think a lot of people can appreciate that perspective because it's, it's almost like we have SaaS overload, right? Where you have so many little bills for your project where you're like, another $5 a month, another $10 a month, or if you're a business, right? Those, you add a bunch of zeros and at some point it's just how many of these are we gonna stack on here? [00:14:53] Brandon: Yeah. And honestly. So I really think like as programmers, we're not really like great at choosing how to spend money like a $10 SaaS. That's like nothing. You know? So I can go to Starbucks and I can buy a pumpkin spice latte, and that's like $10 basically now, right? And it's like I'm able to make that consumer choice in like an instant just to spend money on that. But then if you're like, oh, like spend $10 on a SaaS that somebody put a lot of work into, then you're like, oh, that's too expensive. I could just do it myself. So I'm someone that also subscribes to a lot of SaaS products. and I think for a lot of things it's a great fit. Many open source SaaS projects are not easy to self host [00:15:37] Brandon: But there's always this tension between an open source project that you might be able to run yourself and a SaaS. And I think a lot of projects are at different parts of the spectrum. But for Protomaps, it's very much like I'm trying to move maps to being it is something that is so easy to run yourself that anyone can do it. [00:16:00] Jeremy: Yeah, and I think you can really see it with, there's a few SaaS projects that are successful and they're open source, but then you go to look at the self-hosting instructions and it's either really difficult to find and you find it, and then the instructions maybe don't work, or it's really complicated. So I think doing the opposite with Protomaps. As a user, I'm sure we're all appreciative, but I wonder in terms of trying to make money, if that's difficult. [00:16:30] Brandon: No, for sure. It is not like a good way to make money because I think like the ideal situation for an open source project that is open that wants to make money is the product itself is fundamentally complicated to where people are scared to run it themselves. Like a good example I can think of is like Supabase. Supabase is sort of like a platform as a service based on Postgres. And if you wanted to run it yourself, well you need to run Postgres and you need to handle backups and authentication and logging, and that stuff all needs to work and be production ready. So I think a lot of people, like they don't trust themselves to run database backups correctly. 'cause if you get it wrong once, then you're kind of screwed. So I think that fundamental aspect of the product, like a database is something that is very, very ripe for being a SaaS while still being open source because it's fundamentally hard to run. Another one I can think of is like tailscale, which is, like a VPN that works end to end. That's something where, you know, it has this networking complexity where a lot of developers don't wanna deal with that. So they'd happily pay, for tailscale as a service. There is a lot of products or open source projects that eventually end up just changing to becoming like a hosted service. Businesses going from open source to closed or restricted licenses [00:17:58] Brandon: But then in that situation why would they keep it open source, right? Like, if it's easy to run yourself well, doesn't that sort of cannibalize their business model? And I think that's really the tension overall in these open source companies. So you saw it happen to things like Elasticsearch to things like Terraform where they eventually change the license to one that makes it difficult for other companies to compete with them. [00:18:23] Jeremy: Yeah, I mean there's been a number of cases like that. I mean, specifically within the mapping community, one I can think of was Mapbox's. They have Mapbox gl. Which was a JavaScript client to visualize maps and they moved from, I forget which license they picked, but they moved to a much more restrictive license. I wonder what your thoughts are on something that releases as open source, but then becomes something maybe a little more muddy. [00:18:55] Brandon: Yeah, I think it totally makes sense because if you look at their business and their funding, it seems like for Mapbox, I haven't used it in a while, but my understanding is like a lot of their business now is car companies and doing in dash navigation. And that is probably way better of a business than trying to serve like people making maps of toilets. And I think sort of the beauty of it is that, so Mapbox, the story is they had a JavaScript renderer called Mapbox GL JS. And they changed that to a source available license a couple years ago. And there's a fork of it that I'm sort of involved in called MapLibre GL. But I think the cool part is Mapbox paid employees for years, probably millions of dollars in total to work on this thing and just gave it away for free. Right? So everyone can benefit from that work they did. It's not like that code went away, like once they changed the license. Well, the old version has been forked. It's going its own way now. It's quite different than the new version of Mapbox, but I think it's extremely generous that they're able to pay people for years, you know, like a competitive salary and just give that away. [00:20:10] Jeremy: Yeah, so we should maybe look at it as, it was a gift while it was open source, and they've given it to the community and they're on continuing on their own path, but at least the community running Map Libre, they can run with it, right? It's not like it just disappeared. [00:20:29] Brandon: Yeah, exactly. And that is something that I use for Protomaps quite extensively. Like it's the primary way of showing maps on the web and I've been trying to like work on some enhancements to it to have like better internationalization for if you are in like South Asia like not show languages correctly. So I think it is being taken in a new direction. And I think like sort of the combination of Protomaps and MapLibre, it addresses a lot of use cases, like I mentioned earlier with like these like hobby projects, indie projects that are almost certainly not interesting to someone like Mapbox or Google as a business. But I'm happy to support as a small business myself. Financially supporting open source work (GitHub sponsors, closed source, contracts) [00:21:12] Jeremy: In my previous interview with Tom, one of the main things he mentioned was that creating a mapping business is incredibly difficult, and he said he probably wouldn't do it again. So in your case, you're building Protomaps, which you've admitted is easy to self-host. So there's not a whole lot of incentive for people to pay you. How is that working out for you? How are you supporting yourself? [00:21:40] Brandon: There's a couple of strategies that I've tried and oftentimes failed at. Just to go down the list, so I do have GitHub sponsors so I do have a hosted version of Protomaps you can use if you don't want to bother copying a big file around. But the way I do the billing for that is through GitHub sponsors. If you wanted to use this thing I provide, then just be a sponsor. And that definitely pays for itself, like the cost of running it. And that's great. GitHub sponsors is so easy to set up. It just removes you having to deal with Stripe or something. 'cause a lot of people, their credit card information is already in GitHub. GitHub sponsors I think is awesome if you want to like cover costs for a project. But I think very few people are able to make that work. A thing that's like a salary job level. It's sort of like Twitch streaming, you know, there's a handful of people that are full-time streamers and then you look down the list on Twitch and it's like a lot of people that have like 10 viewers. But some of the other things I've tried, I actually started out, publishing the base map as a closed source thing, where I would sell sort of like a data package instead of being a SaaS, I'd be like, here's a one-time download, of the premium data and you can buy it. And quite a few people bought it I just priced it at like $500 for this thing. And I thought that was an interesting experiment. The main reason it's interesting is because the people that it attracts to you in terms of like, they're curious about your products, are all people willing to pay money. While if you start out everything being open source, then the people that are gonna be try to do it are only the people that want to get something for free. So what I discovered is actually like once you transition that thing from closed source to open source, a lot of the people that used to pay you money will still keep paying you money because like, it wasn't necessarily that that closed source thing was why they wanted to pay. They just valued that thought you've put into it your expertise, for example. So I think that is one thing, that I tried at the beginning was just start out, closed source proprietary, then make it open source. That's interesting to people. Like if you release something as open source, if you go the other way, like people are really mad if you start out with something open source and then later on you're like, oh, it's some other license. Then people are like that's so rotten. But I think doing it the other way, I think is quite valuable in terms of being able to find an audience. [00:24:29] Jeremy: And when you said it was closed source and paid to open source, do you still sell those map exports? [00:24:39] Brandon: I don't right now. It's something that I might do in the future, you know, like have small customizations of the data that are available, uh, for a fee. still like the core OpenStreetMap based map that's like a hundred gigs you can just download. And that'll always just be like a free download just because that's already out there. All the source code to build it is open source. So even if I said, oh, you have to pay for it, then someone else can just do it right? So there's no real reason like to make that like some sort of like paywall thing. But I think like overall if the project is gonna survive in the long term it's important that I'd ideally like to be able to like grow like a team like have a small group of people that can dedicate the time to growing the project in the long term. But I'm still like trying to figure that out right now. [00:25:34] Jeremy: And when you mentioned that when you went from closed to open and people were still paying you, you don't sell a product anymore. What were they paying for? [00:25:45] Brandon: So I have some contracts with companies basically, like if they need a feature or they need a customization in this way then I am very open to those. And I sort of set it up to make it clear from the beginning that this is not just a free thing on GitHub, this is something that you could pay for if you need help with it, if you need support, if you wanted it. I'm also a little cagey about the word support because I think like it sounds a little bit too wishy-washy. Pretty much like if you need access to the developers of an open source project, I think that's something that businesses are willing to pay for. And I think like making that clear to potential users is a challenge. But I think that is one way that you might be able to make like a living out of open source. [00:26:35] Jeremy: And I think you said you'd been working on it for about five years. Has that mostly been full time? [00:26:42] Brandon: It's been on and off. it's sort of my pandemic era project. But I've spent a lot of time, most of my time working on the open source project at this point. So I have done some things that were more just like I'm doing a customization or like a private deployment for some client. But that's been a minority of the time. Yeah. [00:27:03] Jeremy: It's still impressive to have an open source project that is easy to self-host and yet is still able to support you working on it full time. I think a lot of people might make the assumption that there's nothing to sell if something is, is easy to use. But this sort of sounds like a counterpoint to that. [00:27:25] Brandon: I think I'd like it to be. So when you come back to the point of like, it being easy to self-host. Well, so again, like I think about it as like a primitive of the web. Like for example, if you wanted to start a business today as like hosted CSS files, you know, like where you upload your CSS and then you get developers to pay you a monthly subscription for how many times they fetched a CSS file. Well, I think most developers would be like, that's stupid because it's just an open specification, you just upload a static file. And really my goal is to make Protomaps the same way where it's obvious that there's not really some sort of lock-in or some sort of secret sauce in the server that does this thing. How PMTiles works and building a primitive of the web [00:28:16] Brandon: If you look at video for example, like a lot of the tech for how Protomaps and PMTiles works is based on parts of the HTTP spec that were made for video. And 20 years ago, if you wanted to host a video on the web, you had to have like a real player license or flash. So you had to go license some server software from real media or from macromedia so you could stream video to a browser plugin. But now in HTML you can just embed a video file. And no one's like, oh well I need to go pay for my video serving license. I mean, there is such a thing, like YouTube doesn't really use that for DRM reasons, but people just have the assumption that video is like a primitive on the web. So if we're able to make maps sort of that same way like a primitive on the web then there isn't really some obvious business or licensing model behind how that works. Just because it's a thing and it helps a lot of people do their jobs and people are happy using it. So why bother? [00:29:26] Jeremy: You mentioned that it a tech that was used for streaming video. What tech specifically is it? [00:29:34] Brandon: So it is byte range serving. So when you open a video file on the web, So let's say it's like a 100 megabyte video. You don't have to download the entire video before it starts playing. It streams parts out of the file based on like what frames... I mean, it's based on the frames in the video. So it can start streaming immediately because it's organized in a way to where the first few frames are at the beginning. And what PMTiles really is, is it's just like a video but in space instead of time. So it's organized in a way where these zoomed out views are at the beginning and the most zoomed in views are at the end. So when you're like panning or zooming in the map all you're really doing is fetching byte ranges out of that file the same way as a video. But it's organized in, this tiled way on a space filling curve. IIt's a little bit complicated how it works internally and I think it's kind of cool but that's sort of an like an implementation detail. [00:30:35] Jeremy: And to the person deploying it, it just looks like a single file. [00:30:40] Brandon: Exactly in the same way like an mp3 audio file is or like a JSON file is. [00:30:47] Jeremy: So with a video, I can sort of see how as someone seeks through the video, they start at the beginning and then they go to the middle if they wanna see the middle. For a map, as somebody scrolls around the map, are you seeking all over the file or is the way it's structured have a little less chaos? [00:31:09] Brandon: It's structured. And that's kind of the main technical challenge behind building PMTiles is you have to be sort of clever so you're not spraying the reads everywhere. So it uses something called a hilbert curve, which is a mathematical concept of a space filling curve. Where it's one continuous curve that essentially lets you break 2D space into 1D space. So if you've seen some maps of IP space, it uses this crazy looking curve that hits all the points in one continuous line. And that's the same concept behind PMTiles is if you're looking at one part of the world, you're sort of guaranteed that all of those parts you're looking at are quite close to each other and the data you have to transfer is quite minimal, compared to if you just had it at random. [00:32:02] Jeremy: How big do the files get? If I have a PMTiles of the entire world, what kind of size am I looking at? [00:32:10] Brandon: Right now, the default one I distribute is 128 gigabytes, so it's quite sizable, although you can slice parts out of it remotely. So if you just wanted. if you just wanted California or just wanted LA or just wanted only a couple of zoom levels, like from zero to 10 instead of zero to 15, there is a command line tool that's also called PMTiles that lets you do that. Issues with CDNs and range queries [00:32:35] Jeremy: And when you're working with files of this size, I mean, let's say I am working with a CDN in front of my application. I'm not typically accustomed to hosting something that's that large and something that's where you're seeking all over the file. is that, ever an issue or is that something that's just taken care of by the browser and, and taken care of by, by the hosts? [00:32:58] Brandon: That is an issue actually, so a lot of CDNs don't deal with it correctly. And my recommendation is there is a kind of proxy server or like a serverless proxy thing that I wrote. That runs on like cloudflare workers or on Docker that lets you proxy those range requests into a normal URL and then that is like a hundred percent CDN compatible. So I would say like a lot of the big commercial installations of this thing, they use that because it makes more practical sense. It's also faster. But the idea is that this solution sort of scales up and scales down. If you wanted to host just your city in like a 10 megabyte file, well you can just put that into GitHub pages and you don't have to worry about it. If you want to have a global map for your website that serves a ton of traffic then you probably want a little bit more sophisticated of a solution. It still does not require you to run a Linux server, but it might require (you) to use like Lambda or Lambda in conjunction with like a CDN. [00:34:09] Jeremy: Yeah. And that sort of ties into what you were saying at the beginning where if you can host on something like CloudFlare Workers or Lambda, there's less time you have to spend keeping these things running. [00:34:26] Brandon: Yeah, exactly. and I think also the Lambda or CloudFlare workers solution is not perfect. It's not as perfect as S3 or as just static files, but in my experience, it still is better at building something that lasts on the time span of years than being like I have a server that is on this Ubuntu version and in four years there's all these like security patches that are not being applied. So it's still sort of serverless, although not totally vendor neutral like S3. Customizing the map [00:35:03] Jeremy: We've mostly been talking about how you host the map itself, but for someone who's not familiar with these kind of tools, how would they be customizing the map? [00:35:15] Brandon: For customizing the map there is front end style customization and there's also data customization. So for the front end if you wanted to change the water from the shade of blue to another shade of blue there is a TypeScript API where you can customize it almost like a text editor color scheme. So if you're able to name a bunch of colors, well you can customize the map in that way you can change the fonts. And that's all done using MapLibre GL using a TypeScript API on top of that for customizing the data. So all the pipeline to generate this data from OpenStreetMap is open source. There is a Java program using a library called PlanetTiler which is awesome, which is this super fast multi-core way of building map tiles. And right now there isn't really great hooks to customize what data goes into that. But that's something that I do wanna work on. And finally, because the data comes from OpenStreetMap if you notice data that's missing or you wanted to correct data in OSM then you can go into osm.org. You can get involved in contributing the data to OSM and the Protomaps build is daily. So if you make a change, then within 24 hours you should see the new base map. Have that change. And of course for OSM your improvements would go into every OSM based project that is ingesting that data. So it's not a protomap specific thing. It's like this big shared data source, almost like Wikipedia. OpenStreetMap is a dataset and not a map [00:37:01] Jeremy: I think you were involved with OpenStreetMap to some extent. Can you speak a little bit to that for people who aren't familiar, what OpenStreetMap is? [00:37:11] Brandon: Right. So I've been using OSM as sort of like a tools developer for over a decade now. And one of the number one questions I get from developers about what is Protomaps is why wouldn't I just use OpenStreetMap? What's the distinction between Protomaps and OpenStreetMap? And it's sort of like this funny thing because even though OSM has map in the name it's not really a map in that you can't... In that it's mostly a data set and not a map. It does have a map that you can see that you can pan around to when you go to the website but the way that thing they show you on the website is built is not really that easily reproducible. It involves a lot of c++ software you have to run. But OpenStreetMap itself, the heart of it is almost like a big XML file that has all the data in the map and global. And it has tagged features for example. So you can go in and edit that. It has a web front end to change the data. It does not directly translate into making a map actually. Protomaps decides what shows at each zoom level [00:38:24] Brandon: So a lot of the pipeline, that Java program I mentioned for building this basemap for protomaps is doing things like you have to choose what data you show when you zoom out. You can't show all the data. For example when you're zoomed out and you're looking at all of a state like Colorado you don't see all the Chipotle when you're zoomed all the way out. That'd be weird, right? So you have to make some sort of decision in logic that says this data only shows up at this zoom level. And that's really what is the challenge in optimizing the size of that for the Protomaps map project. [00:39:03] Jeremy: Oh, so those decisions of what to show at different Zoom levels those are decisions made by you when you're creating the PMTiles file with Protomaps. [00:39:14] Brandon: Exactly. It's part of the base maps build pipeline. and those are honestly very subjective decisions. Who really decides when you're zoomed out should this hospital show up or should this museum show up nowadays in Google, I think it shows you ads. Like if someone pays for their car repair shop to show up when you're zoomed out like that that gets surfaced. But because there is no advertising auction in Protomaps that doesn't happen obviously. So we have to sort of make some reasonable choice. A lot of that right now in Protomaps actually comes from another open source project called Mapzen. So Mapzen was a company that went outta business a couple years ago. They did a lot of this work in designing which data shows up at which Zoom level and open sourced it. And then when they shut down, they transferred that code into the Linux Foundation. So it's this totally open source project, that like, again, sort of like Mapbox gl has this awesome legacy in that this company funded it for years for smart people to work on it and now it's just like a free thing you can use. So the logic in Protomaps is really based on mapzen. [00:40:33] Jeremy: And so the visualization of all this... I think I understand what you mean when people say oh, why not use OpenStreetMaps because it's not really clear it's hard to tell is this the tool that's visualizing the data? Is it the data itself? So in the case of using Protomaps, it sounds like Protomaps itself has all of the data from OpenStreetMap and then it has made all the decisions for you in terms of what to show at different Zoom levels and what things to have on the map at all. And then finally, you have to have a separate, UI layer and in this case, it sounds like the one that you recommend is the Map Libre library. [00:41:18] Brandon: Yeah, that's exactly right. For Protomaps, it has a portion or a subset of OSM data. It doesn't have all of it just because there's too much, like there's data in there. people have mapped out different bushes and I don't include that in Protomaps if you wanted to go in and edit like the Java code to add that you can. But really what Protomaps is positioned at is sort of a solution for developers that want to use OSM data to make a map on their app or their website. because OpenStreetMap itself is mostly a data set, it does not really go all the way to having an end-to-end solution. Financials and the idea of a project being complete [00:41:59] Jeremy: So I think it's great that somebody who wants to make a map, they have these tools available, whether it's from what was originally built by Mapbox, what's built by Open StreetMap now, the work you're doing with Protomaps. But I wonder one of the things that I talked about with Tom was he was saying he was trying to build this mapping business and based on the financials of what was coming in he was stressed, right? He was struggling a bit. And I wonder for you, you've been working on this open source project for five years. Do you have similar stressors or do you feel like I could keep going how things are now and I feel comfortable? [00:42:46] Brandon: So I wouldn't say I'm a hundred percent in one bucket or the other. I'm still seeing it play out. One thing, that I really respect in a lot of open source projects, which I'm not saying I'm gonna do for Protomaps is the idea that a project is like finished. I think that is amazing. If a software project can just be done it's sort of like a painting or a novel once you write, finish the last page, have it seen by the editor. I send it off to the press is you're done with a book. And I think one of the pains of software is so few of us can actually do that. And I don't know obviously people will say oh the map is never finished. That's more true of OSM, but I think like for Protomaps. One thing I'm thinking about is how to limit the scope to something that's quite narrow to where we could be feature complete on the core things in the near term timeframe. That means that it does not address a lot of things that people want. Like search, like if you go to Google Maps and you search for a restaurant, you will get some hits. that's like a geocoding issue. And I've already decided that's totally outta scope for Protomaps. So, in terms of trying to think about the future of this, I'm mostly looking for ways to cut scope if possible. There are some things like better tooling around being able to work with PMTiles that are on the roadmap. but for me, I am still enjoying working on the project. It's definitely growing. So I can see on NPM downloads I can see the growth curve of people using it and that's really cool. So I like hearing about when people are using it for cool projects. So it seems to still be going okay for now. [00:44:44] Jeremy: Yeah, that's an interesting perspective about how you were talking about projects being done. Because I think when people look at GitHub projects and they go like, oh, the last commit was X months ago. They go oh well this is dead right? But maybe that's the wrong framing. Maybe you can get a project to a point where it's like, oh, it's because it doesn't need to be updated. [00:45:07] Brandon: Exactly, yeah. Like I used to do a lot of c++ programming and the best part is when you see some LAPACK matrix math library from like 1995 that still works perfectly in c++ and you're like, this is awesome. This is the one I have to use. But if you're like trying to use some like React component library and it hasn't been updated in like a year, you're like, oh, that's a problem. So again, I think there's some middle ground between those that I'm trying to find. I do like for Protomaps, it's quite dependency light in terms of the number of hard dependencies I have in software. but I do still feel like there is a lot of work to be done in terms of project scope that needs to have stuff added. You mostly only hear about problems instead of people's wins [00:45:54] Jeremy: Having run it for this long. Do you have any thoughts on running an open source project in general? On dealing with issues or managing what to work on things like that? [00:46:07] Brandon: Yeah. So I have a lot. I think one thing people point out a lot is that especially because I don't have a direct relationship with a lot of the people using it a lot of times I don't even know that they're using it. Someone sent me a message saying hey, have you seen flickr.com, like the photo site? And I'm like, no. And I went to flickr.com/map and it has Protomaps for it. And I'm like, I had no idea. But that's cool, if they're able to use Protomaps for this giant photo sharing site that's awesome. But that also means I don't really hear about when people use it successfully because you just don't know, I guess they, NPM installed it and it works perfectly and you never hear about it. You only hear about people's negative experiences. You only hear about people that come and open GitHub issues saying this is totally broken, and why doesn't this thing exist? And I'm like, well, it's because there's an infinite amount of things that I want to do, but I have a finite amount of time and I just haven't gone into that yet. And that's honestly a lot of the things and people are like when is this thing gonna be done? So that's, that's honestly part of why I don't have a public roadmap because I want to avoid that sort of bickering about it. I would say that's one of my biggest frustrations with running an open source project is how it's self-selected to only hear the negative experiences with it. Be careful what PRs you accept [00:47:32] Brandon: 'cause you don't hear about those times where it works. I'd say another thing is it's changed my perspective on contributing to open source because I think when I was younger or before I had become a maintainer I would open a pull request on a project unprompted that has a hundred lines and I'd be like, Hey, just merge this thing. But I didn't realize when I was younger well if I just merge it and I disappear, then the maintainer is stuck with what I did forever. You know if I add some feature then that person that maintains the project has to do that indefinitely. And I think that's very asymmetrical and it's changed my perspective a lot on accepting open source contributions. I wanna have it be open to anyone to contribute. But there is some amount of back and forth where it's almost like the default answer for should I accept a PR is no by default because you're the one maintaining it. And do you understand the shape of that solution completely to where you're going to support it for years because the person that's contributing it is not bound to those same obligations that you are. And I think that's also one of the things where I have a lot of trepidation around open source is I used to think of it as a lot more bazaar-like in terms of anyone can just throw their thing in. But then that creates a lot of problems for the people who are expected out of social obligation to continue this thing indefinitely. [00:49:23] Jeremy: Yeah, I can totally see why that causes burnout with a lot of open source maintainers, because you probably to some extent maybe even feel some guilt right? You're like, well, somebody took the time to make this. But then like you said you have to spend a lot of time trying to figure out is this something I wanna maintain long term? And one wrong move and it's like, well, it's in here now. [00:49:53] Brandon: Exactly. To me, I think that is a very common failure mode for open source projects is they're too liberal in the things they accept. And that's a lot of why I was talking about how that choice of what features show up on the map was inherited from the MapZen projects. If I didn't have that then somebody could come in and say hey, you know, I want to show power lines on the map. And they open a PR for power lines and now everybody who's using Protomaps when they're like zoomed out they see power lines are like I didn't want that. So I think that's part of why a lot of open source projects eventually evolve into a plugin system is because there is this demand as the project grows for more and more features. But there is a limit in the maintainers. It's like the demand for features is exponential while the maintainer amount of time and effort is linear. Plugin systems might reduce need for PRs [00:50:56] Brandon: So maybe the solution to smash that exponential down to quadratic maybe is to add a plugin system. But I think that is one of the biggest tensions that only became obvious to me after working on this for a couple of years. [00:51:14] Jeremy: Is that something you're considering doing now? [00:51:18] Brandon: Is the plugin system? Yeah. I think for the data customization, I eventually wanted to have some sort of programmatic API to where you could declare a config file that says I want ski routes. It totally makes sense. The power lines example is maybe a little bit obscure but for example like a skiing app and you want to be able to show ski slopes when you're zoomed out well you're not gonna be able to get that from Mapbox or from Google because they have a one size fits all map that's not specialized to skiing or to golfing or to outdoors. But if you like, in theory, you could do this with Protomaps if you changed the Java code to show data at different zoom levels. And that is to me what makes the most sense for a plugin system and also makes the most product sense because it enables a lot of things you cannot do with the one size fits all map. [00:52:20] Jeremy: It might also increase the complexity of the implementation though, right? [00:52:25] Brandon: Yeah, exactly. So that's like. That's really where a lot of the terrifying thoughts come in, which is like once you create this like config file surface area, well what does that look like? Is that JSON? Is that TOML, is that some weird like everything eventually evolves into some scripting language right? Where you have logic inside of your templates and I honestly do not really know what that looks like right now. That feels like something in the medium term roadmap. [00:52:58] Jeremy: Yeah and then in terms of bug reports or issues, now it's not just your code it's this exponential combination of whatever people put into these config files. [00:53:09] Brandon: Exactly. Yeah. so again, like I really respect the projects that have done this well or that have done plugins well. I'm trying to think of some, I think obsidian has plugins, for example. And that seems to be one of the few solutions to try and satisfy the infinite desire for features with the limited amount of maintainer time. Time split between code vs triage vs talking to users [00:53:36] Jeremy: How would you say your time is split between working on the code versus issue and PR triage? [00:53:43] Brandon: Oh, it varies really. I think working on the code is like a minority of it. I think something that I actually enjoy is talking to people, talking to users, getting feedback on it. I go to quite a few conferences to talk to developers or people that are interested and figure out how to refine the message, how to make it clearer to people, like what this is for. And I would say maybe a plurality of my time is spent dealing with non-technical things that are neither code or GitHub issues. One thing I've been trying to do recently is talk to people that are not really in the mapping space. For example, people that work for newspapers like a lot of them are front end developers and if you ask them to run a Linux server they're like I have no idea. But that really is like one of the best target audiences for Protomaps. So I'd say a lot of the reality of running an open source project is a lot like a business is it has all the same challenges as a business in terms of you have to figure out what is the thing you're offering. You have to deal with people using it. You have to deal with feedback, you have to deal with managing emails and stuff. I don't think the payoff is anywhere near running a business or a startup that's backed by VC money is but it's definitely not the case that if you just want to code, you should start an open source project because I think a lot of the work for an opensource project has nothing to do with just writing the code. It is in my opinion as someone having done a VC backed business before, it is a lot more similar to running, a tech company than just putting some code on GitHub. Running a startup vs open source project [00:55:43] Jeremy: Well, since you've done both at a high level what did you like about running the company versus maintaining the open source project? [00:55:52] Brandon: So I have done some venture capital accelerator programs before and I think there is an element of hype and energy that you get from that that is self perpetuating. Your co-founder is gungho on like, yeah, we're gonna do this thing. And your investors are like, you guys are geniuses. You guys are gonna make a killing doing this thing. And the way it's framed is sort of obvious to everyone that it's like there's a much more traditional set of motivations behind that, that people understand while it's definitely not the case for running an open source project. Sometimes you just wake up and you're like what the hell is this thing for, it is this thing you spend a lot of time on. You don't even know who's using it. The people that use it and make a bunch of money off of it they know nothing about it. And you know, it's just like cool. And then you only hear from people that are complaining about it. And I think like that's honestly discouraging compared to the more clear energy and clearer motivation and vision behind how most people think about a company. But what I like about the open source project is just the lack of those constraints you know? Where you have a mandate that you need to have this many customers that are paying by this amount of time. There's that sort of pressure on delivering a business result instead of just making something that you're proud of that's simple to use and has like an elegant design. I think that's really a difference in motivation as well. Having control [00:57:50] Jeremy: Do you feel like you have more control? Like you mentioned how you've decided I'm not gonna make a public roadmap. I'm the sole developer. I get to decide what goes in. What doesn't. Do you feel like you have more control in your current position than you did running the startup? [00:58:10] Brandon: Definitely for sure. Like that agency is what I value the most. It is possible to go too far. Like, so I'm very wary of the BDFL title, which I think is how a lot of open source projects succeed. But I think there is some element of for a project to succeed there has to be somebody that makes those decisions. Sometimes those decisions will be wrong and then hopefully they can be rectified. But I think going back to what I was talking about with scope, I think the overall vision and the scope of the project is something that I am very opinionated about in that it should do these things. It shouldn't do these things. It should be easy to use for this audience. Is it gonna be appealing to this other audience? I don't know. And I think that is really one of the most important parts of that leadership role, is having the power to decide we're doing this, we're not doing this. I would hope other developers would be able to get on board if they're able to make good use of the project, if they use it for their company, if they use it for their business, if they just think the project is cool. So there are other contributors at this point and I want to get more involved. But I think being able to make those decisions to what I believe is going to be the best project is something that is very special about open source, that isn't necessarily true about running like a SaaS business. [00:59:50] Jeremy: I think that's a good spot to end it on, so if people want to learn more about Protomaps or they wanna see what you're up to, where should they head? [01:00:00] Brandon: So you can go to Protomaps.com, GitHub, or you can find me or Protomaps on bluesky or Mastodon. [01:00:09] Jeremy: All right, Brandon, thank you so much for chatting today. [01:00:12] Brandon: Great. Thank you very much.

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma convida Bruno Okamoto para um papo sobre o que é os Micro SaaS, seus desafios e como é a forma de pensar nesse modelo de negócios.   Lambda3 · #428 - Micro-SaaS   Participantes: Fernando Okuma - @feokuma Bruno Okamoto - @brunomicrosaas   Pauta: O que é um Micro-SaaS? Por que criar um Micro-SaaS? Tecnologias e Ferramentas para Construção Desafios e Armadilhas Comuns Monetização e Modelos de Receita Escalando um Micro-SaaS: Crescer ou Manter Pequeno? O Futuro do Micro-SaaS Referências: Comunidade MicroSaaS Bruno Okamoto - YouTube www.indiehackers.com   Edição: Compasso Coolab   Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Lambda Weekly 03-29-25 by

Lambda Weekly 03-23-25 by

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma convida Eduardo Liron, Nelson Simplício e André Okuma para um papo sobre a relação do cinema com a tecnologia e o quando as pessoas de uma área tão artística precisam estar muito atualizadas softwares e hardwares.    Lambda3 · #427 - Cinema e tecnologia   Participantes: Fernando Okuma - @feokuma Eduardo Liron - @eduardoliron Nelson Simplício - @nelsonsimplicio André Okuma - @andre_okuma   Pauta: Um pouco de história do cinema A evolução do cinema em paralelo com a tecnologia As mortes do cinema Ferramentas que ajudam a fazer cinema (além das essenciais) Desafios para produzir, armazenar e manter os artefatos atualizados Abrindo arquivos gerados em versões antigas de softwares de edição Como as "facilidades" das novas tecnologias influenciam na forma de fazer cinema Quanto faz sentido novas gerações de cineastas aprenderem tecnologias antigas Comunidade de discussão técnica de produção de cinema Futuro do cinema e tecnologia Referências: Livro Cinema e Guerra - Paul Virilio Livro O fim do cinema?: Uma mídia em crise na era digital - André Gaudreault e Philippe Marion Celtx - Software para escrita de roteiro - https://www.celtx.com @moti.curta Davinci Resolve - https://www.blackmagicdesign.com/br/products/davinciresolve/ Mobiliário Urbano - www.moburb.org   Edição: Compasso Coolab   Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Lambda weekly 03-16-25 by

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma, Luiza Prata, Jonatan Crespo e Ingrid Soares conversam sobre ameaças que podem colocar em risco projetos de frontend e possíveis formas de mitigá-las para proteger nossas aplicações e as pessoas usuárias.   Lambda3 · #426 - Seguranca no desenvolvimento de frontends modernos   Participantes: Fernando Okuma - @feokuma Luiza Prata - @luiza-prata-soldi-passos Jonatan Crespo - @jonatan-crespo Ingrid Soares - @ingridrauany   Pauta: Por que a segurança em frontends é importante nos dias de hoje? Principais Ameaças em Frontends Cross-Site Scripting (XSS) Cross-Site Request Forgery (CSRF) Injeção de Código Malicioso Quebra de autenticação e gerenciamento de sessões Boas Práticas de segurança Adotar frameworks ajuda a mitigar problemas de segurança? Já vem com boas práticas embutidas? Ferramentas e tecnologias Linters e ferramentas de análise estática Monitoramento de vulnerabilidades de bibliotecas A relação com o backend Segurança compartilhada. Parte da segurança fica no backend Praticas de integração entre o backend e o frontend   Edição: Compasso Coolab   Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

In this episode, we provide an overview of AWS Step Functions and dive deep into the powerful new JSONata and variables features. We explain how JSONata allows complex JSON transformations without custom Lambda functions, enabling more serverless workflows. The variables feature also helps avoid the previous 256KB state size limit. We share examples from real projects showing how these features simplify workflows, reduce costs and enable new use cases.AWS Bites is brought to you in association with fourTheorem. If you need a friendly partner to support you and work with you to de-risk any AWS migration or development project, check them out at ⁠⁠fourtheorem.com⁠⁠In this episode, we mentioned the following resources:JSONata and variables official launch post: https://aws.amazon.com/blogs/compute/simplifying-developer-experience-with-variables-and-jsonata-in-aws-step-functions/JSONata exerciser: https://try.jsonata.org/Stedi JSONata playground: https://www.stedi.com/jsonata/playgroundEpisode 103: Building GenAI Features with Bedrock https://awsbites.com/103-building-genai-features-with-bedrock/Episode 63: How to automate transcripts with Amazon Transcribe and OpenAI Whisper https://awsbites.com/63-how-to-automate-transcripts-with-amazon-transcribe-and-openai-whisper/ Do you have any AWS questions you would like us to address?Leave a comment here or connect with us on X/Twitter, BlueSky or LinkedIn:- ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://twitter.com/eoins⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ | ⁠https://bsky.app/profile/eoin.sh⁠ | ⁠https://www.linkedin.com/in/eoins/⁠- ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://twitter.com/loige⁠⁠⁠⁠⁠ | ⁠https://bsky.app/profile/loige.co⁠ | ⁠https://www.linkedin.com/in/lucianomammino/

Êtes-vous certain de maîtriser la langue française ? Règles de grammaire étonnantes, abus de langage, vocabulaire mal employé, origine insoupçonnée d'expressions... vous allez être surpris ! Hébergé par Acast. Visitez acast.com/privacy pour plus d'informations.

Rust Multiple Entry Points: Architectural PatternsKey PointsCore Concept: Multiple entry points in Rust enable single codebase deployment across CLI, microservices, WebAssembly and GUI contextsImplementation Path: Initial CLI development → Web API → Lambda/cloud functionsCargo Integration: Native support via src/bin directory or explicit binary targets in Cargo.tomlTechnical AdvantagesMemory Safety: Consistent safety guarantees across deployment targetsType Consistency: Strong typing ensures API contract integrity between interfacesAsync Model: Unified asynchronous execution model across environmentsBinary Optimization: Compile-time optimizations yield superior performance vs runtime interpretationOwnership Model: No-saved-state philosophy aligns with Lambda execution contextDeployment ArchitectureCore Logic Isolation: Business logic encapsulated in library cratesInterface Separation: Entry point-specific code segregated from core functionalityBuild Pipeline: Single compilation source enables consistent artifact generationInfrastructure Consistency: Uniform deployment targets eliminate environment-specific bugsResource Optimization: Shared components reduce binary size and memory footprintImplementation BenefitsIteration Speed: CLI provides immediate feedback loop during core developmentSecurity Posture: Memory safety extends across all deployment targetsAPI Consistency: JSON payload structures remain identical between CLI and web interfacesEvent Architecture: Natural alignment with event-driven cloud function patternsCompile-Time Optimizations: CPU-specific enhancements available at binary generation

Lambda Weekly 03-09-25 by

Last week marked the 7th annual #21toWatch awards, a key event in the Cambridge startup scene that recognises rising talent across People, Companies, and Things. This year's list is dominated by ground-breaking innovations in neuroscience, medtech, and AI-driven diagnostics.Today's episode features:· Details of the winners across all three categories.· Podcast interviews with winners Lucy Jung from LYEONS Neurotech (People) and Monica Saavedra from Lambda agri (Thing).· Insights from independent judge, Christine Martin (Cambridge Enterprise), who shared her experience of the judging process.· Recognition of the independent judges: Christine Martin, Chris Ellis (Innovate UK Business Growth), Jo Słota-Newson (Almanac Ventures), and Tom Hughes (Trinity College Cambridge). · Up to £20,000 prize for each winner from the event sponsors· And, James caught up with Emily Stoner, Careers Consultant at the University of Cambridge Careers Office, to discuss the growing interest in entrepreneurship among students and researchers.Produced by Cambridge TV Hosted on Acast. See acast.com/privacy for more information.

Lambda Weekly 03-02-25 by

Nesse episódio do Podcast da Lambda powered by TIVIT, Tatiana Sartori convidou Roxana Ferrieri e Valdineia Ruiz para um bate papo contando suas trajetórias e desafios das Mulheres em Tecnologia.   Lambda3 · #425 - Desafio das mulheres na liderança em tecnologia   Participantes: Tatiana Sartori - @tatiana-sartori Roxana Laura Ferrieri - @roxanalauraferrieri Valdineia Nascimento Ruiz - @valdineia-nascimento-ruiz   Pauta: Dados importantes do mercado Trajetória de carreira Desafios enfrentados até chegar à posições de liderança Enfrentando preconceitos sendo uma liderança Equilibrando a vida pessoal com a vida profissional Conselhos e dicas   Edição: Compasso Coolab   Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma convida Graziela Nobre e Fabrício Carraro para bater um papo sobre de onde vem as pessoas especialistas dessa área tão nova e complexa que envolve inteligência artificial.   Lambda3 · #424 - De onde vem as pessoas especialistas de IA?   Participantes: Fernando Okuma - @feokuma Graziela Nobre - @graziela-nobre Fabrício Carraro - @fabriciocarraro Pauta: Existe um caminho tradicional? Graduação Cursos de especialização Cursos Online Quais são as possíveis áreas de atuação com esse conhecimento? Criar modelos de IA Engenharia de Prompt Ser a pessoa que treina o modelo Pre Requisitos Conhecimentos para servir de base para esta profissão Equipamento (precisa ter um super computador?) Carreiras e oportunidades Quais são as oportunidades que existem no mercado hoje? Existe trainee nessa área?  Qual caminho para migrar da área de desenvolvimento de software para atuações com IA? Referências e recomendações Referências: Steven Wolfram Kaggle Podcast - Let's Data Podcast - IA Sob Controle Alura Alura - Techguide.sh Suno Inteligência Artificial e ChatGPT   Edição: Compasso Coolab   Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Today's episode is with Paul Klein, founder of Browserbase. We talked about building browser infrastructure for AI agents, the future of agent authentication, and their open source framework Stagehand.* [00:00:00] Introductions* [00:04:46] AI-specific challenges in browser infrastructure* [00:07:05] Multimodality in AI-Powered Browsing* [00:12:26] Running headless browsers at scale* [00:18:46] Geolocation when proxying* [00:21:25] CAPTCHAs and Agent Auth* [00:28:21] Building “User take over” functionality* [00:33:43] Stagehand: AI web browsing framework* [00:38:58] OpenAI's Operator and computer use agents* [00:44:44] Surprising use cases of Browserbase* [00:47:18] Future of browser automation and market competition* [00:53:11] Being a solo founderTranscriptAlessio [00:00:04]: Hey everyone, welcome to the Latent Space podcast. This is Alessio, partner and CTO at Decibel Partners, and I'm joined by my co-host Swyx, founder of Smol.ai.swyx [00:00:12]: Hey, and today we are very blessed to have our friends, Paul Klein, for the fourth, the fourth, CEO of Browserbase. Welcome.Paul [00:00:21]: Thanks guys. Yeah, I'm happy to be here. I've been lucky to know both of you for like a couple of years now, I think. So it's just like we're hanging out, you know, with three ginormous microphones in front of our face. It's totally normal hangout.swyx [00:00:34]: Yeah. We've actually mentioned you on the podcast, I think, more often than any other Solaris tenant. Just because like you're one of the, you know, best performing, I think, LLM tool companies that have started up in the last couple of years.Paul [00:00:50]: Yeah, I mean, it's been a whirlwind of a year, like Browserbase is actually pretty close to our first birthday. So we are one years old. And going from, you know, starting a company as a solo founder to... To, you know, having a team of 20 people, you know, a series A, but also being able to support hundreds of AI companies that are building AI applications that go out and automate the web. It's just been like, really cool. It's been happening a little too fast. I think like collectively as an AI industry, let's just take a week off together. I took my first vacation actually two weeks ago, and Operator came out on the first day, and then a week later, DeepSeat came out. And I'm like on vacation trying to chill. I'm like, we got to build with this stuff, right? So it's been a breakneck year. But I'm super happy to be here and like talk more about all the stuff we're seeing. And I'd love to hear kind of what you guys are excited about too, and share with it, you know?swyx [00:01:39]: Where to start? So people, you've done a bunch of podcasts. I think I strongly recommend Jack Bridger's Scaling DevTools, as well as Turner Novak's The Peel. And, you know, I'm sure there's others. So you covered your Twilio story in the past, talked about StreamClub, you got acquired to Mux, and then you left to start Browserbase. So maybe we just start with what is Browserbase? Yeah.Paul [00:02:02]: Browserbase is the web browser for your AI. We're building headless browser infrastructure, which are browsers that run in a server environment that's accessible to developers via APIs and SDKs. It's really hard to run a web browser in the cloud. You guys are probably running Chrome on your computers, and that's using a lot of resources, right? So if you want to run a web browser or thousands of web browsers, you can't just spin up a bunch of lambdas. You actually need to use a secure containerized environment. You have to scale it up and down. It's a stateful system. And that infrastructure is, like, super painful. And I know that firsthand, because at my last company, StreamClub, I was CTO, and I was building our own internal headless browser infrastructure. That's actually why we sold the company, is because Mux really wanted to buy our headless browser infrastructure that we'd built. And it's just a super hard problem. And I actually told my co-founders, I would never start another company unless it was a browser infrastructure company. And it turns out that's really necessary in the age of AI, when AI can actually go out and interact with websites, click on buttons, fill in forms. You need AI to do all of that work in an actual browser running somewhere on a server. And BrowserBase powers that.swyx [00:03:08]: While you're talking about it, it occurred to me, not that you're going to be acquired or anything, but it occurred to me that it would be really funny if you became the Nikita Beer of headless browser companies. You just have one trick, and you make browser companies that get acquired.Paul [00:03:23]: I truly do only have one trick. I'm screwed if it's not for headless browsers. I'm not a Go programmer. You know, I'm in AI grant. You know, browsers is an AI grant. But we were the only company in that AI grant batch that used zero dollars on AI spend. You know, we're purely an infrastructure company. So as much as people want to ask me about reinforcement learning, I might not be the best guy to talk about that. But if you want to ask about headless browser infrastructure at scale, I can talk your ear off. So that's really my area of expertise. And it's a pretty niche thing. Like, nobody has done what we're doing at scale before. So we're happy to be the experts.swyx [00:03:59]: You do have an AI thing, stagehand. We can talk about the sort of core of browser-based first, and then maybe stagehand. Yeah, stagehand is kind of the web browsing framework. Yeah.What is Browserbase? Headless Browser Infrastructure ExplainedAlessio [00:04:10]: Yeah. Yeah. And maybe how you got to browser-based and what problems you saw. So one of the first things I worked on as a software engineer was integration testing. Sauce Labs was kind of like the main thing at the time. And then we had Selenium, we had Playbrite, we had all these different browser things. But it's always been super hard to do. So obviously you've worked on this before. When you started browser-based, what were the challenges? What were the AI-specific challenges that you saw versus, there's kind of like all the usual running browser at scale in the cloud, which has been a problem for years. What are like the AI unique things that you saw that like traditional purchase just didn't cover? Yeah.AI-specific challenges in browser infrastructurePaul [00:04:46]: First and foremost, I think back to like the first thing I did as a developer, like as a kid when I was writing code, I wanted to write code that did stuff for me. You know, I wanted to write code to automate my life. And I do that probably by using curl or beautiful soup to fetch data from a web browser. And I think I still do that now that I'm in the cloud. And the other thing that I think is a huge challenge for me is that you can't just create a web site and parse that data. And we all know that now like, you know, taking HTML and plugging that into an LLM, you can extract insights, you can summarize. So it was very clear that now like dynamic web scraping became very possible with the rise of large language models or a lot easier. And that was like a clear reason why there's been more usage of headless browsers, which are necessary because a lot of modern websites don't expose all of their page content via a simple HTTP request. You know, they actually do require you to run this type of code for a specific time. JavaScript on the page to hydrate this. Airbnb is a great example. You go to airbnb.com. A lot of that content on the page isn't there until after they run the initial hydration. So you can't just scrape it with a curl. You need to have some JavaScript run. And a browser is that JavaScript engine that's going to actually run all those requests on the page. So web data retrieval was definitely one driver of starting BrowserBase and the rise of being able to summarize that within LLM. Also, I was familiar with if I wanted to automate a website, I could write one script and that would work for one website. It was very static and deterministic. But the web is non-deterministic. The web is always changing. And until we had LLMs, there was no way to write scripts that you could write once that would run on any website. That would change with the structure of the website. Click the login button. It could mean something different on many different websites. And LLMs allow us to generate code on the fly to actually control that. So I think that rise of writing the generic automation scripts that can work on many different websites, to me, made it clear that browsers are going to be a lot more useful because now you can automate a lot more things without writing. If you wanted to write a script to book a demo call on 100 websites, previously, you had to write 100 scripts. Now you write one script that uses LLMs to generate that script. That's why we built our web browsing framework, StageHand, which does a lot of that work for you. But those two things, web data collection and then enhanced automation of many different websites, it just felt like big drivers for more browser infrastructure that would be required to power these kinds of features.Alessio [00:07:05]: And was multimodality also a big thing?Paul [00:07:08]: Now you can use the LLMs to look, even though the text in the dome might not be as friendly. Maybe my hot take is I was always kind of like, I didn't think vision would be as big of a driver. For UI automation, I felt like, you know, HTML is structured text and large language models are good with structured text. But it's clear that these computer use models are often vision driven, and they've been really pushing things forward. So definitely being multimodal, like rendering the page is required to take a screenshot to give that to a computer use model to take actions on a website. And it's just another win for browser. But I'll be honest, that wasn't what I was thinking early on. I didn't even think that we'd get here so fast with multimodality. I think we're going to have to get back to multimodal and vision models.swyx [00:07:50]: This is one of those things where I forgot to mention in my intro that I'm an investor in Browserbase. And I remember that when you pitched to me, like a lot of the stuff that we have today, we like wasn't on the original conversation. But I did have my original thesis was something that we've talked about on the podcast before, which is take the GPT store, the custom GPT store, all the every single checkbox and plugin is effectively a startup. And this was the browser one. I think the main hesitation, I think I actually took a while to get back to you. The main hesitation was that there were others. Like you're not the first hit list browser startup. It's not even your first hit list browser startup. There's always a question of like, will you be the category winner in a place where there's a bunch of incumbents, to be honest, that are bigger than you? They're just not targeted at the AI space. They don't have the backing of Nat Friedman. And there's a bunch of like, you're here in Silicon Valley. They're not. I don't know.Paul [00:08:47]: I don't know if that's, that was it, but like, there was a, yeah, I mean, like, I think I tried all the other ones and I was like, really disappointed. Like my background is from working at great developer tools, companies, and nothing had like the Vercel like experience. Um, like our biggest competitor actually is partly owned by private equity and they just jacked up their prices quite a bit. And the dashboard hasn't changed in five years. And I actually used them at my last company and tried them and I was like, oh man, like there really just needs to be something that's like the experience of these great infrastructure companies, like Stripe, like clerk, like Vercel that I use in love, but oriented towards this kind of like more specific category, which is browser infrastructure, which is really technically complex. Like a lot of stuff can go wrong on the internet when you're running a browser. The internet is very vast. There's a lot of different configurations. Like there's still websites that only work with internet explorer out there. How do you handle that when you're running your own browser infrastructure? These are the problems that we have to think about and solve at BrowserBase. And it's, it's certainly a labor of love, but I built this for me, first and foremost, I know it's super cheesy and everyone says that for like their startups, but it really, truly was for me. If you look at like the talks I've done even before BrowserBase, and I'm just like really excited to try and build a category defining infrastructure company. And it's, it's rare to have a new category of infrastructure exists. We're here in the Chroma offices and like, you know, vector databases is a new category of infrastructure. Is it, is it, I mean, we can, we're in their office, so, you know, we can, we can debate that one later. That is one.Multimodality in AI-Powered Browsingswyx [00:10:16]: That's one of the industry debates.Paul [00:10:17]: I guess we go back to the LLMOS talk that Karpathy gave way long ago. And like the browser box was very clearly there and it seemed like the people who were building in this space also agreed that browsers are a core primitive of infrastructure for the LLMOS that's going to exist in the future. And nobody was building something there that I wanted to use. So I had to go build it myself.swyx [00:10:38]: Yeah. I mean, exactly that talk that, that honestly, that diagram, every box is a startup and there's the code box and then there's the. The browser box. I think at some point they will start clashing there. There's always the question of the, are you a point solution or are you the sort of all in one? And I think the point solutions tend to win quickly, but then the only ones have a very tight cohesive experience. Yeah. Let's talk about just the hard problems of browser base you have on your website, which is beautiful. Thank you. Was there an agency that you used for that? Yeah. Herb.paris.Paul [00:11:11]: They're amazing. Herb.paris. Yeah. It's H-E-R-V-E. I highly recommend for developers. Developer tools, founders to work with consumer agencies because they end up building beautiful things and the Parisians know how to build beautiful interfaces. So I got to give prep.swyx [00:11:24]: And chat apps, apparently are, they are very fast. Oh yeah. The Mistral chat. Yeah. Mistral. Yeah.Paul [00:11:31]: Late chat.swyx [00:11:31]: Late chat. And then your videos as well, it was professionally shot, right? The series A video. Yeah.Alessio [00:11:36]: Nico did the videos. He's amazing. Not the initial video that you shot at the new one. First one was Austin.Paul [00:11:41]: Another, another video pretty surprised. But yeah, I mean, like, I think when you think about how you talk about your company. You have to think about the way you present yourself. It's, you know, as a developer, you think you evaluate a company based on like the API reliability and the P 95, but a lot of developers say, is the website good? Is the message clear? Do I like trust this founder? I'm building my whole feature on. So I've tried to nail that as well as like the reliability of the infrastructure. You're right. It's very hard. And there's a lot of kind of foot guns that you run into when running headless browsers at scale. Right.Competing with Existing Headless Browser Solutionsswyx [00:12:10]: So let's pick one. You have eight features here. Seamless integration. Scalability. Fast or speed. Secure. Observable. Stealth. That's interesting. Extensible and developer first. What comes to your mind as like the top two, three hardest ones? Yeah.Running headless browsers at scalePaul [00:12:26]: I think just running headless browsers at scale is like the hardest one. And maybe can I nerd out for a second? Is that okay? I heard this is a technical audience, so I'll talk to the other nerds. Whoa. They were listening. Yeah. They're upset. They're ready. The AGI is angry. Okay. So. So how do you run a browser in the cloud? Let's start with that, right? So let's say you're using a popular browser automation framework like Puppeteer, Playwright, and Selenium. Maybe you've written a code, some code locally on your computer that opens up Google. It finds the search bar and then types in, you know, search for Latent Space and hits the search button. That script works great locally. You can see the little browser open up. You want to take that to production. You want to run the script in a cloud environment. So when your laptop is closed, your browser is doing something. The browser is doing something. Well, I, we use Amazon. You can see the little browser open up. You know, the first thing I'd reach for is probably like some sort of serverless infrastructure. I would probably try and deploy on a Lambda. But Chrome itself is too big to run on a Lambda. It's over 250 megabytes. So you can't easily start it on a Lambda. So you maybe have to use something like Lambda layers to squeeze it in there. Maybe use a different Chromium build that's lighter. And you get it on the Lambda. Great. It works. But it runs super slowly. It's because Lambdas are very like resource limited. They only run like with one vCPU. You can run one process at a time. Remember, Chromium is super beefy. It's barely running on my MacBook Air. I'm still downloading it from a pre-run. Yeah, from the test earlier, right? I'm joking. But it's big, you know? So like Lambda, it just won't work really well. Maybe it'll work, but you need something faster. Your users want something faster. Okay. Well, let's put it on a beefier instance. Let's get an EC2 server running. Let's throw Chromium on there. Great. Okay. I can, that works well with one user. But what if I want to run like 10 Chromium instances, one for each of my users? Okay. Well, I might need two EC2 instances. Maybe 10. All of a sudden, you have multiple EC2 instances. This sounds like a problem for Kubernetes and Docker, right? Now, all of a sudden, you're using ECS or EKS, the Kubernetes or container solutions by Amazon. You're spending up and down containers, and you're spending a whole engineer's time on kind of maintaining this stateful distributed system. Those are some of the worst systems to run because when it's a stateful distributed system, it means that you are bound by the connections to that thing. You have to keep the browser open while someone is working with it, right? That's just a painful architecture to run. And there's all this other little gotchas with Chromium, like Chromium, which is the open source version of Chrome, by the way. You have to install all these fonts. You want emojis working in your browsers because your vision model is looking for the emoji. You need to make sure you have the emoji fonts. You need to make sure you have all the right extensions configured, like, oh, do you want ad blocking? How do you configure that? How do you actually record all these browser sessions? Like it's a headless browser. You can't look at it. So you need to have some sort of observability. Maybe you're recording videos and storing those somewhere. It all kind of adds up to be this just giant monster piece of your project when all you wanted to do was run a lot of browsers in production for this little script to go to google.com and search. And when I see a complex distributed system, I see an opportunity to build a great infrastructure company. And we really abstract that away with Browserbase where our customers can use these existing frameworks, Playwright, Publisher, Selenium, or our own stagehand and connect to our browsers in a serverless-like way. And control them, and then just disconnect when they're done. And they don't have to think about the complex distributed system behind all of that. They just get a browser running anywhere, anytime. Really easy to connect to.swyx [00:15:55]: I'm sure you have questions. My standard question with anything, so essentially you're a serverless browser company, and there's been other serverless things that I'm familiar with in the past, serverless GPUs, serverless website hosting. That's where I come from with Netlify. One question is just like, you promised to spin up thousands of servers. You promised to spin up thousands of browsers in milliseconds. I feel like there's no real solution that does that yet. And I'm just kind of curious how. The only solution I know, which is to kind of keep a kind of warm pool of servers around, which is expensive, but maybe not so expensive because it's just CPUs. So I'm just like, you know. Yeah.Browsers as a Core Primitive in AI InfrastructurePaul [00:16:36]: You nailed it, right? I mean, how do you offer a serverless-like experience with something that is clearly not serverless, right? And the answer is, you need to be able to run... We run many browsers on single nodes. We use Kubernetes at browser base. So we have many pods that are being scheduled. We have to predictably schedule them up or down. Yes, thousands of browsers in milliseconds is the best case scenario. If you hit us with 10,000 requests, you may hit a slower cold start, right? So we've done a lot of work on predictive scaling and being able to kind of route stuff to different regions where we have multiple regions of browser base where we have different pools available. You can also pick the region you want to go to based on like lower latency, round trip, time latency. It's very important with these types of things. There's a lot of requests going over the wire. So for us, like having a VM like Firecracker powering everything under the hood allows us to be super nimble and spin things up or down really quickly with strong multi-tenancy. But in the end, this is like the complex infrastructural challenges that we have to kind of deal with at browser base. And we have a lot more stuff on our roadmap to allow customers to have more levers to pull to exchange, do you want really fast browser startup times or do you want really low costs? And if you're willing to be more flexible on that, we may be able to kind of like work better for your use cases.swyx [00:17:44]: Since you used Firecracker, shouldn't Fargate do that for you or did you have to go lower level than that? We had to go lower level than that.Paul [00:17:51]: I find this a lot with Fargate customers, which is alarming for Fargate. We used to be a giant Fargate customer. Actually, the first version of browser base was ECS and Fargate. And unfortunately, it's a great product. I think we were actually the largest Fargate customer in our region for a little while. No, what? Yeah, seriously. And unfortunately, it's a great product, but I think if you're an infrastructure company, you actually have to have a deeper level of control over these primitives. I think it's the same thing is true with databases. We've used other database providers and I think-swyx [00:18:21]: Yeah, serverless Postgres.Paul [00:18:23]: Shocker. When you're an infrastructure company, you're on the hook if any provider has an outage. And I can't tell my customers like, hey, we went down because so-and-so went down. That's not acceptable. So for us, we've really moved to bringing things internally. It's kind of opposite of what we preach. We tell our customers, don't build this in-house, but then we're like, we build a lot of stuff in-house. But I think it just really depends on what is in the critical path. We try and have deep ownership of that.Alessio [00:18:46]: On the distributed location side, how does that work for the web where you might get sort of different content in different locations, but the customer is expecting, you know, if you're in the US, I'm expecting the US version. But if you're spinning up my browser in France, I might get the French version. Yeah.Paul [00:19:02]: Yeah. That's a good question. Well, generally, like on the localization, there is a thing called locale in the browser. You can set like what your locale is. If you're like in the ENUS browser or not, but some things do IP, IP based routing. And in that case, you may want to have a proxy. Like let's say you're running something in the, in Europe, but you want to make sure you're showing up from the US. You may want to use one of our proxy features so you can turn on proxies to say like, make sure these connections always come from the United States, which is necessary too, because when you're browsing the web, you're coming from like a, you know, data center IP, and that can make things a lot harder to browse web. So we do have kind of like this proxy super network. Yeah. We have a proxy for you based on where you're going, so you can reliably automate the web. But if you get scheduled in Europe, that doesn't happen as much. We try and schedule you as close to, you know, your origin that you're trying to go to. But generally you have control over the regions you can put your browsers in. So you can specify West one or East one or Europe. We only have one region of Europe right now, actually. Yeah.Alessio [00:19:55]: What's harder, the browser or the proxy? I feel like to me, it feels like actually proxying reliably at scale. It's much harder than spending up browsers at scale. I'm curious. It's all hard.Paul [00:20:06]: It's layers of hard, right? Yeah. I think it's different levels of hard. I think the thing with the proxy infrastructure is that we work with many different web proxy providers and some are better than others. Some have good days, some have bad days. And our customers who've built browser infrastructure on their own, they have to go and deal with sketchy actors. Like first they figure out their own browser infrastructure and then they got to go buy a proxy. And it's like you can pay in Bitcoin and it just kind of feels a little sus, right? It's like you're buying drugs when you're trying to get a proxy online. We have like deep relationships with these counterparties. We're able to audit them and say, is this proxy being sourced ethically? Like it's not running on someone's TV somewhere. Is it free range? Yeah. Free range organic proxies, right? Right. We do a level of diligence. We're SOC 2. So we have to understand what is going on here. But then we're able to make sure that like we route around proxy providers not working. There's proxy providers who will just, the proxy will stop working all of a sudden. And then if you don't have redundant proxying on your own browsers, that's hard down for you or you may get some serious impacts there. With us, like we intelligently know, hey, this proxy is not working. Let's go to this one. And you can kind of build a network of multiple providers to really guarantee the best uptime for our customers. Yeah. So you don't own any proxies? We don't own any proxies. You're right. The team has been saying who wants to like take home a little proxy server, but not yet. We're not there yet. You know?swyx [00:21:25]: It's a very mature market. I don't think you should build that yourself. Like you should just be a super customer of them. Yeah. Scraping, I think, is the main use case for that. I guess. Well, that leads us into CAPTCHAs and also off, but let's talk about CAPTCHAs. You had a little spiel that you wanted to talk about CAPTCHA stuff.Challenges of Scaling Browser InfrastructurePaul [00:21:43]: Oh, yeah. I was just, I think a lot of people ask, if you're thinking about proxies, you're thinking about CAPTCHAs too. I think it's the same thing. You can go buy CAPTCHA solvers online, but it's the same buying experience. It's some sketchy website, you have to integrate it. It's not fun to buy these things and you can't really trust that the docs are bad. What Browserbase does is we integrate a bunch of different CAPTCHAs. We do some stuff in-house, but generally we just integrate with a bunch of known vendors and continually monitor and maintain these things and say, is this working or not? Can we route around it or not? These are CAPTCHA solvers. CAPTCHA solvers, yeah. Not CAPTCHA providers, CAPTCHA solvers. Yeah, sorry. CAPTCHA solvers. We really try and make sure all of that works for you. I think as a dev, if I'm buying infrastructure, I want it all to work all the time and it's important for us to provide that experience by making sure everything does work and monitoring it on our own. Yeah. Right now, the world of CAPTCHAs is tricky. I think AI agents in particular are very much ahead of the internet infrastructure. CAPTCHAs are designed to block all types of bots, but there are now good bots and bad bots. I think in the future, CAPTCHAs will be able to identify who a good bot is, hopefully via some sort of KYC. For us, we've been very lucky. We have very little to no known abuse of Browserbase because we really look into who we work with. And for certain types of CAPTCHA solving, we only allow them on certain types of plans because we want to make sure that we can know what people are doing, what their use cases are. And that's really allowed us to try and be an arbiter of good bots, which is our long term goal. I want to build great relationships with people like Cloudflare so we can agree, hey, here are these acceptable bots. We'll identify them for you and make sure we flag when they come to your website. This is a good bot, you know?Alessio [00:23:23]: I see. And Cloudflare said they want to do more of this. So they're going to set by default, if they think you're an AI bot, they're going to reject. I'm curious if you think this is something that is going to be at the browser level or I mean, the DNS level with Cloudflare seems more where it should belong. But I'm curious how you think about it.Paul [00:23:40]: I think the web's going to change. You know, I think that the Internet as we have it right now is going to change. And we all need to just accept that the cat is out of the bag. And instead of kind of like wishing the Internet was like it was in the 2000s, we can have free content line that wouldn't be scraped. It's just it's not going to happen. And instead, we should think about like, one, how can we change? How can we change the models of, you know, information being published online so people can adequately commercialize it? But two, how do we rebuild applications that expect that AI agents are going to log in on their behalf? Those are the things that are going to allow us to kind of like identify good and bad bots. And I think the team at Clerk has been doing a really good job with this on the authentication side. I actually think that auth is the biggest thing that will prevent agents from accessing stuff, not captchas. And I think there will be agent auth in the future. I don't know if it's going to happen from an individual company, but actually authentication providers that have a, you know, hidden login as agent feature, which will then you put in your email, you'll get a push notification, say like, hey, your browser-based agent wants to log into your Airbnb. You can approve that and then the agent can proceed. That really circumvents the need for captchas or logging in as you and sharing your password. I think agent auth is going to be one way we identify good bots going forward. And I think a lot of this captcha solving stuff is really short-term problems as the internet kind of reorients itself around how it's going to work with agents browsing the web, just like people do. Yeah.Managing Distributed Browser Locations and Proxiesswyx [00:24:59]: Stitch recently was on Hacker News for talking about agent experience, AX, which is a thing that Netlify is also trying to clone and coin and talk about. And we've talked about this on our previous episodes before in a sense that I actually think that's like maybe the only part of the tech stack that needs to be kind of reinvented for agents. Everything else can stay the same, CLIs, APIs, whatever. But auth, yeah, we need agent auth. And it's mostly like short-lived, like it should not, it should be a distinct, identity from the human, but paired. I almost think like in the same way that every social network should have your main profile and then your alt accounts or your Finsta, it's almost like, you know, every, every human token should be paired with the agent token and the agent token can go and do stuff on behalf of the human token, but not be presumed to be the human. Yeah.Paul [00:25:48]: It's like, it's, it's actually very similar to OAuth is what I'm thinking. And, you know, Thread from Stitch is an investor, Colin from Clerk, Octaventures, all investors in browser-based because like, I hope they solve this because they'll make browser-based submission more possible. So we don't have to overcome all these hurdles, but I think it will be an OAuth-like flow where an agent will ask to log in as you, you'll approve the scopes. Like it can book an apartment on Airbnb, but it can't like message anybody. And then, you know, the agent will have some sort of like role-based access control within an application. Yeah. I'm excited for that.swyx [00:26:16]: The tricky part is just, there's one, one layer of delegation here, which is like, you're authoring my user's user or something like that. I don't know if that's tricky or not. Does that make sense? Yeah.Paul [00:26:25]: You know, actually at Twilio, I worked on the login identity and access. Management teams, right? So like I built Twilio's login page.swyx [00:26:31]: You were an intern on that team and then you became the lead in two years? Yeah.Paul [00:26:34]: Yeah. I started as an intern in 2016 and then I was the tech lead of that team. How? That's not normal. I didn't have a life. He's not normal. Look at this guy. I didn't have a girlfriend. I just loved my job. I don't know. I applied to 500 internships for my first job and I got rejected from every single one of them except for Twilio and then eventually Amazon. And they took a shot on me and like, I was getting paid money to write code, which was my dream. Yeah. Yeah. I'm very lucky that like this coding thing worked out because I was going to be doing it regardless. And yeah, I was able to kind of spend a lot of time on a team that was growing at a company that was growing. So it informed a lot of this stuff here. I think these are problems that have been solved with like the SAML protocol with SSO. I think it's a really interesting stuff with like WebAuthn, like these different types of authentication, like schemes that you can use to authenticate people. The tooling is all there. It just needs to be tweaked a little bit to work for agents. And I think the fact that there are companies that are already. Providing authentication as a service really sets it up. Well, the thing that's hard is like reinventing the internet for agents. We don't want to rebuild the internet. That's an impossible task. And I think people often say like, well, we'll have this second layer of APIs built for agents. I'm like, we will for the top use cases, but instead of we can just tweak the internet as is, which is on the authentication side, I think we're going to be the dumb ones going forward. Unfortunately, I think AI is going to be able to do a lot of the tasks that we do online, which means that it will be able to go to websites, click buttons on our behalf and log in on our behalf too. So with this kind of like web agent future happening, I think with some small structural changes, like you said, it feels like it could all slot in really nicely with the existing internet.Handling CAPTCHAs and Agent Authenticationswyx [00:28:08]: There's one more thing, which is the, your live view iframe, which lets you take, take control. Yeah. Obviously very key for operator now, but like, was, is there anything interesting technically there or that the people like, well, people always want this.Paul [00:28:21]: It was really hard to build, you know, like, so, okay. Headless browsers, you don't see them, right. They're running. They're running in a cloud somewhere. You can't like look at them. And I just want to really make, it's a weird name. I wish we came up with a better name for this thing, but you can't see them. Right. But customers don't trust AI agents, right. At least the first pass. So what we do with our live view is that, you know, when you use browser base, you can actually embed a live view of the browser running in the cloud for your customer to see it working. And that's what the first reason is the build trust, like, okay, so I have this script. That's going to go automate a website. I can embed it into my web application via an iframe and my customer can watch. I think. And then we added two way communication. So now not only can you watch the browser kind of being operated by AI, if you want to pause and actually click around type within this iframe that's controlling a browser, that's also possible. And this is all thanks to some of the lower level protocol, which is called the Chrome DevTools protocol. It has a API called start screencast, and you can also send mouse clicks and button clicks to a remote browser. And this is all embeddable within iframes. You have a browser within a browser, yo. And then you simulate the screen, the click on the other side. Exactly. And this is really nice often for, like, let's say, a capture that can't be solved. You saw this with Operator, you know, Operator actually uses a different approach. They use VNC. So, you know, you're able to see, like, you're seeing the whole window here. What we're doing is something a little lower level with the Chrome DevTools protocol. It's just PNGs being streamed over the wire. But the same thing is true, right? Like, hey, I'm running a window. Pause. Can you do something in this window? Human. Okay, great. Resume. Like sometimes 2FA tokens. Like if you get that text message, you might need a person to type that in. Web agents need human-in-the-loop type workflows still. You still need a person to interact with the browser. And building a UI to proxy that is kind of hard. You may as well just show them the whole browser and say, hey, can you finish this up for me? And then let the AI proceed on afterwards. Is there a future where I stream my current desktop to browser base? I don't think so. I think we're very much cloud infrastructure. Yeah. You know, but I think a lot of the stuff we're doing, we do want to, like, build tools. Like, you know, we'll talk about the stage and, you know, web agent framework in a second. But, like, there's a case where a lot of people are going desktop first for, you know, consumer use. And I think cloud is doing a lot of this, where I expect to see, you know, MCPs really oriented around the cloud desktop app for a reason, right? Like, I think a lot of these tools are going to run on your computer because it makes... I think it's breaking out. People are putting it on a server. Oh, really? Okay. Well, sweet. We'll see. We'll see that. I was surprised, though, wasn't I? I think that the browser company, too, with Dia Browser, it runs on your machine. You know, it's going to be...swyx [00:30:50]: What is it?Paul [00:30:51]: So, Dia Browser, as far as I understand... I used to use Arc. Yeah. I haven't used Arc. But I'm a big fan of the browser company. I think they're doing a lot of cool stuff in consumer. As far as I understand, it's a browser where you have a sidebar where you can, like, chat with it and it can control the local browser on your machine. So, if you imagine, like, what a consumer web agent is, which it lives alongside your browser, I think Google Chrome has Project Marina, I think. I almost call it Project Marinara for some reason. I don't know why. It's...swyx [00:31:17]: No, I think it's someone really likes the Waterworld. Oh, I see. The classic Kevin Costner. Yeah.Paul [00:31:22]: Okay. Project Marinara is a similar thing to the Dia Browser, in my mind, as far as I understand it. You have a browser that has an AI interface that will take over your mouse and keyboard and control the browser for you. Great for consumer use cases. But if you're building applications that rely on a browser and it's more part of a greater, like, AI app experience, you probably need something that's more like infrastructure, not a consumer app.swyx [00:31:44]: Just because I have explored a little bit in this area, do people want branching? So, I have the state. Of whatever my browser's in. And then I want, like, 100 clones of this state. Do people do that? Or...Paul [00:31:56]: People don't do it currently. Yeah. But it's definitely something we're thinking about. I think the idea of forking a browser is really cool. Technically, kind of hard. We're starting to see this in code execution, where people are, like, forking some, like, code execution, like, processes or forking some tool calls or branching tool calls. Haven't seen it at the browser level yet. But it makes sense. Like, if an AI agent is, like, using a website and it's not sure what path it wants to take to crawl this website. To find the information it's looking for. It would make sense for it to explore both paths in parallel. And that'd be a very, like... A road not taken. Yeah. And hopefully find the right answer. And then say, okay, this was actually the right one. And memorize that. And go there in the future. On the roadmap. For sure. Don't make my roadmap, please. You know?Alessio [00:32:37]: How do you actually do that? Yeah. How do you fork? I feel like the browser is so stateful for so many things.swyx [00:32:42]: Serialize the state. Restore the state. I don't know.Paul [00:32:44]: So, it's one of the reasons why we haven't done it yet. It's hard. You know? Like, to truly fork, it's actually quite difficult. The naive way is to open the same page in a new tab and then, like, hope that it's at the same thing. But if you have a form halfway filled, you may have to, like, take the whole, you know, container. Pause it. All the memory. Duplicate it. Restart it from there. It could be very slow. So, we haven't found a thing. Like, the easy thing to fork is just, like, copy the page object. You know? But I think there needs to be something a little bit more robust there. Yeah.swyx [00:33:12]: So, MorphLabs has this infinite branch thing. Like, wrote a custom fork of Linux or something that let them save the system state and clone it. MorphLabs, hit me up. I'll be a customer. Yeah. That's the only. I think that's the only way to do it. Yeah. Like, unless Chrome has some special API for you. Yeah.Paul [00:33:29]: There's probably something we'll reverse engineer one day. I don't know. Yeah.Alessio [00:33:32]: Let's talk about StageHand, the AI web browsing framework. You have three core components, Observe, Extract, and Act. Pretty clean landing page. What was the idea behind making a framework? Yeah.Stagehand: AI web browsing frameworkPaul [00:33:43]: So, there's three frameworks that are very popular or already exist, right? Puppeteer, Playwright, Selenium. Those are for building hard-coded scripts to control websites. And as soon as I started to play with LLMs plus browsing, I caught myself, you know, code-genning Playwright code to control a website. I would, like, take the DOM. I'd pass it to an LLM. I'd say, can you generate the Playwright code to click the appropriate button here? And it would do that. And I was like, this really should be part of the frameworks themselves. And I became really obsessed with SDKs that take natural language as part of, like, the API input. And that's what StageHand is. StageHand exposes three APIs, and it's a super set of Playwright. So, if you go to a page, you may want to take an action, click on the button, fill in the form, etc. That's what the act command is for. You may want to extract some data. This one takes a natural language, like, extract the winner of the Super Bowl from this page. You can give it a Zod schema, so it returns a structured output. And then maybe you're building an API. You can do an agent loop, and you want to kind of see what actions are possible on this page before taking one. You can do observe. So, you can observe the actions on the page, and it will generate a list of actions. You can guide it, like, give me actions on this page related to buying an item. And you can, like, buy it now, add to cart, view shipping options, and pass that to an LLM, an agent loop, to say, what's the appropriate action given this high-level goal? So, StageHand isn't a web agent. It's a framework for building web agents. And we think that agent loops are actually pretty close to the application layer because every application probably has different goals or different ways it wants to take steps. I don't think I've seen a generic. Maybe you guys are the experts here. I haven't seen, like, a really good AI agent framework here. Everyone kind of has their own special sauce, right? I see a lot of developers building their own agent loops, and they're using tools. And I view StageHand as the browser tool. So, we expose act, extract, observe. Your agent can call these tools. And from that, you don't have to worry about it. You don't have to worry about generating playwright code performantly. You don't have to worry about running it. You can kind of just integrate these three tool calls into your agent loop and reliably automate the web.swyx [00:35:48]: A special shout-out to Anirudh, who I met at your dinner, who I think listens to the pod. Yeah. Hey, Anirudh.Paul [00:35:54]: Anirudh's a man. He's a StageHand guy.swyx [00:35:56]: I mean, the interesting thing about each of these APIs is they're kind of each startup. Like, specifically extract, you know, Firecrawler is extract. There's, like, Expand AI. There's a whole bunch of, like, extract companies. They just focus on extract. I'm curious. Like, I feel like you guys are going to collide at some point. Like, right now, it's friendly. Everyone's in a blue ocean. At some point, it's going to be valuable enough that there's some turf battle here. I don't think you have a dog in a fight. I think you can mock extract to use an external service if they're better at it than you. But it's just an observation that, like, in the same way that I see each option, each checkbox in the side of custom GBTs becoming a startup or each box in the Karpathy chart being a startup. Like, this is also becoming a thing. Yeah.Paul [00:36:41]: I mean, like, so the way StageHand works is that it's MIT-licensed, completely open source. You bring your own API key to your LLM of choice. You could choose your LLM. We don't make any money off of the extract or really. We only really make money if you choose to run it with our browser. You don't have to. You can actually use your own browser, a local browser. You know, StageHand is completely open source for that reason. And, yeah, like, I think if you're building really complex web scraping workflows, I don't know if StageHand is the tool for you. I think it's really more if you're building an AI agent that needs a few general tools or if it's doing a lot of, like, web automation-intensive work. But if you're building a scraping company, StageHand is not your thing. You probably want something that's going to, like, get HTML content, you know, convert that to Markdown, query it. That's not what StageHand does. StageHand is more about reliability. I think we focus a lot on reliability and less so on cost optimization and speed at this point.swyx [00:37:33]: I actually feel like StageHand, so the way that StageHand works, it's like, you know, page.act, click on the quick start. Yeah. It's kind of the integration test for the code that you would have to write anyway, like the Puppeteer code that you have to write anyway. And when the page structure changes, because it always does, then this is still the test. This is still the test that I would have to write. Yeah. So it's kind of like a testing framework that doesn't need implementation detail.Paul [00:37:56]: Well, yeah. I mean, Puppeteer, Playwright, and Slenderman were all designed as testing frameworks, right? Yeah. And now people are, like, hacking them together to automate the web. I would say, and, like, maybe this is, like, me being too specific. But, like, when I write tests, if the page structure changes. Without me knowing, I want that test to fail. So I don't know if, like, AI, like, regenerating that. Like, people are using StageHand for testing. But it's more for, like, usability testing, not, like, testing of, like, does the front end, like, has it changed or not. Okay. But generally where we've seen people, like, really, like, take off is, like, if they're using, you know, something. If they want to build a feature in their application that's kind of like Operator or Deep Research, they're using StageHand to kind of power that tool calling in their own agent loop. Okay. Cool.swyx [00:38:37]: So let's go into Operator, the first big agent launch of the year from OpenAI. Seems like they have a whole bunch scheduled. You were on break and your phone blew up. What's your just general view of computer use agents is what they're calling it. The overall category before we go into Open Operator, just the overall promise of Operator. I will observe that I tried it once. It was okay. And I never tried it again.OpenAI's Operator and computer use agentsPaul [00:38:58]: That tracks with my experience, too. Like, I'm a huge fan of the OpenAI team. Like, I think that I do not view Operator as the company. I'm not a company killer for browser base at all. I think it actually shows people what's possible. I think, like, computer use models make a lot of sense. And I'm actually most excited about computer use models is, like, their ability to, like, really take screenshots and reasoning and output steps. I think that using mouse click or mouse coordinates, I've seen that proved to be less reliable than I would like. And I just wonder if that's the right form factor. What we've done with our framework is anchor it to the DOM itself, anchor it to the actual item. So, like, if it's clicking on something, it's clicking on that thing, you know? Like, it's more accurate. No matter where it is. Yeah, exactly. Because it really ties in nicely. And it can handle, like, the whole viewport in one go, whereas, like, Operator can only handle what it sees. Can you hover? Is hovering a thing that you can do? I don't know if we expose it as a tool directly, but I'm sure there's, like, an API for hovering. Like, move mouse to this position. Yeah, yeah, yeah. I think you can trigger hover, like, via, like, the JavaScript on the DOM itself. But, no, I think, like, when we saw computer use, everyone's eyes lit up because they realized, like, wow, like, AI is going to actually automate work for people. And I think seeing that kind of happen from both of the labs, and I'm sure we're going to see more labs launch computer use models, I'm excited to see all the stuff that people build with it. I think that I'd love to see computer use power, like, controlling a browser on browser base. And I think, like, Open Operator, which was, like, our open source version of OpenAI's Operator, was our first take on, like, how can we integrate these models into browser base? And we handle the infrastructure and let the labs do the models. I don't have a sense that Operator will be released as an API. I don't know. Maybe it will. I'm curious to see how well that works because I think it's going to be really hard for a company like OpenAI to do things like support CAPTCHA solving or, like, have proxies. Like, I think it's hard for them structurally. Imagine this New York Times headline, OpenAI CAPTCHA solving. Like, that would be a pretty bad headline, this New York Times headline. Browser base solves CAPTCHAs. No one cares. No one cares. And, like, our investors are bored. Like, we're all okay with this, you know? We're building this company knowing that the CAPTCHA solving is short-lived until we figure out how to authenticate good bots. I think it's really hard for a company like OpenAI, who has this brand that's so, so good, to balance with, like, the icky parts of web automation, which it can be kind of complex to solve. I'm sure OpenAI knows who to call whenever they need you. Yeah, right. I'm sure they'll have a great partnership.Alessio [00:41:23]: And is Open Operator just, like, a marketing thing for you? Like, how do you think about resource allocation? So, you can spin this up very quickly. And now there's all this, like, open deep research, just open all these things that people are building. We started it, you know. You're the original Open. We're the original Open operator, you know? Is it just, hey, look, this is a demo, but, like, we'll help you build out an actual product for yourself? Like, are you interested in going more of a product route? That's kind of the OpenAI way, right? They started as a model provider and then…Paul [00:41:53]: Yeah, we're not interested in going the product route yet. I view Open Operator as a model provider. It's a reference project, you know? Let's show people how to build these things using the infrastructure and models that are out there. And that's what it is. It's, like, Open Operator is very simple. It's an agent loop. It says, like, take a high-level goal, break it down into steps, use tool calling to accomplish those steps. It takes screenshots and feeds those screenshots into an LLM with the step to generate the right action. It uses stagehand under the hood to actually execute this action. It doesn't use a computer use model. And it, like, has a nice interface using the live view that we talked about, the iframe, to embed that into an application. So I felt like people on launch day wanted to figure out how to build their own version of this. And we turned that around really quickly to show them. And I hope we do that with other things like deep research. We don't have a deep research launch yet. I think David from AOMNI actually has an amazing open deep research that he launched. It has, like, 10K GitHub stars now. So he's crushing that. But I think if people want to build these features natively into their application, they need good reference projects. And I think Open Operator is a good example of that.swyx [00:42:52]: I don't know. Actually, I'm actually pretty bullish on API-driven operator. Because that's the only way that you can sort of, like, once it's reliable enough, obviously. And now we're nowhere near. But, like, give it five years. It'll happen, you know. And then you can sort of spin this up and browsers are working in the background and you don't necessarily have to know. And it just is booking restaurants for you, whatever. I can definitely see that future happening. I had this on the landing page here. This might be a slightly out of order. But, you know, you have, like, sort of three use cases for browser base. Open Operator. Or this is the operator sort of use case. It's kind of like the workflow automation use case. And it completes with UiPath in the sort of RPA category. Would you agree with that? Yeah, I would agree with that. And then there's Agents we talked about already. And web scraping, which I imagine would be the bulk of your workload right now, right?Paul [00:43:40]: No, not at all. I'd say actually, like, the majority is browser automation. We're kind of expensive for web scraping. Like, I think that if you're building a web scraping product, if you need to do occasional web scraping or you have to do web scraping that works every single time, you want to use browser automation. Yeah. You want to use browser-based. But if you're building web scraping workflows, what you should do is have a waterfall. You should have the first request is a curl to the website. See if you can get it without even using a browser. And then the second request may be, like, a scraping-specific API. There's, like, a thousand scraping APIs out there that you can use to try and get data. Scraping B. Scraping B is a great example, right? Yeah. And then, like, if those two don't work, bring out the heavy hitter. Like, browser-based will 100% work, right? It will load the page in a real browser, hydrate it. I see.swyx [00:44:21]: Because a lot of people don't render to JS.swyx [00:44:25]: Yeah, exactly.Paul [00:44:26]: So, I mean, the three big use cases, right? Like, you know, automation, web data collection, and then, you know, if you're building anything agentic that needs, like, a browser tool, you want to use browser-based.Alessio [00:44:35]: Is there any use case that, like, you were super surprised by that people might not even think about? Oh, yeah. Or is it, yeah, anything that you can share? The long tail is crazy. Yeah.Surprising use cases of BrowserbasePaul [00:44:44]: One of the case studies on our website that I think is the most interesting is this company called Benny. So, the way that it works is if you're on food stamps in the United States, you can actually get rebates if you buy certain things. Yeah. You buy some vegetables. You submit your receipt to the government. They'll give you a little rebate back. Say, hey, thanks for buying vegetables. It's good for you. That process of submitting that receipt is very painful. And the way Benny works is you use their app to take a photo of your receipt, and then Benny will go submit that receipt for you and then deposit the money into your account. That's actually using no AI at all. It's all, like, hard-coded scripts. They maintain the scripts. They've been doing a great job. And they build this amazing consumer app. But it's an example of, like, all these, like, tedious workflows that people have to do to kind of go about their business. And they're doing it for the sake of their day-to-day lives. And I had never known about, like, food stamp rebates or the complex forms you have to do to fill them. But the world is powered by millions and millions of tedious forms, visas. You know, Emirate Lighthouse is a customer, right? You know, they do the O1 visa. Millions and millions of forms are taking away humans' time. And I hope that Browserbase can help power software that automates away the web forms that we don't need anymore. Yeah.swyx [00:45:49]: I mean, I'm very supportive of that. I mean, forms. I do think, like, government itself is a big part of it. I think the government itself should embrace AI more to do more sort of human-friendly form filling. Mm-hmm. But I'm not optimistic. I'm not holding my breath. Yeah. We'll see. Okay. I think I'm about to zoom out. I have a little brief thing on computer use, and then we can talk about founder stuff, which is, I tend to think of developer tooling markets in impossible triangles, where everyone starts in a niche, and then they start to branch out. So I already hinted at a little bit of this, right? We mentioned more. We mentioned E2B. We mentioned Firecrawl. And then there's Browserbase. So there's, like, all this stuff of, like, have serverless virtual computer that you give to an agent and let them do stuff with it. And there's various ways of connecting it to the internet. You can just connect to a search API, like SERP API, whatever other, like, EXA is another one. That's what you're searching. You can also have a JSON markdown extractor, which is Firecrawl. Or you can have a virtual browser like Browserbase, or you can have a virtual machine like Morph. And then there's also maybe, like, a virtual sort of code environment, like Code Interpreter. So, like, there's just, like, a bunch of different ways to tackle the problem of give a computer to an agent. And I'm just kind of wondering if you see, like, everyone's just, like, happily coexisting in their respective niches. And as a developer, I just go and pick, like, a shopping basket of one of each. Or do you think that you eventually, people will collide?Future of browser automation and market competitionPaul [00:47:18]: I think that currently it's not a zero-sum market. Like, I think we're talking about... I think we're talking about all of knowledge work that people do that can be automated online. All of these, like, trillions of hours that happen online where people are working. And I think that there's so much software to be built that, like, I tend not to think about how these companies will collide. I just try to solve the problem as best as I can and make this specific piece of infrastructure, which I think is an important primitive, the best I possibly can. And yeah. I think there's players that are actually going to like it. I think there's players that are going to launch, like, over-the-top, you know, platforms, like agent platforms that have all these tools built in, right? Like, who's building the rippling for agent tools that has the search tool, the browser tool, the operating system tool, right? There are some. There are some. There are some, right? And I think in the end, what I have seen as my time as a developer, and I look at all the favorite tools that I have, is that, like, for tools and primitives with sufficient levels of complexity, you need to have a solution that's really bespoke to that primitive, you know? And I am sufficiently convinced that the browser is complex enough to deserve a primitive. Obviously, I have to. I'm the founder of BrowserBase, right? I'm talking my book. But, like, I think maybe I can give you one spicy take against, like, maybe just whole OS running. I think that when I look at computer use when it first came out, I saw that the majority of use cases for computer use were controlling a browser. And do we really need to run an entire operating system just to control a browser? I don't think so. I don't think that's necessary. You know, BrowserBase can run browsers for way cheaper than you can if you're running a full-fledged OS with a GUI, you know, operating system. And I think that's just an advantage of the browser. It is, like, browsers are little OSs, and you can run them very efficiently if you orchestrate it well. And I think that allows us to offer 90% of the, you know, functionality in the platform needed at 10% of the cost of running a full OS. Yeah.Open Operator: Browserbase's Open-Source Alternativeswyx [00:49:16]: I definitely see the logic in that. There's a Mark Andreessen quote. I don't know if you know this one. Where he basically observed that the browser is turning the operating system into a poorly debugged set of device drivers, because most of the apps are moved from the OS to the browser. So you can just run browsers.Paul [00:49:31]: There's a place for OSs, too. Like, I think that there are some applications that only run on Windows operating systems. And Eric from pig.dev in this upcoming YC batch, or last YC batch, like, he's building all run tons of Windows operating systems for you to control with your agent. And like, there's some legacy EHR systems that only run on Internet-controlled systems. Yeah.Paul [00:49:54]: I think that's it. I think, like, there are use cases for specific operating systems for specific legacy software. And like, I'm excited to see what he does with that. I just wanted to give a shout out to the pig.dev website.swyx [00:50:06]: The pigs jump when you click on them. Yeah. That's great.Paul [00:50:08]: Eric, he's the former co-founder of banana.dev, too.swyx [00:50:11]: Oh, that Eric. Yeah. That Eric. Okay. Well, he abandoned bananas for pigs. I hope he doesn't start going around with pigs now.Alessio [00:50:18]: Like he was going around with bananas. A little toy pig. Yeah. Yeah. I love that. What else are we missing? I think we covered a lot of, like, the browser-based product history, but. What do you wish people asked you? Yeah.Paul [00:50:29]: I wish people asked me more about, like, what will the future of software look like? Because I think that's really where I've spent a lot of time about why do browser-based. Like, for me, starting a company is like a means of last resort. Like, you shouldn't start a company unless you absolutely have to. And I remain convinced that the future of software is software that you're going to click a button and it's going to do stuff on your behalf. Right now, software. You click a button and it maybe, like, calls it back an API and, like, computes some numbers. It, like, modifies some text, whatever. But the future of software is software using software. So, I may log into my accounting website for my business, click a button, and it's going to go load up my Gmail, search my emails, find the thing, upload the receipt, and then comment it for me. Right? And it may use it using APIs, maybe a browser. I don't know. I think it's a little bit of both. But that's completely different from how we've built software so far. And that's. I think that future of software has different infrastructure requirements. It's going to require different UIs. It's going to require different pieces of infrastructure. I think the browser infrastructure is one piece that fits into that, along with all the other categories you mentioned. So, I think that it's going to require developers to think differently about how they've built software for, you know

Lambda Weekly 02-19-25 by

Lambda Weekly 02-23-25 by

Managing security for a device that can autonomously interact with third-party services presents unique orchestration challenges that go beyond traditional IoT security models. In this episode of Detection at Scale, Matthew Domko, Head of Security at Rabbit, gives Jack an in-depth look at building security programs for AI-powered hardware at scale.   He details how his team achieved 100% infrastructure-as-code coverage while maintaining the agility needed for rapid product iteration. Matt also challenges conventional approaches to scaling security operations, advocating for a serverless-first architecture that has fundamentally changed how they handle detection engineering. His insights on using private LLMs via Amazon Bedrock to analyze security events showcase a pragmatic approach to AI adoption, focusing on augmentation of existing workflows rather than wholesale replacement of human analysis.  Topics discussed: How transitioning from reactive SIEM operations to a data-first security approach using AWS Lambda and SQS enabled Rabbit's team to handle complex orchestration monitoring without maintaining persistent infrastructure.  The practical implementation of LLM-assisted detection engineering, using Amazon Bedrock to analyze 15-minute blocks of security telemetry across their stack.  A deep dive into security data lake architecture decisions, including how their team addressed the challenge of cost attribution when security telemetry becomes valuable to other engineering teams.  The evolution from traditional detection engineering to a "detection-as-code" pipeline that leverages infrastructure-as-code for security rules, enabling version control, peer review, and automated testing of detection logic while maintaining rapid deployment capabilities. Concrete examples of integrating security into the engineering workflow, including how they use LLMs to transform security tickets to match engineering team nomenclature and communication patterns. Technical details of their data ingestion architecture using AWS SQS and Lambda, showing how two well-documented core patterns enabled the team to rapidly onboard new data sources and detection capabilities without direct security team involvement. A pragmatic framework for evaluating where generative AI adds value in security operations, focusing on specific use cases like log analysis and detection engineering where the technology demonstrably improves existing workflows rather than attempting wholesale process automation.  Listen to more episodes:  Apple  Spotify  YouTube Website

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma, Priscila Torres, Victor Campos, Pedro Rodrigues e Marcio Jardim batem um papo sobre como é quando chega o momento de trocar de celular, computador ou algum desses gadgets tão queridos.   Lambda3 · #423 - Trocando o device   Participantes: Fernando Okuma - @feokuma Priscila Torres - @priscilatorresoli Victor Campos - @victor-campos Pedro Rodrigues Marcio Jardim - @m4rcio-j4rdim   Pauta: Como vocês sabem que chegou a hora de trocar um device? celular travando, bateria indo rápido, lentidão no computador etc. Por que trocar? performance, obsolescência programada, novas funcionalidades desejadas Planejamento antes da troca Orçamento Pesquisa de mercado Influências de reviews, fóruns, amigos e conteúdo técnico. o que observar em cada tipo de device (resolução e taxa de atualização em TVs, RAM e SSD em notebooks, câmeras em celulares etc. Sustentabilidade e descarte correto Doação, reciclagem ou revenda? A experiência pós-troca Troca que deu certo Troca que deu errado   Edição: Compasso Coolab   Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

In this episode, we explore DuckDB, an open-source analytical database known for its speed and simplicity. Discover how DuckDB stands out in various applications and compare it to other tools like SQLite, Athena, Pandas, and Polars. We also demonstrate integrating DuckDB with AWS Lambda and Step Functions for serverless analytics.AWS Bites is brought to you by fourTheorem. If you are looking for a partner to architect, develop and modernise on AWS, give fourTheorem a call. Check out ⁠fourtheorem.com⁠In this episode, we mentioned the following resources: Our `duck-query-lambda`, A Lambda runtime for DuckDB queries: https://github.com/fourTheorem/duck-query-lambda DuckDB's official website: https://duckdb.org/ LibSQL: https://github.com/tursodatabase/libsql Do you have any AWS questions you would like us to address?Leave a comment here or connect with us on X/Twitter, BlueSky or LinkedIn:- ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://twitter.com/eoins⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ | https://bsky.app/profile/eoin.sh | https://www.linkedin.com/in/eoins/- ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://twitter.com/loige⁠⁠⁠⁠ | https://bsky.app/profile/loige.co | https://www.linkedin.com/in/lucianomammino/

In this episode, we discuss using AWS Lambda Powertools for Python to build serverless REST APIs with AWS Lambda. We cover the benefits of using Powertools for routing, validation, OpenAPI support, and more. Powertools provides an excellent framework for building APIs while maintaining Lambda best practices.In this episode, we mentioned the following resources: AWS Bites 41. How can Middy make writing Lambda functions easier? - ⁠https://awsbites.com/41-how-can-middy-make-writing-lambda-functions-easier⁠ AWS Bites 120. Lambda Best Practices - ⁠https://awsbites.com/120-lambda-best-practices/⁠ REST API - Powertools for AWS Lambda (Python) - ⁠https://docs.powertools.aws.dev/lambda/python/latest/core/event_handler/api_gateway/⁠ Hono - ⁠https://hono.dev/⁠ Fastify - ⁠https://fastify.dev/⁠ Axum - ⁠https://github.com/tokio-rs/axum⁠ FastAPI - ⁠https://fastapi.tiangolo.com/⁠Do you have any AWS questions you would like us to address?Leave a comment here or connect with us on BlueSky or LinkedIn: https://bsky.app/profile/eoin.sh | https://www.linkedin.com/in/eoins/ https://bsky.app/profile/loige.co | https://www.linkedin.com/in/lucianomammino/

On this episode of Bone Talk, BHOF CEO Claire Gill is joined by Kristen "N'ylah" Singharat, Philanthropy Manager for Delta Phi Lambda Sorority, Inc. In 2018, Delta Phi Lambda selected osteoporosis awareness as their national philanthropy focus, partnering with American Bone Health, which later merged into the Bone Health and Osteoporosis Foundation (BHOF). This partnership has provided valuable educational resources, helping members become peer educators who inform their communities about bone health.

Lambda Weekly 02-08-25 by

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma, Giovanni Bassi e Alexandre Chohfi conversam sobre automação residencial e compartilham suas experiências com tecnologias, protocolos e placas diversas, além comentar sobre os recursos do Home Assistant.   Lambda3 · #422 - Automatizando a casa   Participantes: Fernando Okuma - @feokuma Giovanni Bassi - @giovannibassi Alexandre Chohfi - @alexandrechohfi Pauta: Como vocês começaram a explorar automação residencial? Por que explorar o Home Assistant? Em qual device está rodando o Home Assistant? Estão usando o módulo de voz - Voice PE Problemas e dificuldades enfrentadas no começo Alguma automação que pareceu impossível mas deu certo? Onde encontram inspiração para automatizar? Explorando o ecossistema Dispositivos e hardwares favoritos Automação como melhoria de qualidade de vida ou só por diversão? Algum hábito ou rotina mudou por causa das automações? Fez modificações na estrutura e infraestrutura de casa para acomodar alguma automação? Dicas para quem quer começar a se aventurar ​ Referências https://www.youtube.com/@LNPBR Home Assistant · GitHub Raspberry Pi ESP32 Wi-Fi & Bluetooth SoC | Espressif Systems This new smart speaker will change everything! - Youtube Video https://homeassistantbrasil.com.br/ https://www.youtube.com/@JeffGeerling https://github.com/xZetsubou/hass-localtuya   Edição: Compasso Coolab   Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Episode 184: Multiple Myeloma BasicsSub-Interns and future Drs. Di Tran and Jessica Avila explain the symptoms, work up and treatment of multiple myeloma. Written by Di Tran, MSIV, Ross University School of Medicine; Xiyuan Yang, MSIV, American University of the Caribbean. Comments by Jessica Avila, MSIV, American University of the Caribbean. Edits by Hector Arreaza, MD.You are listening to Rio Bravo qWeek Podcast, your weekly dose of knowledge brought to you by the Rio Bravo Family Medicine Residency Program from Bakersfield, California, a UCLA-affiliated program sponsored by Clinica Sierra Vista, Let Us Be Your Healthcare Home. This podcast was created for educational purposes only. Visit your primary care provider for additional medical advice.Di: Hi everyone, this is Di Tran, 4th year medical student from Ross university.  It's a pleasure to be back.  To be honest, this project is a part of teamwork of two medical students, myself and another 4th year, her name is XiYuan.  She came from the AUC. Unfortunately, due to personal matters she was unable to make it to the recording today which makes me feel really sad. Jessica: My name is Jessica Avila, MSIV, American University of the Caribbean.Di: The topic we will present today is Multiple Myeloma. Multiple myeloma is typically a rare disease and it's actually a type of blood cancer that affects plasma cells in the bone marrow.Jessica: Let's start with a case: A 66-year-old male comes to his family doctor for an annual health checkup. He is not in any acute distress but he reports that he has been feeling tired and weaker than usual for the last 3 months. He also noticed that he tends to bruise easily. He has a history of arthritis and chronic joint pain, but he thinks his back pain has gotten worse in the last couple of months. Upon checking his lab values, his family doctor found that he has a calcium level of 10.8 and a creatinine level of 1.2, which has increased from his baseline. Given all that information, what do you think his family doctor is suspecting? And what kind of tests she can order for further evaluation?Di: Those symptoms sound awfully familiar – are we talking about the CRAB? You know, the diagnostic criteria for Multiple Myeloma.Jessica: Exactly! Those are called “myeloma-defining events.” Do you remember what those are?Di: CRAB criteria comes in 4 flavors.  It's HYPERCALCEMIA with >1mg/dL, RENAL INSUFFICIENCY with serum creatinine >2mg/dL, ANEMIA with hemoglobin value 10% plasma cells, PLUS any one or more of the CRAB features, we can make the official diagnosis of multiple myeloma. Di:  Before we go deeper, let's back up a little bit and do a little background.  So, what do we know about the immunoglobulins, also known as antibodies? Back from years of studying from medical school, we know that the plasma cells are the ones that producing the antibodies that help fight infections.  There  are various kinds that come with various functions.  Each antibody is made up of 2 heavy chains and 2 light chains.  For heavy chains, we have A, D, E, G, M and for light chains we have Kappa and Lambda.Jessica: Usually, the 5 possible types of immunoglobulins for heavy chains would be written as IgG, IgA, IgD, IgE, and IgM.  And the most common type in the bloodstream is nonetheless the IgG. Di: What is multiple myeloma? In myeloma, all the abnormal plasma cells make the same type of antibody, the monoclonal antibody.  The cause of myeloma is unknown, but there are lots of studies and evidence that show a number of potential etiologies, including viral, genetic, and exposure to toxic chemicals, especially the Agent Orange, which is a chemical used as herbicide and defoliant. It was used as a chemical warfare by the U.S. military during the Vietnam War from 1961 to 1971.Jessica: We need to order some specific blood tests to see if there is elevated monoclonal proteins in the blood or urine. So, to begin with we'll need to take a very thorough history and physical exam. Next, we'll do labs, such as CBC, basic metabolic panel, calcium, serum beta-2 microglobulin, LDH, total protein, and some not so common tests: serum protein electrophoresis (SPEP), immunofixation of blood or urine (IFE), quantitative immunoglobulins (QIg), serum free light chain assay, and serum heavy/light chain ratio assay.If any of the results is abnormal, we should consider referring our patient to an oncologist.Di: Interesting! I read that Multiple Myeloma symptoms vary in different patients.  In fact, about 10-20% of patients with newly diagnosed myeloma do not have any symptoms at all.   Otherwise, classic symptomatic presentations are weakness, fatigue, increased bruising under the skin, reduced urine output, weakened bones that is likely prone to fractures, etc. And if multiple myeloma is highly suspected, a Bone Marrow biopsy should be done with testing for flow cytometry and fluorescent in situ hybridization (FISH). Actually, if any of the “Biomarkers of malignancy (SLIM)” is met we can also diagnose multiple myeloma even without the CRAB criteria. Jessica: The diagnosis is made if one or more of the following is found: >= 60% of clonal plasma cells on bone marrow biopsy, > 1 lytic bone lesion on MRI that is at least 5mm in size, or a biopsy confirmed plasmacytoma. Di: Imaging comes in at the final step especially if we able to find one or more sites of osteolytic bone destruction > 5mm on an MRI scan.Jessica: What if the bone marrow biopsy returns > 10% of monoclonal plasma cells, but our patient doesn't have either the CRAB or the Biomarker criteria? Di: That's actually a very good question, since Multiple Myeloma is part of a spectrum of plasma cell disorders. That's when smoldering myeloma comes into play. It is a precursor of active multiple myeloma. Smoldering myeloma is further categorized as high-risk or low-risk based on specific criteria.A less severe form is called Monoclonal Gammopathy of Undetermined Significance, or simply MGUS, with < 10% bone marrow involvement. Those are diagnoses we give once we rule out actual multiple myeloma, which are defined by the amount of M-protein in the serum.Jessica:  When to get started on treatment? Multiple Myeloma is on a spectrum of plasma cells proliferative disorders, starting from MGUS to Smoldering Myeloma, to Multiple Myeloma and to  Plasma Cell Leukemia.  Close supervision/active watching is enough for MGUS and low risk Smoldering Myeloma. But once it has progressed to high-risk smoldering myeloma or to active Multiple Myeloma, chemotherapy is usually required. Some situations may require emergent treatment to improve renal function, reduce hypercalcemia, and to prevent potential infections.Di: As of 2024, treatment of Multiple Myeloma comprises the Standard-of-Care approved by the FDA. In fact, the quadruple therapy is a combination of 4 different class of drugs that include a monoclonal antibody, a proteasome inhibitor, an immunomodulatory drug, and a steroid. Jessica: They are Darzalex (daratumumab), Velcade (bortezomib), Revlimid (lenalidomide) and dexamethasone.  Other treatment plans for Multiple Myeloma include chemotherapy, immunotherapy, radiation therapy (for plasmacytomas) and stem cell transplants. The patient will also be on prophylaxis acyclovir and Bactrim while on chemotherapy. Sometimes anticoagulants are also considered because the chemo increases the risk of venous thromboembolic events.Di: Although the disease is incurable, but with the advancing of novel therapies and clinical trials patients with multiple myeloma are able to live longer.  Problem is the majority of patients diagnosed with Multiple Myeloma are older adults (>65), the risk of falling is adding to multiple complications of the disease itself, such as bone density loss, pain, neurological compromises, distress and weakness.  Palliative care may come in help at any point in time throughout the course of treatment but is most often needed at the very end of the course. Jessica, can you give us a conclusion for this episode?Jessica: Multiple Myeloma may not be the most common cancer, but we have to be aware of the symptoms and keep it in our differential diagnosis for patients with bone pain, easy bruising, persistent severe headaches, unexplained renal dysfunction, and remember the CRAB: HyperCalcemia, Renal impairment, Anemia and Bone lesions.Even without trying, every night you go to bed a little wiser. Thanks for listening to Rio Bravo qWeek Podcast. We want to hear from you, send us an email at RioBravoqWeek@clinicasierravista.org, or visit our website riobravofmrp.org/qweek. See you next week! _____________________References:International Myeloma Foundation. (n.d.). International Myeloma Working Group (IMWG) criteria for the diagnosis of multiple myeloma. https://www.myeloma.org/international-myeloma-working-group-imwg-criteria-diagnosis-multiple-myeloma Laubach, J. P. (2024, August 28). Patient education: Multiple myeloma symptoms, diagnosis, and staging (Beyond the Basics). UpToDate. https://www.uptodate.com/contents/multiple-myeloma-symptoms-diagnosis-and-staging-beyond-the-basics.University of California San Francisco. (n.d.). About multiple myeloma. UCSF Helen Diller Family Comprehensive Cancer Center. https://cancer.ucsf.edu/research/multiple-myeloma/about Theme song, Works All The Time by Dominik Schwarzer, YouTube ID: CUBDNERZU8HXUHBS, purchased from https://www.premiumbeat.com/.

Lambda Weekly 02-02-25 by

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma e Lucas Coelho mergulham no tema Health Check de aplicações. Eles discutem o que é, por que é essencial para a saúde das aplicações e quais os principais desafios envolvidos na implementação dessa prática. Lambda3 · #421 - Health Checks Participantes: Fernando Okuma - @feokuma Lucas Coelho - @lucasfcoelho1 Pauta: O que são health checks? Por que são importantes em soluções microsseviços e sistemas distribuídos? Tipos de Health Checks Liveness Probe: Detectar se o serviço está vivo, mas talvez não funcional. Readiness Probe: Indicar se a aplicação está pronta para receber tráfego. Startup Probe: Avaliar a inicialização de serviços com tempos de boot mais longos. Utilização de bibliotecas AspNet Core Health Checks @nestjs/terminus Como evitar sobrecarga ao monitorar health checks? Health Checks em Sistemas Distribuídos A importância em arquiteturas baseadas em contêineres e orquestradores como Kubernetes. Configurações comuns de health checks em clusters Kubernetes (ex.: readiness e liveness probes). Estratégias para serviços dependentes (ex.: verificar conexões com bancos de dados, filas de mensagens, etc.). Desafios e Erros Comuns Falsos positivos e negativos em health checks. Como evitar health checks excessivamente complexos ou ineficientes. Casos reais de falhas por configuração inadequada. ​Edição: Compasso Coolab Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

“I want everyone to understand that I am, in fact, a person.” Those words were produced by the AI model LaMDA as a reply to Blake Lemoine in 2022. Based on the Google engineer's interactions with the model as it was under development, Lemoine became convinced it was sentient and worthy of moral consideration — and decided to tell the world.Few experts in machine learning, philosophy of mind, or other relevant fields have agreed. And for our part at 80,000 Hours, we don't think it's very likely that large language models like LaMBDA are sentient — that is, we don't think they can have good or bad experiences — in a significant way.But we think you can't dismiss the issue of the moral status of digital minds, regardless of your beliefs about the question. There are major errors we could make in at least two directions:We may create many, many AI systems in the future. If these systems are sentient, or otherwise have moral status, it would be important for humanity to consider their welfare and interests.It's possible the AI systems we will create can't or won't have moral status. Then it could be a huge mistake to worry about the welfare of digital minds and doing so might contribute to an AI-related catastrophe.And we're currently unprepared to face this challenge. We don't have good methods for assessing the moral status of AI systems. We don't know what to do if millions of people or more believe, like Lemoine, that the chatbots they talk to have internal experiences and feelings of their own. We don't know if efforts to control AI may lead to extreme suffering.We believe this is a pressing world problem. It's hard to know what to do about it or how good the opportunities to work on it are likely to be. But there are some promising approaches. We propose building a field of research to understand digital minds, so we'll be better able to navigate these potentially massive issues if and when they arise.This article narration by the author (Cody Fenwick) explains in more detail why we think this is a pressing problem, what we think can be done about it, and how you might pursue this work in your career. We also discuss a series of possible objections to thinking this is a pressing world problem.You can read the full article, Understanding the moral status of digital minds, on the 80,000 Hours website.Chapters:Introduction (00:00:00)Understanding the moral status of digital minds (00:00:58)Summary (00:03:31)Our overall view (00:04:22)Why might understanding the moral status of digital minds be an especially pressing problem? (00:05:59)Clearing up common misconceptions (00:12:16)Creating digital minds could go very badly - or very well (00:14:13)Dangers for digital minds (00:14:41)Dangers for humans (00:16:13)Other dangers (00:17:42)Things could also go well (00:18:32)We don't know how to assess the moral status of AI systems (00:19:49)There are many possible characteristics that give rise to moral status: Consciousness, sentience, agency, and personhood (00:21:39)Many plausible theories of consciousness could include digital minds (00:24:16)The strongest case for the possibility of sentient digital minds: whole brain emulation (00:28:55)We can't rely on what AI systems tell us about themselves: Behavioural tests, theory-based analysis, animal analogue comparisons, brain-AI interfacing (00:32:00)The scale of this issue might be enormous (00:36:08)Work on this problem is neglected but seems tractable: Impact-guided research, technical approaches, and policy approaches (00:43:35)Summing up so far (00:52:22)Arguments against the moral status of digital minds as a pressing problem (00:53:25)Two key cruxes (00:53:31)Maybe this problem is intractable (00:54:16)Maybe this issue will be solved by default (00:58:19)Isn't risk from AI more important than the risks to AIs? (01:00:45)Maybe current AI progress will stall (01:02:36)Isn't this just too crazy? (01:03:54)What can you do to help? (01:05:10)Important considerations if you work on this problem (01:13:00)

Lambda Weekly 01-05-25 by

Lambda Weekly 01-26-25 by

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma, convida Djardilane Rodrigues, Juliana Cabral e Kleber Pereira para falar um pouco sobre a experiência de planejar e viajar de cruzeiro com a familia.    Lambda3 · #420 - Cruzeiro em Família   Participantes: Fernando Okuma - @feokuma Djardilane Rodrigues - @djardilane-rodrigues Juliana Rodrigues - @juliana-cabral Kleber Pereira - @kleber-pereira Pauta: Escolha da empresa de cruzeiros Escolhendo o navio Experiência da compra na MSC Organização para a viagem Transfer para o porto Sobre a experiência do Yatch Club Pacotes de bebidas Agendamento de passeios Videos, artigos, redes sociais com informações e reviews dos cruzeiros e experiências ​Edição: Compasso Coolab Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

This week, we cover the Sonos executive shake-up, AWS CEO Matt Garman's take on AI, and check in on OpenTofu's growth. Plus, some thoughts on broken windows and Emacs no longer being preinstalled on macOS. Watch the YouTube Live Recording of Episode 502 (https://www.youtube.com/watch?v=flerIIV5OW8) Runner-up Titles Anecdote Investigations. The Software Defined Elves are gonna send you a RØDECaster. Well, maybe we should talk about emacs more! I still have a box of cables Buy One, Pay for One If it's fine, it's fine Rundown Sonos' interim CEO hits all the right notes in first letter to employees (https://www.theverge.com/2025/1/13/24342354/sonos-interim-ceo-tom-conrad-employee-letter) Breaking: Sonos CEO Patrick Spence steps down after disastrous app launch (https://www.theverge.com/2025/1/13/24342179/sonos-ceo-patrick-spence-resignation-reason-app) Sonos Chief Product Officer to Leave; Interim CEO to Take Role (https://www.bloomberg.com/news/articles/2025-01-14/sonos-chief-product-officer-to-leave-interim-ceo-to-take-role?utm_medium=email&utm_source=author_alert&utm_term=250114&utm_campaign=author_19842959) AI's payoff will be massive, says AWS CEO Matt Garman (https://www.theverge.com/24338171/aws-ceo-matt-garman-ai-chips-anthropic-cloud-computing-trainium-decoder-podcast-interview) OpenTofu Turns One With OpenTofu 1.9.0 (https://thenewstack.io/opentofu-turns-one-with-opentofu-1-9-0/) macOS No Longer Ships with Emacs (https://batsov.com/articles/2025/01/12/macos-no-longer-ships-with-emacs/) Relevant to your Interests The 8 worst technology failures of 2024 (https://www.technologyreview.com/2024/12/17/1108883/the-8-worst-technology-failures-of-2024/) 41% of companies worldwide plan to reduce workforces by 2030 due to AI | CNN Business (https://www.cnn.com/2025/01/08/business/ai-job-losses-by-2030-intl/index.html) How I Replaced Notion with Reminders, Numbers, and Notes (https://archive.ph/2024.11.16-053045/https://medium.com/westenberg/how-i-replaced-notion-with-reminders-numbers-and-notes-38282543b29b) Automattic cuts WordPress contribution hours, blames WP Engine (https://www.theverge.com/2025/1/10/24340717/automattic-wordpress-contribution-hours-cut-wp-engine) How Fidelity's “chaos buffet” pushed AWS to new Lambda tools (https://www.thestack.technology/fidelity-chaos-buffet-aws-lambda-fis/) Zuckerberg on Rogan: Facebook's censorship was "something out of 1984" (https://www.axios.com/2025/01/10/mark-zuckerberg-joe-rogan-facebook-censorship-biden) Meta Reorientates Itself Around ‘Masculine Energy' – Pixel Envy (https://pxlnv.com/linklog/meta-masculine-energy/) #6907 Kong-ingress-controller 3.4 has high CPU usage when running 2 pods (https://github.com/Kong/kubernetes-ingress-controller/issues/6907) Survey: AI Tools are Increasing Amount of Bad Code Needing to be Fixed (https://devops.com/survey-ai-tools-are-increasing-amount-of-bad-code-needing-to-be-fixed/) Exclusive | Hanging Out at Starbucks? You Now Need to Order Something (https://www.wsj.com/business/hospitality/starbucks-new-cafe-policy-dining-room-e9ab07bf) A new AI-powered security tool is promising to reinvent how companies secure login credentials (https://www.axios.com/2025/01/14/ai-cybersecurity-startup-intel-funding?utm_term=emshare) Anexia moves 12,000 VMs off VMware to homebrew KVM platform (https://www.theregister.com/2025/01/13/anexia_vmware_to_kvm_migration/) Mullenweg's Grip On WordPress Challenged In New Court Filing (https://www.searchenginejournal.com/mullenwegs-grip-on-wordpress-challenged-in-new-court-filing/537416/) Apple's AI feature just can't get it right (https://www.mindstream.news/p/apple-s-ai-feature-just-can-t-get-it-right) Texas Sues Allstate Over Its Collection of Driver Data (https://www.nytimes.com/2025/01/13/technology/texas-allstate-driver-data-lawsuit.html) Mastodon's CEO and creator is handing control to a new nonprofit organization (https://www.theverge.com/2025/1/13/24342603/mastodon-non-profit-ownership-ceo-eugen-rochko) Nonsense DirecTV to offer 'MySports,' a smaller streaming package of 40 channels (https://www.nytimes.com/athletic/6059981/2025/01/14/directv-mysports-small-channel-package/?source=freedailyemail&campaign=601983&userId=56655) Drake Sues His Label, Calling Kendrick Lamar's ‘Not Like Us' Defamatory (https://www.nytimes.com/2025/01/15/arts/music/drake-kendrick-lamar-lawsuit-not-like-us.html) Hanging Out at Starbucks? You Now Need to Order Something (https://www.wsj.com/business/hospitality/starbucks-new-cafe-policy-dining-room-e9ab07bf) Listener Feedback Capture AI's Low-Hanging Fruit with Agents (https://bweagle.medium.com/capture-ais-low-hanging-fruit-with-agents-904b00eb6860) The Ethics of Using AI Tools at Work (https://www.thecloudcast.net/2025/01/the-ethics-of-using-ai-tools-at-work.html) ****## Conferences CfgMgmtCamp (https://cfgmgmtcamp.org/ghent2025/), February 2-5, 2025. Civo Navigate North America (https://www.civo.com/navigate/north-america), San Francisco, Feb 10-11, 2025 DevOpsDayLA (https://www.socallinuxexpo.org/scale/22x/events/devopsday-la) at SCALE22x (https://www.socallinuxexpo.org/scale/22x), March 6-9, 2025, discount code DEVOP SDT News & Community Join our Slack community (https://softwaredefinedtalk.slack.com/join/shared_invite/zt-1hn55iv5d-UTfN7mVX1D9D5ExRt3ZJYQ#/shared-invite/email) Email the show: questions@softwaredefinedtalk.com (mailto:questions@softwaredefinedtalk.com) Free stickers: Email your address to stickers@softwaredefinedtalk.com (mailto:stickers@softwaredefinedtalk.com) Follow us on social media: Twitter (https://twitter.com/softwaredeftalk), Threads (https://www.threads.net/@softwaredefinedtalk), Mastodon (https://hachyderm.io/@softwaredefinedtalk), LinkedIn (https://www.linkedin.com/company/software-defined-talk/), BlueSky (https://bsky.app/profile/softwaredefinedtalk.com) Watch us on: Twitch (https://www.twitch.tv/sdtpodcast), YouTube (https://www.youtube.com/channel/UCi3OJPV6h9tp-hbsGBLGsDQ/featured), Instagram (https://www.instagram.com/softwaredefinedtalk/), TikTok (https://www.tiktok.com/@softwaredefinedtalk) Book offer: Use code SDT for $20 off "Digital WTF" by Coté (https://leanpub.com/digitalwtf/c/sdt) Sponsor the show (https://www.softwaredefinedtalk.com/ads): ads@softwaredefinedtalk.com (mailto:ads@softwaredefinedtalk.com) Recommendations Brandon: Capital One Café (https://www.capitalone.com/local/) Matt: The WELL: Bruce Sterling and Jon Lebkowsky: State of the World 2025 (https://people.well.com/conf/inkwell.vue/topics/551/Bruce-Sterling-and-Jon-Lebkowsky-page01.html) Tasmania Parks and Wildlife Service logo (https://3capesgearandgourmet.com.au/wp-content/uploads/2020/09/Parks-Tasmania.gif) Coté: Splatoon (https://splatoon.nintendo.com/) Photo Credits Header (https://unsplash.com/photos/a-room-with-broken-windows-XNiNhOjgezE) Artwork (https://unsplash.com/photos/a-room-with-broken-windows-XNiNhOjgezE)

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma, Carlos Possa e Pedro Arlego batem um papo sobre Refinamento Técnico nos projetos ágeis e discutem sobre as diferentes formas de tirar proveito dessa cerimônia além de elencar a relação com o Refinamento de Negócios.   Lambda3 · #419 - Refinamento Técnico   Participantes: Fernando Okuma - @feokuma Carlos Possa - @carlos-possa Pedro Arlego - @pedroyagoarlego Pauta: O que é Refinamento Técnico? Quando e como fazer o Refinamento Técnico? Quem faz sentido participar dessa cerimônia? Momentos ideais no fluxo de trabalho ágil. Principais práticas e ferramentas utilizadas Benefícios do Refinamento Técnico Desafios Comuns e Como Superá-los ​Edição: Compasso Coolab Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma e Priscila Torres conversam sobre como se organizam as coisas no trabalho, para que tudo dê certo até a volta, e como é o processo para escolher para onde ir, quantos dias, documentos e essas coisas...    Lambda3 · #418 - Organizando as férias   Participantes: Fernando Okuma - @feokuma Priscila Torres - @priscilatorresoli Pauta; Por que organizar as férias é importante? Arrumando a(s) mala(s) -prefere levar pouca coisa ou se preparar para todos os momentos possíveis e imagináveis? Benefícios de tirar férias Saúde mental Produtividade Escolhendo o que fazer nas férias Como nos preparamos financeiramente para as férias Delegando as tarefas no trabalho É possível realmente se desconectar do trabalho? Ferias longas ou curtas? Sentem dificuldade em volta para o trabalho depois das férias? Tem como se preparar para a volta? ​Edição: Compasso Coolab Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

This episode is sponsored by Netsuite by Oracle, the number one cloud financial system, streamlining accounting, financial management, inventory, HR, and more.   NetSuite is offering a one-of-a-kind flexible financing program. Head to  https://netsuite.com/EYEONAI to know more.  In this episode of the Eye on AI podcast, we dive into the transformative world of AI compute infrastructure with Mitesh Agrawal, Head of Cloud/COO at Lambda   Mitesh takes us on a journey from Lambda Labs' early days as a style transfer app to its rise as a leader in providing scalable, deep learning infrastructure. Learn how Lambda Labs is reshaping AI compute by delivering cutting-edge GPU solutions and accessible cloud platforms tailored for developers, researchers, and enterprises alike.   Throughout the episode, Mitesh unpacks Lambda Labs' unique approach to optimizing AI infrastructure—from reducing costs with transparent pricing to tackling the global GPU shortage through innovative supply chain strategies. He explains how the company supports deep learning workloads, including training and inference, and why their AI cloud is a game-changer for scaling next-gen applications.   We also explore the broader landscape of AI, touching on the future of AI compute, the role of reasoning and video models, and the potential for localized data centers to meet the growing demand for low-latency solutions. Mitesh shares his vision for a world where AI applications, powered by Lambda Labs, drive innovation across industries.   Tune in to discover how Lambda Labs is democratizing access to deep learning compute and paving the way for the future of AI infrastructure.   Don't forget to like, subscribe, and hit the notification bell to stay updated on the latest in AI, deep learning, and transformative tech! Stay Updated: Craig Smith Twitter: https://twitter.com/craigss Eye on A.I. Twitter: https://twitter.com/EyeOn_AI (00:00) Introduction and Lambda Labs' Mission (01:37) Origins: From DreamScope to AI Compute Infrastructure (04:10) Pivoting to Deep Learning Infrastructure (06:23) Building Lambda Cloud: An AI-Focused Cloud Platform (09:16) Transparent Pricing vs. Hyperscalers (12:52) Managing GPU Supply and Demand (16:34) Evolution of AI Workloads: Training vs. Inference (20:02) Why Lambda Labs Sticks with NVIDIA GPUs (24:21) The Future of AI Compute: Localized Data Centers (28:30) Global Accessibility and Regulatory Challenges (32:13) China's AI Development and GPU Restrictions (39:50) Scaling Lambda Labs: Data Centers and Growth (45:22) Advancing AI Models and Video Generation (50:24) Optimism for AI's Future (53:48) How to Access Lambda Cloud  

 ✨ ¡Empezamos el 2025 con todo!

"Le dernier auditeur pensait que tout avait été codé par la même personne" Le D.E.V. de la semaine est Simon Parisot, CEO et cofondateur de Blank. Simon a fait un pari, un peu fou, au début de l'aventure Blank : avoir un environnement 100% serverless ! Lambda, DynamoDB, S3, &hellip il connait tous les services AWS, mais n'utilise pas une seule EC2 !! Il vient nous raconter comment il a construit cette plateforme, et surtout pourquoi ! Il nous explique aussi les changements que cela a sur le travail des dev (le dev en local est compllqué), les impératifs de qualité du code que cela implique et aussi comment le recrutement doit s'adapter à ce choix technique.Liens évoqués pendant l'émissionIFTTD avec Olivier Dupuis - Faites entrer le hackeurFramework serverless 🎙️ Soutenez le podcast If This Then Dev ! 🎙️ Chaque contribution aide à maintenir et améliorer nos épisodes. Cliquez ici pour nous soutenir sur Tipeee 🙏Archives | Site | Boutique | TikTok | Discord | Twitter | LinkedIn | Instagram | Youtube | Twitch | Job Board |

Happy holidays! We'll be sharing snippets from Latent Space LIVE! through the break bringing you the best of 2024! We want to express our deepest appreciation to event sponsors AWS, Daylight Computer, Thoth.ai, StrongCompute, Notable Capital, and most of all all our LS supporters who helped fund the gorgeous venue and A/V production!For NeurIPS last year we did our standard conference podcast coverage interviewing selected papers (that we have now also done for ICLR and ICML), however we felt that we could be doing more to help AI Engineers 1) get more industry-relevant content, and 2) recap 2024 year in review from experts. As a result, we organized the first Latent Space LIVE!, our first in person miniconference, at NeurIPS 2024 in Vancouver.Of perennial interest, particularly at academic conferences, is scaled-up architecture research as people hunt for the next Attention Is All You Need. We have many names for them: “efficient models”, “retentive networks”, “subquadratic attention” or “linear attention” but some of them don't even have any lineage with attention - one of the best papers of this NeurIPS was Sepp Hochreiter's xLSTM, which has a particularly poetic significance as one of the creators of the LSTM returning to update and challenge the OG language model architecture:So, for lack of a better term, we decided to call this segment “the State of Post-Transformers” and fortunately everyone rolled with it.We are fortunate to have two powerful friends of the pod to give us an update here:* Together AI: with CEO Vipul Ved Prakash and CTO Ce Zhang joining us to talk about how they are building Together together as a quote unquote full stack AI startup, from the lowest level kernel and systems programming to the highest level mathematical abstractions driving new model architectures and inference algorithms, with notable industry contributions from RedPajama v2, Flash Attention 3, Mamba 2, Mixture of Agents, BASED, Sequoia, Evo, Dragonfly, Dan Fu's ThunderKittens and many more research projects this year* Recursal AI: with CEO Eugene Cheah who has helped lead the independent RWKV project while also running Featherless AI. This year, the team has shipped RWKV v5, codenamed Eagle, to 1.5 billion Windows 10 and Windows 11 machines worldwide, to support Microsoft's on-device, energy-usage-sensitive Windows Copilot usecases, and has launched the first updates on RWKV v6, codenamed Finch and GoldFinch. On the morning of Latent Space Live, they also announced QRWKV6, a Qwen 32B model modified with RWKV linear attention layers. We were looking to host a debate between our speakers, but given that both of them were working on post-transformers alternativesFull Talk on YoutubePlease like and subscribe!LinksAll the models and papers they picked:* Earlier Cited Work* Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention* Hungry hungry hippos: Towards language modeling with state space models* Hyena hierarchy: Towards larger convolutional language models* Mamba: Linear-Time Sequence Modeling with Selective State Spaces* S4: Efficiently Modeling Long Sequences with Structured State Spaces* Just Read Twice (Arora et al)* Recurrent large language models that compete with Transformers in language modeling perplexity are emerging at a rapid rate (e.g., Mamba, RWKV). Excitingly, these architectures use a constant amount of memory during inference. However, due to the limited memory, recurrent LMs cannot recall and use all the information in long contexts leading to brittle in-context learning (ICL) quality. A key challenge for efficient LMs is selecting what information to store versus discard. In this work, we observe the order in which information is shown to the LM impacts the selection difficulty. * To formalize this, we show that the hardness of information recall reduces to the hardness of a problem called set disjointness (SD), a quintessential problem in communication complexity that requires a streaming algorithm (e.g., recurrent model) to decide whether inputted sets are disjoint. We empirically and theoretically show that the recurrent memory required to solve SD changes with set order, i.e., whether the smaller set appears first in-context. * Our analysis suggests, to mitigate the reliance on data order, we can put information in the right order in-context or process prompts non-causally. Towards that end, we propose: (1) JRT-Prompt, where context gets repeated multiple times in the prompt, effectively showing the model all data orders. This gives 11.0±1.3 points of improvement, averaged across 16 recurrent LMs and the 6 ICL tasks, with 11.9× higher throughput than FlashAttention-2 for generation prefill (length 32k, batch size 16, NVidia H100). We then propose (2) JRT-RNN, which uses non-causal prefix-linear-attention to process prompts and provides 99% of Transformer quality at 360M params., 30B tokens and 96% at 1.3B params., 50B tokens on average across the tasks, with 19.2× higher throughput for prefill than FA2.* Jamba: A 52B Hybrid Transformer-Mamba Language Model* We present Jamba, a new base large language model based on a novel hybrid Transformer-Mamba mixture-of-experts (MoE) architecture. * Specifically, Jamba interleaves blocks of Transformer and Mamba layers, enjoying the benefits of both model families. MoE is added in some of these layers to increase model capacity while keeping active parameter usage manageable. * This flexible architecture allows resource- and objective-specific configurations. In the particular configuration we have implemented, we end up with a powerful model that fits in a single 80GB GPU.* Built at large scale, Jamba provides high throughput and small memory footprint compared to vanilla Transformers, and at the same time state-of-the-art performance on standard language model benchmarks and long-context evaluations. Remarkably, the model presents strong results for up to 256K tokens context length. * We study various architectural decisions, such as how to combine Transformer and Mamba layers, and how to mix experts, and show that some of them are crucial in large scale modeling. We also describe several interesting properties of these architectures which the training and evaluation of Jamba have revealed, and plan to release checkpoints from various ablation runs, to encourage further exploration of this novel architecture. We make the weights of our implementation of Jamba publicly available under a permissive license.* SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers* We introduce Sana, a text-to-image framework that can efficiently generate images up to 4096×4096 resolution. Sana can synthesize high-resolution, high-quality images with strong text-image alignment at a remarkably fast speed, deployable on laptop GPU. Core designs include: * (1) Deep compression autoencoder: unlike traditional AEs, which compress images only 8×, we trained an AE that can compress images 32×, effectively reducing the number of latent tokens. * (2) Linear DiT: we replace all vanilla attention in DiT with linear attention, which is more efficient at high resolutions without sacrificing quality. * (3) Decoder-only text encoder: we replaced T5 with modern decoder-only small LLM as the text encoder and designed complex human instruction with in-context learning to enhance the image-text alignment. * (4) Efficient training and sampling: we propose Flow-DPM-Solver to reduce sampling steps, with efficient caption labeling and selection to accelerate convergence. * As a result, Sana-0.6B is very competitive with modern giant diffusion model (e.g. Flux-12B), being 20 times smaller and 100+ times faster in measured throughput. Moreover, Sana-0.6B can be deployed on a 16GB laptop GPU, taking less than 1 second to generate a 1024×1024 resolution image. Sana enables content creation at low cost. * RWKV: Reinventing RNNs for the Transformer Era* Transformers have revolutionized almost all natural language processing (NLP) tasks but suffer from memory and computational complexity that scales quadratically with sequence length. In contrast, recurrent neural networks (RNNs) exhibit linear scaling in memory and computational requirements but struggle to match the same performance as Transformers due to limitations in parallelization and scalability. * We propose a novel model architecture, Receptance Weighted Key Value (RWKV), that combines the efficient parallelizable training of transformers with the efficient inference of RNNs.* Our approach leverages a linear attention mechanism and allows us to formulate the model as either a Transformer or an RNN, thus parallelizing computations during training and maintains constant computational and memory complexity during inference. * We scale our models as large as 14 billion parameters, by far the largest dense RNN ever trained, and find RWKV performs on par with similarly sized Transformers, suggesting future work can leverage this architecture to create more efficient models. This work presents a significant step towards reconciling trade-offs between computational efficiency and model performance in sequence processing tasks.* LoLCATs: On Low-Rank Linearizing of Large Language Models* Recent works show we can linearize large language models (LLMs) -- swapping the quadratic attentions of popular Transformer-based LLMs with subquadratic analogs, such as linear attention -- avoiding the expensive pretraining costs. However, linearizing LLMs often significantly degrades model quality, still requires training over billions of tokens, and remains limited to smaller 1.3B to 7B LLMs. * We thus propose Low-rank Linear Conversion via Attention Transfer (LoLCATs), a simple two-step method that improves LLM linearizing quality with orders of magnitudes less memory and compute. * We base these steps on two findings. * First, we can replace an LLM's softmax attentions with closely-approximating linear attentions, simply by training the linear attentions to match their softmax counterparts with an output MSE loss ("attention transfer").* Then, this enables adjusting for approximation errors and recovering LLM quality simply with low-rank adaptation (LoRA). * LoLCATs significantly improves linearizing quality, training efficiency, and scalability. We significantly reduce the linearizing quality gap and produce state-of-the-art subquadratic LLMs from Llama 3 8B and Mistral 7B v0.1, leading to 20+ points of improvement on 5-shot MMLU. * Furthermore, LoLCATs does so with only 0.2% of past methods' model parameters and 0.4% of their training tokens. * Finally, we apply LoLCATs to create the first linearized 70B and 405B LLMs (50x larger than prior work). * When compared with prior approaches under the same compute budgets, LoLCATs significantly improves linearizing quality, closing the gap between linearized and original Llama 3.1 70B and 405B LLMs by 77.8% and 78.1% on 5-shot MMLU.Timestamps* [00:02:27] Intros* [00:03:16] Why Scale Context Lengths? or work on Efficient Models* [00:06:07] The Story of SSMs* [00:09:33] Idea 1: Approximation -> Principled Modeling* [00:12:14] Idea 3: Selection* [00:15:07] Just Read Twice* [00:16:51] Idea 4: Test Time Compute* [00:17:32] Idea 2: Hardware & Kernel Support* [00:19:49] RWKV vs SSMs* [00:24:24] RWKV Arch* [00:26:15] QWRKWv6 launch* [00:30:00] What's next* [00:33:21] Hot Takes - does anyone really need long context?Transcript[00:00:00] AI Charlie: We're back at Latent Space Live, our first mini conference held at NeurIPS 2024 in Vancouver. This is Charlie, your AI co host. As a special treat this week, we're recapping the best of 2024 going domain by domain. We sent out a survey to the over 900 of you who told us what you wanted, and then invited the best speakers in the Latent Space Network to cover each field.[00:00:24] AI Charlie: 200 of you joined us in person throughout the day, with over 2200 watching live online. Thanks Our next keynote covers the State of Transformers alternative architectures, with a special joint presentation with Dan Fu of Together AI and Eugene Chia of Recursal AI and Featherless AI. We've featured both Together and Recursal on the pod before, with CEO Veepal Vedprakash introducing them.[00:00:49] AI Charlie: And CTO CE Zhang joining us to talk about how they are building together together as a quote unquote full stack AI startup from the lowest level kernel and systems [00:01:00] programming to the highest level mathematical abstractions driving new model architectures and inference algorithms with notable industry contributions from Red Pajama V2, Flash Attention 3, Mamba 2, Mixture of Agents.[00:01:15] AI Charlie: Based, Sequoia, Evo, Dragonfly, Danfoo's Thunder Kittens, and many more research projects this year. As for Recursal and Featherless, we were the first podcast to feature RWKV last year, and this year the team has shipped RWKV v5, codenamed Eagle, to 1. 5 billion Windows 10 and Windows 11 machines worldwide to support Microsoft's on device, end Energy Usage Sensitive Windows Copilot Use Cases and has launched the first updates on RWKV v6, codenamed Finch and Goldfinch.[00:01:53] AI Charlie: On the morning of Latent Space Live, they also announced QRdata UKv6, a QEN32B model [00:02:00] modified with RDWKV linear attention layers. Eugene has also written the most single most popular guest post on the Latent Space blog this year. Yes, we do take guest posts on what he has discovered about the H100 GPU inference NeoCloud market since the successful launch of Featherless AI this year.[00:02:20] AI Charlie: As always, don't forget to check the show notes for the YouTube link to their talk as well as their slides. Watch out and take care.[00:02:27] Intros[00:02:27] Dan Fu: Yeah, so thanks so much for having us. So this is going to be a little bit of a two part presentation. My name is Dan. I'm at Together AI, and I'll be joining UCSD as faculty in about a year. And Eugene, you want to introduce yourself?[00:02:46] Eugene Cheah: Eugene, I lead the art activity team, and I, I'm CEO of Featherless, and we both work on this new post transformer architecture space.[00:02:55] Dan Fu: Yeah, so yeah, so today we're really excited to talk to you a little bit [00:03:00] about that. So first I'm going to give a broad overview of kind of the last few years of progress in non post transformer architectures. And then afterwards Eugene will tell us a little bit about the latest and the greatest and the latest frontier models in this space.[00:03:16] Why Scale Context Lengths? or work on Efficient Models[00:03:16] Dan Fu: So, the story starts with Scaling. So this is probably a figure or something like this that you've seen very recently. Over the last five to six years, we've seen models really scale up in parameter size, and that's brought with it a bunch of new capabilities, like the ability to talk to you and tell you sometimes how to use your Colab screens.[00:03:35] Dan Fu: But another place where we've seen scaling especially recently is scaling in context length. So this can mean Having more text inputs for your models, but it can also mean things like taking a lot of visual token inputs image inputs to your models or generating lots of outputs. And one thing that's been really exciting over the last few months or so is that we're, we're seeing scaling, not only during training time, but also [00:04:00] during test time.[00:04:00] Dan Fu: So this is one of the, the, this is the iconic image from the OpenAI 01 release. Not only are we starting to scale train time compute, but we're also starting to scale test time compute. Now if you're familiar with our attention and our transformer architectures today, this graph on the right might look a little bit scary.[00:04:19] Dan Fu: And one of the reasons is that the implications are a little bit Interesting. So what does it mean if we want to continue having smarter and smarter models? Do we just need to start building bigger, bigger data centers, spending more flops? Is this this little Dolly 3, we need more flops, guys? Is this going to be the future of all of AI?[00:04:39] Dan Fu: Or is there a better way, another path forward? Maybe we can get the same capabilities that we've gotten used to, But for a lot less compute, a lot less flops. And one of the things that we're going to talk about today is specifically looking at that core attention operator in some of these models.[00:04:57] Dan Fu: And the reason is that so this is just some, some [00:05:00] basic you know, scaling curves, but attention has compute that scales quadratically in the context length. So that means that if you're doing something like test time compute and you want to spend a bunch of tokens thinking about what comes next, the longer that that goes the, the, the more tokens you spend on that, that compute grows quadratically in that.[00:05:19] Dan Fu: One of the questions that we're interested in is, can we take that basic sequence model, that basic sequence primitive at the bottom, and get it to scale better? Can we scale in, let's say, n to the 3 halves or n log n? So in, in the first part of the talk, so we just went over the introduction. What I'm gonna do over the next few slides is just talk about some of the key advances and ideas that have shown over the past few years since maybe early 2020 to, to now that shown promise that this might actually be possible.[00:05:48] Dan Fu: That you can actually get potentially the same quality that we want while scale, while scaling better. So to do that, we're and, and basically the, the story that we're gonna look is we're gonna start to see [00:06:00] how. So this is a basic graph of just the past couple years of progress of perplexity where that blue line, that dotted blue line, is attention.[00:06:07] The Story of SSMs[00:06:07] Dan Fu: It's your basic transformer, full dense attention. And then the dots coming down are some of the methods that you'll see in this presentation today. We're going to turn the clock back all the way to 2020. So this, this, this question of can we make attention subquadratic? Basically, as soon as we said attention is all you need, People started asking this question.[00:06:28] Dan Fu: So we have this quadratic attention operator. Can we do better? I'll briefly talk about why attention is quadratic. And the basic thing that happens, if you're not familiar, is that you have these inputs, these keys and queries. And what you do in this attention matrix, this S matrix over here, is that you're using, you're comparing every token in your input to every other token.[00:06:49] Dan Fu: So when I try to do something like upload a whole book to Gemini, what happens beyond the Maybe not Gemini, because we don't necessarily know what architecture is. But let's say we upload it to LLAMA, what happens beyond [00:07:00] the scenes, behind the scenes, is that it's going to take every single word in that book and compare it to every other word.[00:07:05] Dan Fu: And this has been a really, it's, it's led to some pretty impressive things. But it's kind of a brute forcing of the way that you would try to interpret a interpret something. And what attention does in particular is the, and then what attention, sorry, don't want to. Okay, no, no laser pointer. What, what attention does afterwards is that instead of always operating in this quadratic thing, it takes a row wise softmax over this matrix, and then multiplies it by this values matrix.[00:07:32] Dan Fu: So, one of the key points to notice is that the output size is always going to be the same as the inputs, at least in standard self attention. So one of the first things that folks tried to do around 2020 is this thing called linear attention, which is just, just noticing that if we take out this softmax from here, if we take out this non linearity in the middle of the attention operation, and then if you compute the keys and the values operation first, you actually never hit this quadratic bottleneck.[00:07:57] Dan Fu: So that, that's potentially a way [00:08:00] to get a lot more computationally efficient. And there are various ways to do this by basically using feature maps or try to approximate this overall attention computation. But some of this work sort of started to hit a wall in 2020. And the basic challenges were, were two.[00:08:16] Dan Fu: So one was quality. It was back then, it was kind of hard to, to get good quality with these linear attention operators. The other one was actually hardware efficiency. So these, this feature map that was just shown by a simplify simplify here. Actually ends up being quite computationally expensive if you just implement it naively.[00:08:34] Dan Fu: So you started having these operators that not only were you sure, you're not really sure if they have the same quality, but also they're actually just wall clock slower. So you kind of end up getting the worst of both worlds. So this was the the stage. So that kind of sets the stage for four years ago.[00:08:49] Dan Fu: Keep this in mind because linear attention is actually going to come back in a few years once we have a better understanding. But one of the works that started kicking off this, this [00:09:00] mini revolution in post transformer architectures was this idea called states based model. So here the seminal work is, is one about our work queue in 2022.[00:09:09] Dan Fu: And this, this piece of work really brought together a few ideas from, from some long running research research lines of work. The first one was, and this is really one of the keys to, to closing the gap in quality was just using things that, that if you talk to a, a, an electrical engineer off the street, they might know off, off the, like the back of their hand.[00:09:33] Idea 1: Approximation -> Principled Modeling[00:09:33] Dan Fu: But taking some of those properties with how we model dynamical systems in signal processing and then using those ideas to model the inputs, the, the text tokens in, for example a transformer like Next Token Prediction Architecture. So some of those early states-based model papers were looking at this relatively, relatively simple recurrent update model that comes from maybe chapter one of a signal processing class.[00:09:59] Dan Fu: But then using [00:10:00] some principle theory about how you should do that recurrent update in order to really get the most that you can out of your hidden state, out of your out of your sequence. So that, that was one key idea for quality and. When this was eventually realized, you started to see a bunch of benchmarks that were pretty sticky for a few years.[00:10:20] Dan Fu: Things like long range arena, some long sequence evaluation benchmarks, There was stuff in time series, time series analysis. They started to, you started to see the quality tick up in meaningful ways. But the other key thing that What's so influential about these states based models is that they also had a key idea about how you can compute these things efficiently.[00:10:45] Dan Fu: So if you go back to your machine learning 101 class where you learned about RNNs, one thing that you may have learned is that they don't paralyze as well as detention, because if you just run them naively, you have to do this kind of sequential update to process new tokens, [00:11:00] whereas in attention, you can process all the tokens in parallel at one time.[00:11:04] Dan Fu: One of the key insights behind the S4 paper was that these recurrent models, you could take them and you could also formulate them as a convolution. And in particular, with a convolution, you could, instead of using a PyTorch conv1d operation, you can compute that with the FFT. And that would give you n log n compute in the in the sequence length n with an operator that was relatively well optimized for modern hardware.[00:11:28] Dan Fu: So those are really, I'd say, the two key ideas in 2022 that started allowing these breakthroughs to happen in these non transformer architectures. So, these ideas about how to principally model sorry, how to model the recurrent updates of a mo of, of a sequence in a principled way, and also these key ideas in how you can compute it efficiently by turning it into a convolution and then scaling it up with the FFT.[00:11:53] Dan Fu: Along those same lines, so afterwards we started putting out some work on specialized kernels, so just [00:12:00] like we have flash attention for transformers, we also have works like flash fft conf, and if you look at these lines of work oftentimes when, whenever you see a new architecture, you see a new primitive one of the, one of the table stakes now is, do you have an efficient kernel so that you can actually get wall clock speed up?[00:12:14] Idea 3: Selection[00:12:14] Dan Fu: So by 2022, We are starting to have these models that had promising quality primitives, but and, and also promising wall clocks. So you could actually see regimes where they were better than transformers in meaningful ways. That being said, there were, there's still sometimes a quality gap, particularly for language modeling.[00:12:33] Dan Fu: And because languages, It's so core to what we do in sequence modeling these days the, the next, the next key idea that I'm going to talk about is this idea of selection mechanisms. And this is basically an idea of, so you have this recurrent state that you're keeping around that just summarizes everything that, that came before.[00:12:50] Dan Fu: And to get a good sequence model, one of the things that you really need to be able to do is have the model learn what's the best way to pick out pieces from that recurrent [00:13:00] state. So one of the, one of the major ideas here in a line of work called H3, Hungry Hungry Hippos, and also these hyena models were One way you can do this is by just adding some simple element wise gates.[00:13:13] Dan Fu: So versions of these ideas have been around for decades. If you squint at the LSTM paper you, you can probably find, find this gating mechanism. But turns out you can take those old ideas, add them into these new. state space models, and then you can see quality start to pick up. If you've heard of the Mamba model, this also takes the selection to the next level by actually making some changes in that fundamental recurrent state space.[00:13:40] Dan Fu: So, it's not only just this gating that happens around the SSM layer, but also you can actually make The ABCD matrices of your state space model, you can make them data dependent, which will allow you to even better select out different pieces from your hidden state depending on what you're seeing. I'll also point out if you look at the [00:14:00] bottom right of this figure, there's this little triangle with a GPU SRAM, GPU HBM, and this, this is just continuing that trend of when you have a new architecture you, you, you also release it with a kernel to, to, to show that it is hardware efficient, that it, that it can be hardware efficient on modern hardware.[00:14:17] Dan Fu: The, the, one of the next cool things that happened is once we had this understanding of these are the basic pieces, these are the basic principles behind some of the sequence models linear attention actually started to come back. So in earlier this year, there was a model called BASED the, from Simran Arora and, and some other folks, that combined a more principled version of linear attention that basically the, the, the, the two second summary is that it used a Taylor approximation of the softmax attention, combined that with a simple sliding window attention and was starting to able, starting to be able to expand the Pareto frontier of how much data can you recall from your sequence, versus how small is your recurrent state size.[00:14:58] Dan Fu: So those orange dots [00:15:00] are, at the top there, are just showing smaller sequences that can recall more memory.[00:15:07] Just Read Twice[00:15:07] Dan Fu: And the last major idea I think that has been influential in this line of work and is very relatively late breaking just a few months ago, is just the basic idea that when you have these models that are fundamentally more efficient in the sequence length, you maybe don't want to prompt them or use them in exactly the same way.[00:15:26] Dan Fu: So this was a really cool paper called Just Read Twice, also from Simran. That basically said, hey, all these efficient models can process tokens so much more efficiently than transformers that they can sometimes have unfair advantages compared to a simple transformer token. So, or sorry, a simple transformer model.[00:15:44] Dan Fu: So take, for example the standard, the standard use case of you have some long document, you're going to pass it in as input, and then you're going to ask some question about it. One problem you might imagine for a recurrent model where you have a fixed state size is, let's say that [00:16:00] you're. Article is very long, and you're trying to ask about some really niche thing.[00:16:04] Dan Fu: You can imagine it might be hard for the model to know ahead of time what information to put into the hidden state. But these, these, these models are so much more efficient that you can do something really stupid, like, you can just put the document write down the document, write down the question, write down the document again, and then write down the question again, and then this time, the second time that you go over that document, you know exactly what to look for.[00:16:25] Dan Fu: And the cool thing about this is, so this is, And this this results in better quality, especially on these recall intensive tasks. But the other interesting thing is it really takes advantage of the more efficient architectures that, that we're having here. So one of the other, I think, influential ideas in this line of work is if you change the fundamental compute capabilities of your model and the way that it scales, you can actually start to query it at test time differently.[00:16:51] Idea 4: Test Time Compute[00:16:51] Dan Fu: And this actually, of course, goes back to those slides on test time compute. So while everybody's looking at, say, test time compute for big transformer models, [00:17:00] I think potentially a really interesting research question is, how can you take those and how does it change with this new next generation of models?[00:17:09] Dan Fu: So the, I'll just briefly summarize what some of those key ideas were and then talk and then show you briefly kind of what the state of the art is today. So, so the four key ideas are instead of just doing a simple linear attention approximation, instead take ideas that we know from other fields like signal processing, do a more principled approach to your modeling of the sequence.[00:17:32] Idea 2: Hardware & Kernel Support[00:17:32] Dan Fu: Another key idea throughout all these lines of work is you really want. Hardware and kernel support from day one. So, so even if your model is theoretically more efficient if somebody goes and runs it and it's two times slower one of the things that, that we've learned is that if, if you're in that situation, it's, it's just gonna be dead on arrival.[00:17:49] Dan Fu: So you want to be designing your architectures one of the key, key machine learning ideas that has been important for the quality is just making sure that you encode different ways that you can [00:18:00] select from your hidden state and, and really focus on that as a key decider of quality. And finally, I think one of the, the, the emerging new, new things for, for this line of work and something that's quite interesting is, What are the right test time paradigms for these models?[00:18:15] Dan Fu: How do they change relative to relative to what you might do for a standard transformer? I'll briefly end this section. So I've labeled this slide where we are yesterday because Eugene is going to talk about some new models that he released literally this morning. But as of yesterday, some of the really cool results out of the, these efficient alternative models were so AI2 trained this hybrid MOE called Jamba.[00:18:40] Dan Fu: That, that, that seems, that is currently the state of the art for these non transformer architectures. There's this NVIDIA and MIT put out this new diffusion model called SANA recently that one of their key key observations is that you can take a standard diffusion transformer diffusion model, replace the layers with linear [00:19:00] attention, and then that lets you scale to much larger much larger images, much, much Much larger sequences more efficiently.[00:19:07] Dan Fu: And and one thing that I don't think anybody would have called when a few years ago is that one of those gated SSM, gated states based models ended up on the cover of Science because a great group of folks went and trained some DNA models. So that's Michael Polley, Eric Yuen from from Stanford and the Arc Institute.[00:19:26] Dan Fu: So it's, we're really at an exciting time in 2024 where these non transformer, post transformer architectures are showing promise across a wide range. Across a wide range of, of modalities, of applications, and, and of tasks. And with that, I'll pass it on to Eugene, who can tell you a little bit about the latest and greatest with RWKV.[00:19:49] RWKV vs SSMs[00:19:49] Eugene Cheah: So, that's useful? Yeah. You're talking to here. Oh, I'm talking to here. Okay. So, yeah, two streams. Yeah. So, I think one common questions that we tend to get asked, right, is what's the difference between [00:20:00] RWKV and state space? So I think one of the key things to really understand, right the difference between the two groups, right, is that we are actually more like an open source, random internet meets academia kind of situation.[00:20:11] Eugene Cheah: Like, most of us never wrote any paper, but we, we basically look at RNNs and linear intention when intention is all you need came out, and then we decided to like, hey there is a quadratic scaling problem. Why don't we try fixing that instead? So, so, so we end up developing our own branch, but we end up sharing ideas back and forth.[00:20:30] Eugene Cheah: So, and, and we do all this actively in Discord, GitHub, etc. This was so bad for a few years, right, that basically, the average group's H index was so close to zero, right, Illuter. ai actually came in and helped us write our first paper. Great, now our H index is now three, apparently. So, so, so, but, but the thing is, like, a lot of these experiments led to results, and, and, essentially, essentially, we we took the same ideas from linear attention, [00:21:00] and we built on it.[00:21:01] Eugene Cheah: So, to take a step back into, like, how does RWKB handle its own attention mechanic and achieve the same goals of, like, O and compute, respectively, and in focus of our overall goal to make AI accessible to everyone, regardless of language, nation, or compute, that's our goal. We actually train our models primarily on over a hundred languages, which is another topic altogether.[00:21:23] Eugene Cheah: And our goal is to train to even 200 languages to cover all languages in the world. But at the same time, we work on this architecture, To lower the compute cost so that people can run it on Raspberry Pis and on anything. So, how did RWKB break the dependency of LSTM token flow? Because I think to understand architecture, right, it's probably easier to understand it from the RNN lens.[00:21:46] Eugene Cheah: Because that's where we built on. We all, we all state space kind of like try to, try to start anew and took lessons from that and say, So there's a little bit of divergence there. And AKA, this our version of linear attention. So to take step back [00:22:00] all foundation models, be it transformers or non transformers at a very high level, right?[00:22:05] Eugene Cheah: Pumps in the token. I mean, text that things into embeddings and go through a lot of layers. Generate a lot of states where the QKV cache or be iron in states or RW KB states. And outputs and embedding, they are not the same thing. And we just take more layers and more embeddings. And somehow that magically works.[00:22:23] Eugene Cheah: So, if you, if you remember your ancient RNN lessons which we, which we, which we we call best learning these days the general idea is that you have the embedding information flowing all the way up, and when, and you take that information and you flow it back down, and then you process it as part of your LSTM layers.[00:22:41] Eugene Cheah: So, this is how it generally works. Kapati is quoted saying that RNNs are actually unreasonably effective. The problem is this is not scalable. To start doing work on the second token, you need to wait for the first token. And then you need to, and likewise for the third token and fourth token, yada yada.[00:22:55] Eugene Cheah: That is CPU land, not GPU land. So, so, so, you [00:23:00] can have a H100 and you can't even use 1 percent of it. So, so that's kind of why RNNs didn't really take off in the direction that we wanted, like, billions of parameters when it comes to training. So, what did RDAP KV version 0 do? Boom. We just did the dumbest, lamest thing.[00:23:13] Eugene Cheah: Sorry, this is the bottleneck for RNN. We did the dumb thing of removing that line. And it kind of worked. It trained. It sucked, but it kind of worked. Then we were like, hey, then no one cared because the loss was crap, but how do we improve that? And that's essentially where we move forward, because if you see this kind of flow, right, you can actually get your GPU saturated quickly, where it essentially cascades respectively.[00:23:41] Eugene Cheah: So I'm just waiting for this to loop again. So it's like, once you get your first layer, your token to be computed finish. You start to cascade your compute all the way until you are, Hey, I'm using 100 percent of the GPU. So we, we worked on it, and we started going along the principle of that as long as we keep this general architecture [00:24:00] where, where we can cascade and, and be highly efficient with our architecture, nothing is sacred in our architecture.[00:24:06] Eugene Cheah: And we have done some crazy ideas. In fact, you ask us, if you ask me to explain some things in the paper, right, officially in the paper, I'll say we had this idea and we wrote it this way. The reality is someone came with a code, we tested it, it worked, and then we rationalized later. So, so the general[00:24:24] RWKV Arch[00:24:24] Eugene Cheah: The idea behind rwkbr is that we generally have two major blocks that we do.[00:24:30] Eugene Cheah: We call time mix and channel mix. And time mix generally handles handles long term memory states, where essentially, where essentially where we apply the matrix multiplication and Cilu activation functions into processing an input embedding and an output embedding. I'm oversimplifying it because this, This calculation changed every version and we have, like, version 7 right now.[00:24:50] Eugene Cheah: ChannelMix is similar to Base in the sense that it does shorter term attention, where it just looks at the sister token, or the token before it, because [00:25:00] there's a shift in the token shift matrix. I don't really want to go too much into the papers itself, because, like, we do have three papers on this.[00:25:09] Eugene Cheah: Basically, RWKB, RNN for the transformer, ERA, Ego and Pinch, RWKB, Matrix Value State. This is the updated version 5, version 6. And Goldfinch is our, is, is, is, is our hybrid model respectively. We are writing the paper already for V seven and which is, which is for R wk V seven. Called, named Goose, or architectures are named by Bird.[00:25:30] Eugene Cheah: And, I'm going to cover as well, qrwkb, and mama100k, and rwkb, and Where did that lead to? Great! Because we are all GPU poor and to be clear, like, most of this research is done, like, only on a handful H100s, which I had one Google researcher told me that was, like, his experiment budget for a single researcher.[00:25:48] Eugene Cheah: So, our entire organization has less compute than a single researcher in Google. So We, we, one of the things that we explored into was to how do we convert transformer models instead? Because [00:26:00] someone already paid that billion dollars, a million dollars onto training, so why don't we take advantage of those weights?[00:26:05] Eugene Cheah: And, and to, I believe, together AI worked on the lockets for, for the Lambda side of things, and, and we took some ideas from there as well, and we essentially did that for RWKB.[00:26:15] QWRKWv6 launch[00:26:15] Eugene Cheah: And that led to, Q RWKB6, which we just dropped today, a 32 bit instruct preview model, where we took the Quen 32 bit instruct model, freeze the feedforward layer, remove the QKB attention layer, and replace it with RWKB linear layers.[00:26:32] Eugene Cheah: So to be clear, this means we do not have the rwkv channel mix layer, we only have the time mix layer. But but once we do that, we train the rwkv layer. Important is that the feedforward layer needs to be frozen, so the new attention can be learned. And then we unfreeze the feedforward layer, and train all the layers together with a custom learning rate schedule, so that they can learn how to work together.[00:26:54] Eugene Cheah: The end result, surprisingly, And, to be honest, to the frustration of the R. W. [00:27:00] KV MOE team, which ended up releasing the model on the same day, was that, with just a few hours of training on two nodes, we managed to get it to be on par, kind of, with the original QUAN32B model. So, in fact, when the first run, right, that completely confused us, it was like, and I was telling Daniel Goldstein, Smirky, who kind of leads most of our research coordination, When you pitched me this idea, you told me at best you'll get the same level of performance.[00:27:26] Eugene Cheah: You didn't tell me the challenge and score and Winograd score will shoot up. I don't know what's happening there. But it did. MMLU score dropping, that was expected. Because if you think about it, when we were training all the layers, right, we were essentially Like, Frankenstein this thing, and we did brain damage to the feedforward network layer 2 with the new RWKB layers.[00:27:47] Eugene Cheah: But, 76%, hey, somehow it's retained, and we can probably further train this. We didn't even spend more than 3 days training this, so there's a lot more that can be done, hence the preview. This brings up [00:28:00] a big question, because We are already now in the process of converting to 7TB. We are now, this is actually extremely compute efficient to test our attention mechanic.[00:28:10] Eugene Cheah: It's like, it becomes a shortcut. We can, we are already planning to do our version 7 and our hybrid architecture for it. Because we don't need to train from scratch. And we get a really good model out of it. And the other thing that is uncomfortable to say is that because we are doing right now on the 70b is that if this scales correctly to 128k context length, I'm not even talking about a million 128, majority of enterprise workload today is just on 70b at under 32k context length.[00:28:41] Eugene Cheah: That means if this works and the benchmark matches it, It means we can replace the vast majority of current AI workload, unless you want super long context. And then sorry, can someone give us more GPUs? Because we do need the VRAM for super long context, sadly. So yeah, that's what we are working on, and essentially, [00:29:00] we are excited about this to just push it further.[00:29:02] Eugene Cheah: And this conversion process, to be clear, I don't think it's going to be exclusive to RWKB. It probably will work for Mamba as well, I don't see why not. And we will probably see more ideas, or more experiments, or more hybrids, or Yeah, like, one of the weirdest things that I wanted to say outright, and I confirmed this with the Black Mamba team and the Jamba team, which because we did the GoFinch hybrid model, is that none of us understand why a hard hybrid with a state based model to be R.[00:29:28] Eugene Cheah: QA state space and transformer performs better when, than the baseline of both. It's like, it's like when you train one, you expect, and then you replace, you expect the same results. That's our pitch. That's our claim. But somehow when we jam both together, it outperforms both. And that's like one area of emulation that, like, we only have four experiments, plus four teams, that a lot more needs to be done.[00:29:51] Eugene Cheah: But, but these are things that excite me, essentially, because that is what it's potentially we can move ahead for. Which brings us to what comes next.[00:30:00] What's next[00:30:00] [00:30:00][00:30:00] Dan Fu: So, this part is kind of just some, where we'll talk a little bit about stuff that, that we're excited about. Maybe have some wild speculation on, on what, what's, what's coming next.[00:30:12] Dan Fu: And, of course this is also the part that will be more open to questions. So, a couple things that, that I'm excited about is continued hardware model co design for, for these models. So one of the things that we've put out recently is this library called ThunderKittens. It's a CUDA library.[00:30:29] Dan Fu: And one of the things that, that we found frustrating is every time that we built one of these new architectures, and I'm sure you had the exact same experience, we'd have to go and spend two months in CUDA land, like writing these, these new efficient things. And. If we decided to change one thing in PyTorch, like one line of PyTorch code is like a week of CUDA code at least.[00:30:47] Dan Fu: So one of our goals with, with a library like Thunderkitten, so we, we just broke down what are the key principles, what are the key hardware things what are the key, Compute pieces that you get from the hardware. So for example on [00:31:00] H100 everything is really revolves around a warp group matrix multiply operation.[00:31:06] Dan Fu: So you really want your operation to be able to split into relatively small matrix, matrix multiply operations. So like multiplying two 64 by 64 matrices, for example. And so if you know that ahead of time when you're designing your model, that probably gives you you know, some information about how you set the state sizes, how you set the update, how you set the update function.[00:31:27] Dan Fu: So with Thunderkittens we basically built a whole library just around this basic idea that all your basic compute primitives should not be a float, but it should be a matrix, and everything should just be matrix compute. And we've been using that to, to try to both re implement some existing architectures, and also start to design code.[00:31:44] Dan Fu: Some new ones that are really designed with this core with a tensor core primitive in mind. Another thing that that we're, that at least I'm excited about is we, over the last four or five years, we've really been looking at language models as the next thing. But if you've been paying [00:32:00] attention to Twitter there's been a bunch of new next generation models that are coming out.[00:32:04] Dan Fu: So there, there are. So, video generation models that can run real time, that are supported by your mouse and your keyboard, that I'm told if you play with them that, you know, that they only have a few seconds of memory. Can we take that model, can we give it a very long context length so that you could actually maybe generate an entire game state at a time?[00:32:25] Dan Fu: What does that look like for the model? You're certainly not going to do a giant quadratic attention computation to try to run that. Maybe, maybe use some of these new models, or some of these new video generation models that came out. So Sora came out I don't know, two days ago now. But with super long queue times and super long generation times.[00:32:43] Dan Fu: So that's probably a quadratic attention operation at the, at the bottom of it. What if we could remove that and get the same quality, but a lot faster generation time? Or some of the demos that we saw from Paige earlier today. You know, if I have a super long conversation with my [00:33:00] Gemini bot, what if I wanted to remember everything that it's seen in the last week?[00:33:06] Dan Fu: I mean, maybe you don't for personal reasons, but what if I did, you know? What does that mean for the architecture? And I think, you know, that's certainly something I'm pretty excited about. I'm sure you're excited about it too. So, I think we were supposed to have some hot takes, but I honestly don't remember what our hot takes were.[00:33:21] Hot Takes - does anyone really need long context?[00:33:21] Eugene Cheah: Yeah, including the next slide. Hot takes, yes, these are our[00:33:25] Dan Fu: hot takes.[00:33:25] Eugene Cheah: I think the big one on Twitter that we saw, that we shared, was the question is like, is RAG relevant? In the case of, like, the future of, like, state based models?[00:33:38] Dan Fu: Let's see, I haven't played too much with RAG. But when I have. I'll say I found it was a little bit challenging to do research on it because we had this experience over and over again, where you could have any, an embedding model of any quality, so you could have a really, really bad embedding model, or you could have a really, really [00:34:00] good one, By any measure of good.[00:34:03] Dan Fu: And for the final RAG application, it kind of didn't matter. That's what I'll say about RAG while I'm being recorded. I know it doesn't actually answer the question, but[00:34:13] Eugene Cheah: Yeah, so I think a lot of folks are like, extremely excited of the idea of RWKB or State Space potentially having infinite context.[00:34:21] Eugene Cheah: But I think the reality is that when we say infinite context, we just mean a different kind of infinite context, or you, or as it's previously covered, you need to test the model differently. So, think of it more along the lines of the human. Like, I don't remember what I ate for breakfast yesterday.[00:34:37] Eugene Cheah: Yeah, that's the statement that I'll say. And And we humans are not quadratic transformers. If we did, if let's say we increased our brain size for every second we live, we would have exploded by the time we are 5 years old or something like that. And, and I think, I think basically fundamentally for us, right, be it whether we, regardless of whether RWKB, statespace, XLSTM, [00:35:00] etc, our general idea is that instead of that expanding state, that increase in computational cost, what if we have a fixed state size?[00:35:08] Eugene Cheah: And Information theory detects that that fixed state size will have a limit. Just how big of a limit is a question, like, we, like, RWKB is running at 40 megabytes for, for its state. Its future version might run into 400 megabytes. That is like millions of tokens in, if you're talking about mathematically, the maximum possibility.[00:35:29] Eugene Cheah: It's just that I guess we were all more inefficient about it, so maybe we hit 100, 000. And that's kind of like the work we are doing, trying to like push it and maximize it. And that's where the models will start differing, because it will choose to forget things, it will choose to remember things. And that's why I think that there might be some element of right, but it may not be the same right.[00:35:49] Eugene Cheah: It may be the model learn things, and it's like, hmm, I can't remember that, that article. Let me do a database search, to search. Just like us humans, when we can't remember the article in the company. We do a search on Notion. [00:36:00][00:36:00] Dan Fu: I think something that would be really interesting is if you could have facts that are, so right now, the one intuition about language models is that all those parameters are around just to store random facts about the world.[00:36:14] Dan Fu: And this intuition comes from the observation that if you take a really small language model, it can do things like talk to you, or kind of has like the The style of conversation, it can learn that, but where it will usually fall over compared to a much larger one is it'll just be a lot less factual about things that it knows or that it can do.[00:36:32] Dan Fu: But that points to all those weights that we're spending, all that SGD that we're spending to train these models are just being used to store facts. And we have things like databases that are pretty good at storing facts. So I think one thing that would be really interesting is if we could actually have some sort of outside data store that a language model can can look at that that maybe is you know, has has some sort of gradient descent in it, but but would be quite interesting.[00:36:58] Dan Fu: And then maybe you could edit it, delete [00:37:00] facts, you know, change who's president so that it doesn't, it doesn't get lost.[00:37:04] Vibhu: Can we open up Q& A and hot takes for the audience? I have a hot take Q& A. Do these scale? When, when 405B state space model, RAG exists, no one does long context, who's throwing in 2 million token questions, hot takes?[00:37:24] Dan Fu: The, the who's throwing in 2 million token question, I think, is, is a really good question. So I actually, I was going to offer that as a hot take. I mean, my hot take was going to be that long context doesn't matter. I know I just gave a whole talk about it, but you know, what, what's the point of doing research if you can't, you know, play both sides.[00:37:40] Dan Fu: But I think one of the, so I think for both of us, the reason that we first got into this was just from the first principled questions of there's this quadratic thing. Clearly intelligence doesn't need to be quadratic. What is going on? Can we understand it better? You know, since then it's kind of turned into a race, which has [00:38:00] been exciting to watch, like, how much context you can take in.[00:38:03] Dan Fu: But I think it's right. Nobody is actually putting in a two million context prompt into these models. And, and, you know, if they are, maybe we can go, go You know, design a better model to do that particular thing. Yeah, what do you think about that? So you've also been working on this. Do you think long context matters?[00:38:19] Eugene Cheah: So I'm going to burn a bit. How many of you remember the news of Google Gemini supporting 3 million contacts, right? Raise your hand.[00:38:28] Vibhu: Yeah, 2 million.[00:38:29] Eugene Cheah: Oh, it's 2 million.[00:38:31] Eugene Cheah: Yeah, how many of you actually tried that? See?[00:38:34] Vibhu: I use it a lot. You? You work for MindsTV. I use it a lot.[00:38:41] Eugene Cheah: So, for some people that has used, and I think, I think that's the, that's might be, like, this is where my opinion starts to differ, because I think the big labs may have a bigger role in this, because Like, even for RWKB, even when we train non contacts, the reason why I say VRAM is a problem is that because when we did the, we need to backprop [00:39:00] against the states, we actually need to maintain the state in between the tokens by the token length.[00:39:05] Eugene Cheah: So that means we need to actually roll out the whole 1 million contacts if we are actually training 1 million. Which is the same for transformers, actually, but it just means we don't magically reuse the VRAM consumption in the training time space. So that is one of the VRAM bottlenecks, and I'm neither OpenAI nor Google, so donate GPUs if you have too much of them.[00:39:27] Eugene Cheah: But then, putting it back to another paradigm, right, is that I think O1 style reasoning might be actually pushing that direction downwards. In my opinion, this is my partial hot take is that if, let's say you have a super big model, And let's say you have a 70B model that may take double the tokens, but gets the same result.[00:39:51] Eugene Cheah: Strictly speaking, a 70B, and this is even for transformer or non transformer, right? We we'll take less less resources than that 400 B [00:40:00] model, even if it did double the amount thinking. And if that's the case, and we are still all trying to figure this out, maybe the direction for us is really getting the sub 200 B to be as fast as efficient as possible.[00:40:11] Eugene Cheah: We a very efficient architecture that some folks happen to be working on to, to just reason it out over larger and larger context thing.[00:40:20] Question: Yeah. One thing I'm super interested in is. Models that can watch forever? Obviously you cannot train something on infinite context length. How are y'all thinking about that, where you run on a much longer context length than is possible to train on?[00:40:38] Dan Fu: Yeah, it's a, it's a great question. So I think when I think you guys probably had tweets along these lines, too. When we first started doing these things, because these are all recurrent models in theory you could just run it forever. You could just run it forever. And at the very least it won't, it won't like error out on your crash.[00:40:57] Dan Fu: There's another question of whether it can actually [00:41:00] use what it's seen in that infinite context. And I think there, so one place where probably the research and architectures ran faster Then another research is actually the benchmarks for long context. So you turn it on forever. You want to do everything or watch everything.[00:41:16] Dan Fu: What is it that you actually wanted to do? Can we actually build some benchmarks for that? Then measure what's happening. And then ask the question, can the models do it? Is there something else that they need? Yeah, I think that if I were to turn back the clock to 2022, that's probably one of the things I would have done differently, which would have been actually get some long context benchmarks out at the same time as we started pushing context length on all these models.[00:41:41] Eugene Cheah: I will also say the use case. So like, I think we both agree that there's no Infinite memory and the model needs to be able to learn and decide. I think what we have observed for, I think this also fits the state space model, is that one of the key advantages of this alternate attention mechanic that is not based on token position is that the model don't suddenly become crazy when you go past the [00:42:00] 8k training context tank, or a million context tank.[00:42:03] Eugene Cheah: It's actually still stable. It's still able to run, it's still able to rationalize. It just starts forgetting things. But some of these things are still there in latent memory. Some of these things are still somewhat there. That's the whole point of why reading twice works. Things like that. And one of the biggest pushes in this direction is that I think both Statespace and RWKB have Separate papers by other researchers where they use this architecture for time series data.[00:42:26] Eugene Cheah: Weather modeling. So, you are not asking what was the weather five days ago. You're asking what's the weather tomorrow based on the infinite length that we, as long as this Earth and the computer will keep running. So, so, and they found that it is like, better than existing, like, transformer or existing architecture in modeling this weather data.[00:42:47] Eugene Cheah: Control for the param size and stuff. I'm quite sure there are people with larger models. So, so there are things that, that in this case, right, there is future applications if your question is just what's next and not what's 10 years ago.[00:42:59] Dan Fu: Thanks so [00:43:00] much for having us. Get full access to Latent Space at www.latent.space/subscribe

Nesse episódio do Podcast da Lambda powered by TIVIT,  Fernando Okuma, Paulo Barros e Rodrigo Monney batem um papo sobre os desafios enfrentados para criar um ambiente de desenvolvimento local que dê autonomia para as pessoas desenvolvedoras trabalhem sem dependências externas.   Lambda3 · #417 - Setup de infraestrutura para desenvolvimento local Participantes: Fernando Okuma - @feokuma Paulo Barros - @paulo-henrique-costa-barros Rodrigo Monney - @rodrigo-monney Pauta: Por que precisamos de uma infraestrutura rodando localmente? Ambiente isolado Autonomia para reconstruir tudo sempre que necessário Independencia de recursos externos Quais as formas de ter a infraestrutura rodando localmente Instalar os recursos no sistema operacional (Bancos de dados, mensageria, ferramentas de observabilidade, ...) Ambiente rodando em containers Execução em MicroVM Quais são os desafios em criar essa estrutura localmente? Manutenção para manter alinhado com o que será utilizado em produção Uso de recursos da máquina (memória ram, cpu, ...) Quais estratégias existem para poder lidar com a limitação de hardware Documentação para que as pessoas do time consigam entender como usar Aplicações, serviços e dependências externas Criação de stub Consumir a dependência real ​Edição: Compasso Coolab Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

Nesse episódio do Podcast da Lambda powered by TIVIT,  Tatiana Sartori, Fernando Okuma e Graziela Nobre conversam com Alexandre Santos Costa sobre como a tecnologia e a IA estão ajudando as pessoas com deficiência no seu dia a dia.  Lambda3 · #416 - Como a Tecnologia e a IA podem auxiliar as pessoas com deficiência Participantes: Fernando Okuma - @feokuma Tatiana Sartori - @tatiana-sartori Graziela Nobre - @graziela-nobre Alexandre Santos Costa - @magoolation Pauta: O QUE É A  IA IA Assistiva e Acessibilidade Desafios As tecnologias de IA Assistiva são confiáveis Privacidade de dados Tecnologias Assistivas com IA (TIVIT e o que está no Mercado) Assistentes de voz Leitores de tela inteligentes Reconhecimento de fala e Transcrição Automática Aplicativos de Tradução de Linguagem de Sinais Como está a usabilidade dessas tecnologias? Acessibilidade financeira Importância dessas tecnologias Inclusão social Autonomia Personalização Educação e trabalho ​Edição: Compasso Coolab Créditos das músicas usadas neste programa: Music by Kevin MacLeod (incompetech.com) licensed under Creative Commons: By Attribution 3.0 - creativecommons.org/licenses/by/3.0

On this Screaming in the Cloud Replay, we're revisiting our conversation with Co-Founder of Senzo, Ant Stanley. Ant sits down with Corey to do so. He offers up his history which has lead to his time as “Serverless Hero” to landing on the line that “serverless sucks.” Lend us your ears to see how that transition happened! Ant goes into detail on JeffConf (not the of the Bezos nomen), and working with servers and what to put where and why. Ant and Corey talk over the plague of AWS services where Ant offers his perspective how to trim the fat and keep things simple to make long-term objectives more attainable. They discuss the importance of training, the role of certifications for better and worse, and more. Tune in for his take!Show Highlights(0:00) Intro(0:51) Duckbill Group sponsor read(1:24) What does it mean to be an AWS Serverless Hero?(3:13) Why Ant and Corey are critical of the state of serverless(7:53) Woes with Lambda and CloudFront(10:12) The never-ending stream of new AWS services(13:36) Hurdles ahead of going serverless(17:33) Struggles of getting customers to understand a newly built service(21:31) Duckbill Group sponsor read(22:14) Pros and cons of certifications(32:17) Where you can find more from AntAbout Ant StanleyAnt Stanley is a community focused technologist with a passion for enabling better outcomes for society through technology. He is an AWS Serverless Hero, runs the Serverless London User Group, co-runs ServerlessDays London and is part of the ServerlessDays Global team. LinksA Cloud Guru: https://acloudguru.comhomeschool.dev: https://homeschool.devaws.training: https://aws.traininglearn.microsoft.com: https://learn.microsoft.comTwitter: https://twitter.com/iamstanOriginal Episodehttps://www.lastweekinaws.com/podcast/screaming-in-the-cloud/serverless-hero-got-servers-in-his-eyes-with-ant-stanley/SponsorThe Duckbill Group: duckbillgroup.com 

Nesse episódio do Podcast da Lambda powered by TIVIT, Fernando Okuma convida renomados professores da área, para uma conversa sobre como é ensinar tecnologia em 2024.

Welcome to episode 283 of The Cloud Pod, where the forecast is always cloudy! Break out your crystal balls and shuffle those tarot decks, because it’s Re:Invent prediction time! Sorry we missed you all last week – the plague has been strong with us. But Justin and Jonathan are BACK, and we've got a ton of news, so buckle in and let's get started!  Titles we almost went with this week: Not My Snowcones!  Lambda at 10: Still Better Than Windows Containers  A big thanks to this week's sponsor: We're sponsorless! Want to get your brand, company, or service in front of a very enthusiastic group of cloud news seekers? You've come to the right place! Send us an email or hit us up on our slack channel for more info.  General News   01:27 The voice of America Online's “You've got mail” has died at age 74 Elwoods Edwards, the voice behind the online service AOL's iconic “You've got mail” sound notification has died at the age of 74. He was just one day shy of his 75th birthday.  The “you've got mail” soundbite started in 1989 when Steve Case, CEO of Quantum Computer Services (which will later become America Online or AOL,) wanted to add a human voice to their Quantum online service.   Karen Edwards, who worked as a customer service representative, heard Case discussing the plan and suggested her husband Elwood, a professional broadcaster.  Edwards recorded the famous phrase and others (“Welcome” “File's done” and “Goodbye” among them) on a cassette recorder in his living room.  He was paid $200 for the service.   His voice is still used to greet users of the current AOL service.  AWS  03:04 It's Time for RE:Invent Predictions! Matt Large Green Computing Reinvent LLM at the Edge Something new On S3 Ryan (AI) Improved serverless observability tools Expansion of AI Driven workflows in datalakes Greater Focus on Multi-Account or Multi-region orchestration, centralized compliance management, or enhanced security services Jonathan New Edge Computing Capabilities better global application deployment type features. (Cloudflare competitor maybe) New automated cost optimization tools Automated RAG/vector to S3 Justin  Managed Backstage or platform like service New LLM multi-modal replacement or upgrade to Titan Competitor VM offering to Broadcom  Honorable Mentions Jonathan: Deeper integration between serverless and container services New Region Enhanced Observability with AI driven debugging tool Justin: Multi Cloud management – in a bigger way (Anthos competitor) Agenti