Podcasts about LM

Allen covers Energy Capital Partners buying TPI’s blade factories, GE Vernova’s $1.7 billion rescue of LM Wind Power, offshore wind cutting oil burn during a heat wave, Scotland’s Caledonia approval, and 19 states suing the Pentagon over stalled wind reviews. Sign up now for Uptime Tech News, our weekly newsletter on all things wind technology. This episode is sponsored by Weather Guard Lightning Tech. Learn more about Weather Guard’s StrikeTape Wind Turbine LPS retrofit. Follow the show on YouTube, Linkedin and visit Weather Guard on the web. And subscribe to Rosemary’s “Engineering with Rosie” YouTube channel here. Have a question we can answer on the show? Email us! Good Monday everyone. A few months ago, we told you about a Houston bankruptcy court carving up TPI Composites. Well, that story just got a whole lot bigger. On July sixth, TPI walked out of Chapter Eleven. Zero debt. New owners. A private equity firm called Energy Capital Partners picked up TPI’s blade factories in Iowa and Juarez, Mexico for about twenty million dollars. Twenty million, against more than a billion dollars in liabilities. ECP did not stumble into wind blades. They bought Calpine back in twenty eighteen, inherited seventy-seven power plants, and became GE’s biggest private gas turbine customer in the Western Hemisphere. That relationship, forged in gas turbine halls, is what brought them to composite factories. GE Vernova signed a five-year supply deal requiring it to send blade orders to ECP’s factories. GE is ECP’s partner, its customer, and was even the backup buyer if the deal fell through. So TPI lives on, leaner, debt-free, with locked-in demand from one of the biggest turbine makers on earth. But now, the other side of that coin. While ECP picked up two blade factories for twenty million dollars, GE Vernova recently pumped one-point-seven billion dollars into its own blade company, LM Wind Power. LM’s equity had fallen to negative 575 million euros. Revenue dropped ninety-six percent in one year, from 2.1 billion Danish kroner down to just ninety-three million. The Danish workforce, cut to about twenty-five people. LM Wind Power has lost money every single year since GE bought it in twenty seventeen. Nine straight years of red ink. So think about that. Two American blade factories now serve GE Vernova’s onshore business. One in Grand Forks, North Dakota, owned by GE, inside a division losing four hundred million dollars a year. The other in Newton, Iowa, owned by ECP, zero debt, five-year supply deal. The independent contract blade business that TPI Composites built is gone. Vestas took the India and Mexico plants in-house. GE’s supply is locked to ECP. The OEMs and their financial partners now own the factories directly. And that is a new era for wind manufacturing. Now, let us talk about what those blades are doing once they are spinning. Earlier this month, a brutal heat wave hit the eastern United States. Air conditioners running full blast. Grid operators scrambling to keep up. And off the coast of New England, two offshore wind farms stepped up. Vineyard Wind, eight hundred and six megawatts off Massachusetts. Revolution Wind, seven hundred and four megawatts near Rhode Island. Together they pushed hundreds of megawatts into the grid right when people needed it most. And here is the number that matters. Oil-fired power plants met about ten percent of peak demand on July second this year. Last summer, at the height of a similar heat wave, oil plants covered nearly fifteen percent. That is more than a gigawatt less oil burned. The projects that survived lawsuits, survived construction shutdowns, survived lease freezes, are now keeping the lights on in New England. Across the Atlantic, Scotland just approved two massive offshore wind farms. The Caledonia North and South projects in the Moray Firth, up to one hundred and forty turbines spread across one hundred and sixty-five square miles. Enough power for two million homes. Ocean Wind is leading the development with a commitment of about 1.7 billion pounds. And here is what makes this project different. Caledonia South will mix fixed-bottom and floating turbines, up to thirty-nine floaters. That blend of proven and next-generation technology on a single project is something to watch. Back in the United States, nineteen state attorneys general are suing the Department of Defense. The reason, wind project reviews. Federal law says any wind turbine taller than two hundred feet must go through a Defense Department check, to make sure it does not interfere with military radar or flight paths. Last August, the Pentagon stopped reviewing those projects. No explanation. No timeline for starting again. Maryland Attorney General Anthony Brown is leading the coalition, joined by attorneys general from eighteen other states including California, New York, and New Jersey. They want a court to force the Defense Department to start doing its job again. And finally, a story from the sea floor. Down in southern New England, lobster populations have been falling for decades. Back in nineteen ninety-eight, there were about fifty million lobsters in those waters. By twenty twenty-two, fewer than ten million. But something else is moving in. Jonah crabs. Fishermen used to throw them back. Now they are hauling them in by the thousands, selling them as a cheaper option to lobster. And researchers at the University of Rhode Island are finding that offshore wind foundations are acting like artificial reefs. Algae grows first, then barnacles and mussels, then fish and crabs follow. The question scientists are working to answer is whether these structures create new marine life, or just pull it in from the surrounding ocean. Either way, the turbines are not just making electricity. They are making habitat. Now, here is what to watch. This Wednesday, July twenty-second, GE Vernova reports second quarter earnings. And the numbers we just talked about will be in the room. One-point-seven billion dollars pumped into LM Wind Power, a blade company that has lost money nine years straight. Twenty million dollars to let ECP walk away with two factories and a five-year supply deal. GE Vernova is guiding for four hundred million dollars in wind segment losses this year. Meanwhile, its Power and Electrification divisions are printing money, nearly five billion dollars in free cash flow last quarter alone. So the question on that earnings call is simple. If you are spending eighty times more to keep your in-house blade maker alive than a private equity firm paid to buy your contract supplier, how long do you keep doing both? Watch for what GE Vernova says about LM Wind Power’s future, about North American onshore blade strategy, and about whether that 1.7 billion dollar injection was a rescue, or a goodbye. The answer could reshape who makes blades in this industry for the next decade. And that is the state of the wind industry for the 19th of July, twenty twenty-six. Join us for the Uptime Wind Energy Podcast tomorrow.

The Mullet Guild: Episode #25: Dune Part Three I want some spice! We have some spice, right here! On today's episode of LM's The Mullet Guild, the Mullety One checks out the trailer to #DunePartThree and diagnoses with his prescience the effects Paul Muad'dib is having on the universe with his empire, and we have a big Happy Birthday for our cofounder of FPN Kyle Wagner, all here from the sietch at Quinlan's Cantina. #TheMulletGuildADuneFanPodcast #LethalMulletPodcast #DunePartthree #Dune

While Spencer is busy in Boston, we're diving back into the Dear Prudie mines. Join Spencer, Ty, and Andy as they help some folks out with advice on the proper way to leave a child in a hot car, what to do when your ex is a predator, and the ethics of helping out your dead husband's mistress. Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

Laudetur Jesus Christus - Ngợi khen Chúa Giêsu KitôRadio Vatican hằng ngày của Vatican News Tiếng Việt.Nội dung chương trình hôm nay:0:00 Bản tin15:41 Chia sẻ Lời Chúa : Lm. Giuse Trần Sĩ Nghị, SJ, chia sẻ Lời Chúa Chúa Nhật thứ 16 thường niên---Những hình ảnh này thuộc Bộ Truyền Thông của Toà Thánh. Mọi sử dụng những hình ảnh này của bên thứ ba đều bị cấm và dẫn đến việc đánh bản quyền, trừ khi được cho phép bằng giấy tờ của Bộ Truyền Thông. Copyright © Dicasterium pro Communicatione - Giữ mọi bản quyền.

FAN MAIL--We would love YOUR feedback--Send us a Text MessageA 78-year-old man sits in solitary confinement in Hong Kong with a 20-year sentence hanging over him. He could have left. He could have said the words that would have set him free. Jimmy Lai refused to lie, and that single act of conscience explains why authoritarian power reacts so fiercely to one person who won't be bought or frightened into agreement. We trace Lai's story from his rise as a self-made publisher and founder of Apple Daily to the convictions that made him a political prisoner, then widen the lens to a timeless pattern: rulers don't only want silence, they want affirmation. To make the point vivid, we bring in Robert Bolt's A Man for All Seasons and the example of Sir Thomas More, whose quiet “no” threatened a king more than any army ever could. From there, we get personal. Tyranny's business model depends on getting good people to say things they don't believe, and most of the time the pressure arrives in small, ordinary moments. A boss, a crowd, or an algorithm nudges you toward the easy line instead of the true one, and repetition slowly erases your ability to resist. You'll leave with one clear action step for the week, a sharper understanding of why free speech and freedom of the press matter, and a reason to take courage seriously before it costs more than you think. Key Points from the Episode:• the weight of “20 years” for a 78-year-old man and why it reads like a life sentence  • the Thomas More parallel from A Man for All Seasons and the danger of quiet refusal  • who Jimmy Lai is, why he founded Apple Daily, and why he stayed  • why regimes punish unbought people and make examples of them  • conscience as an antidote to tyranny and the everyday pressures to comply  • a simple action step for the next moment you feel pushed to say the comfortable lie  • a warning about treating communism as abstract and a call to sober attention at home  • prayer for Jimmy Lai and others imprisoned for refusing to bow  Want to go deeper?Watch the documentary on Jimmy Lai's story and the fall of Hong Kong: The Hong Konger — freejimmylai.comGo back and listen to where this story started — our 2022 episode: LM #21 — Two Iron Ladies: The Hong KongerAnd keep Jimmy Lai in your prayers tonight, along with every man and woman sitting in a cell right now simply for refusing to lie.Want to leave a review? Click here, and if we earned a five-star review from you — high five and knuckle bumps — we appreciate it greatly, thank you so much!Keep fighting the good fight.Links?

Lethal Mullet Podcast: Episode #317: Repo Man #newlookmullet we return to The Mullet, at Quinlan's Cantina, and a look at the Alex Cox film REPO MAN starring Emilio Estevez, and Harry Dean Stanton, along with Tracy Walter and a host of others. A punk gets picked up by a Repo Man for a job, and decides to join the business for a job, unbeknownst to them one of their jobs is a car with an extraterresstrial in it. A bizarre, road movie, and one that is for the lack of a better word, punky, and with a new wave sensibility. We look at that and more on the show tonight on LM. Site: fpnet.podbean.com Socials: @thelethalmullet #repoman #lethalmulletpodcast #fandompodcastnetwork #eighties

Rosemary previews Pardalote’s new hands-on blade repair course. EverWind’s Ocean Lake, Canada’s largest wind project, will feed a green hydrogen and ammonia plant in Nova Scotia rather than the grid. Plus BP’s exit from an offshore project in Japan, and the wake-effect lawsuit pitting SSE, Equinor, and Vårgrønn against RWE’s Dogger Bank South. Sign up now for Uptime Tech News, our weekly newsletter on all things wind technology. This episode is sponsored by Weather Guard Lightning Tech. Learn more about Weather Guard’s StrikeTape Wind Turbine LPS retrofit. Follow the show on YouTube, Linkedin and visit Weather Guard on the web. And subscribe to Rosemary’s “Engineering with Rosie” YouTube channel here. Have a question we can answer on the show? Email us! The Uptime Wind Energy podcast, brought to you by StrikeTape. Protecting thousands of wind turbines from lightning damage worldwide. Visit striketape.com. And now your hosts Allen Hall 2025: Welcome to the Uptime Wind Energy podcast. I’m your host, Allen Hall. I’m here with Matthew Stead, Yolanda Padron, and Rosemary Barnes is back this week.  Rosemary, you’ve been to a number of training courses over the last couple of weeks. The first off was GWO. What was your experience at GWO training?  Rosemary1: It was the fourth or maybe even fifth time that I’ve done it. Um, I did it a few times in Denmark and then, uh, this is the second time doing it in Australia. also, this was my first time doing first aid in Australia. Last time they did GWO here, but my first aid was still valid from Europe, so I, I didn’t redo it. And it’s like so much about [00:01:00] snakes and spiders and jellyfish But a good, good rule of thumb, not 100% accurate, but good rule of thumb, if it is something from the ocean that stung you, then you put something warm on it, and if it’s something from the land that stung or bit you, then something cold on it, Allen Hall 2025: well, how often do you usually take GWO training? Rosemary1: You gotta do it every two years to be valid. I don’t do it every two years because, um, if you do it every two years, like within two years, then you can do the refresher course. So that’s three days instead of four However, um, because I don’t climb constantly, like often it will be six months or more in between climbs, I’ll just do it before I know that I’ve got a climb. all the other people except for one were technicians who, you know, have been working for a while. So they’re also doing the full course, not the refresher. So they get a little bit more practice than I do. But, um, it’s just not often enough. Y-you know, like every time I go it’s like I, I really feel the need to have the refresher, um, because I’m just not fully on top of it. ‘Cause it’s [00:02:00] not just that you need to know what to do. You need to be able to… Like if you need to use it, you’re gonna be freaking out, you know? This is the worst thing that’s probably ever happened in your life, and now you’ve gotta remember all your training. It’s like you want it to be actually second nature to some extent. So yeah, first day is manual handling, which is v- you know, very– That one’s very easy and I would be happy to never do that again. Like I will always remember that. Um, then you got fire, um, fire safety awareness, and that one’s just fun ’cause you just get to, um, light fires and put stuff out then first aid, which I definitely always want a refresher on. The CPR dummies at this place, they had lights, um, and it lit up green if you were doing it right, and I haven’t used a dummy that was so advanced before, so that was quite good. I realized I wasn’t pressing hard enough. and then yeah, last two days is working at heights training, which is the most intense ’cause you got your harness on all day and, um, you know, climbing up and down and rescuing people. this was Rite Training in Goulburn, and, um, the [00:03:00] instructor’s name was Claire. highly recommend doing that one. Allen Hall 2025: Is that a general requirement in Australia that you have GWO before you can climb? Rosemary1: Like, yeah, they will sometimes, um, let you climb if you are babysat by people. I would not recommend other engineers, like if you’ve never climbed a wind turbine before, like I would really not recommend that you just go up with a team and haven’t done the training because you do need to be able to use a ladder safely and, um, you can, y- you can easily, like even inside the nacelle, you could easily hurt yourself really badly if you’re used to working in an office, uh, you’re upping your danger level by, you know, like many, many, many times by going up a turbine and it’s just something that you gotta take seriously. Allen Hall 2025: How busy are the courses in Australia? Are a lot of technicians trying to get in and get trained?  Rosemary1: No, it’s people that have a job that are getting trained. But there were heaps of techs in this course. There were maybe eight or so, which is also part of the reason why it took a really long time. Allen Hall 2025: So [00:04:00] this week, as we record, y- you’re presenting a blade repair course for engineers and technicians. a completely new area that you’re, uh, going into in terms of offering advice and expertise that it’s really hard to find on the planet. It’s probably a, a, a busy or, or requested course, I would imagine, in Australia, where you just don’t have access to a lot of the manufacturers. Rosemary2: it’s a, it’s a course for just for engineers or technical type people, um, but including hands-on stuff. So the way that I I forced this to come into being was just the last five years. I, um, you know, I started working a lot on wind turbine blade repairs and, um, people would ask me, you know, “Have these repairs been done right?” And the thing is that the only repairs that I had anything to do with when I was working at LM were weirdo ones, right? [00:05:00] Where the normal, like a technician couldn’t, couldn’t handle it. It was outside of, um, yeah, their, their standard, uh, kind of repairs that they can do for whatever reason. and now in the work that we do at Part Load, it’s primarily normal repairs, and I just didn’t know exactly what technicians know. You know, how do they, how do they know whether they can repair it or not? What do they know before they go up there? When are they calling the engineer? Um, all that sort of stuff, like the normal stuff. eventually it became less about me learning, ’cause like I said, I kind of picked up most of it. Um, but now I’ve got staff that I’m training up to be, uh, you know, composites engineers and to work with these kinds of issues. There’s a lot of repetitive tasks involved in what we do when we, like, assess the condition of a wind farm. A lot of what we do is look main- manually looking through photos and thing- if things are classified right or not. I [00:06:00] Found this guy from Direct Wind Services, Jurij Eska. He’s a blade engineer. He’s worked in Europe and then come back to Australia, so a little bit like me. And, um, I just worked with him on a few projects and I’m like, “Oh, okay. Well, this guy, uh, he really gets it.” And I asked him, “How do you, how do you train your technicians? What course do they do? Maybe I can do that course.” And he said, “Oh, we train them ourselves.” And so then I asked him to put this course together. So where we started off the course yesterday, that was, um, uh, an indoor session where I was talking through how are blades designed, uh, certified, tested, manufactured, um, what kinds of manufacturing defects can you see and what do they do about them in the factory? ‘Cause you know that they’re doing a lot of repairs in the factory already before you ever see a, a brand new blade. and then the next three days we’re going to be working on, um, yeah, grinding and [00:07:00] infusions and a bit of a, a bit of theory about, um, composite repairs.  Allen Hall 2025: What do you feel like are those key skill sets that engineers should know how to do, maybe not as well as a, a professional technician that does it a lot, but at least at a beginner’s level should be able to complete them before they start repairing blades on their own and giving advice about how to repair blades? What, what are those key items? Rosemary2: part of it is that I want them to be able to understand what is a bad damage and what’s not a bad damage cause you look a lot at images from the outside, but it’s really about what’s on the inside and how deep it goes is the real thing. So, um, it’ll be about learning, you know, developing some judgment about, um, how bad it can be and how bad it can look on the outside. We’re not gonna be looking at so many real damages ’cause like obviously we’re just dealing with pieces that are in the, um, in the, uh, workshop and Yuri has [00:08:00] made some samples for us, um, purposely made them badly so that we’ve got some, you know, damage to find. Allen Hall 2025: Are you addressing carbon fiber at all? Rosemary2: Uh, I actually haven’t asked about that. I don’t think so. Carbon fiber is, um, is a real pain to work with because it’s conductive. Like, even grinding it makes a bit of a hazardous work environment. We did talk a little bit about the different materials yesterday and, um, about pultrusions. And actually, it turns out Yuri used to work somewhere where they, uh, manufactured pultrusions, and I had always, I was always under the impression that a pultrusion is, you know, like, perfectly s- perfectly straight. That’s the point. And he’s like, “No way.” No way. There’s waviness in the pultrusions  Allen Hall 2025: And on March 3rd through 5th at WOMA 2027, Rosie, you’re gonna give part of this course as part of WOMA, right? Rosemary2: Little, little mini course. We’ll have to decide what, what makes sense to include, ’cause it was… Yeah, I went through really a, a fair [00:09:00]bit about blades yesterday, you know, like why they are shaped the way that they are. So we had to talk about aerodynamics and, um, why they’re made of composite. So we had to talk about, you know, like composite materials, like how, how they, how they work So I don’t know if, uh, people wanna write in comments that m- we should, we should do some sort of, um, poll beforehand to see what are the topics that are most interesting to people, ’cause I think we’ll have a half day, right? So we’ll need to be, we’ll need to be focused. Allen Hall 2025: the description of repairs and what repairs should look like could be tremendously valuable. Everybody who has seen a repair always wonders, “Was that repair done right?” And s- and if you can have some general tools to know, like, “Uh, maybe there’s something not quite right here,” or, “That looks like a solid repair,” that would be a tremendous help to the industry, p- particularly for asset managers Rosemary2: Yeah. And you know what I think is even more useful than being able to pick out when it’s wrong is to be able to know when it’s right. You can– Y-you know, like it is so– [00:10:00] It’s such a relief. Like it takes such a mental load off you when you’re just like, “Yeah, that’s all, that’s all good. That’s normal. Okay, I know that that– I knew that that would happen, so this is not a surprise.” ‘ know, once you know you can make that judgment, you can do it very quickly and focus your attention where it should be, so you don’t need to stress for an hour over every repair. You’re just like, “Yeah. Good, good, good, good, good.” And then, “Mm, please explain why you have chosen to not, not repair this, but just put a Band-Aid over it.” that’s the goal of this training is to get everybody, y-you know, technical people, not people who wanna ever be a blade repair technician. They’ve got their own training that covers what they need to know. But this one is just, yeah, getting people like asset managers or my employees to learn what they need to know about composites, given that they have already got a strong engineering education. So, um, you know, they know a lot of the stuff, but just need to know the composite-specific stuff and wind turbine blade-specific stuff I will run this course again, by the [00:11:00] way, ’cause there was a lot of people who wanted to do it I couldn’t fit in. So it’ll happen at least once. I’ll keep on running it until everybody that wants to do it has, has done it. But, um, yeah, feel free to get in touch  Allen Hall 2025: So if you wanna attend Rosie’s short blade course at WOMA 2027, just visit woma2027.com and register today ​ Allen Hall 2025: [00:12:00] Well, over in Canada, they just approved a, really a wind farm big enough to power a small city, and almost none of the electricity is going to the grid, which is a very interesting aspect to some of the things that are happening in Canada at the minute. So up in Nova Scotia, uh, they’ve conditionally approved the Ocean Lake Wind Project. This’d be the largest wind farm in the province’s history. Up to 158 turbines will rise, uh, generating as much as 1.2 gigawatts of power. But this power is not headed to households in Canada. Nearly all of it will be feeding Everwind Fuels’ green hydrogen and ammonia plant at Point Tupper, where clean electrons will become a fuel that can be shipped across the ocean to Europe. And Matthew, there’s been a lot of [00:13:00] projects like this in Europe that have stopped more recently, particularly in northern Europe and up in Scandinavia, uh, on the hydrogen side. Or at least they’ve slowed them down. Canada seems to be going into that breach maybe to fill that void. And is there a marketplace for this to occur up in Canada?  Matthew Stead: Yeah, I think it’s very interesting. Um, you know, like you say, a number of canceled projects, and in Australia there’s been numerous canceled projects. So I like, um, the analogy or use of the term hopium rather than hydrogen, um, where, um, everyone’s hoping hydrogen will be the answer. Um, although, you know, what I, what I’ve read and understood is that, um, you know, the commercials just don’t really stack up and, um, yeah. So in terms of South Australia anyway, um, there was some major, um, hydrogen, uh, development planned with, um, you know, it, it never stacked up. So, you know, it sounds like a great [00:14:00] idea, um, but I’m not sure that the commercials will ever stack up unless you’ve got that guaranteed offtake for the, for the ammonium Allen Hall 2025: Yolanda, what kind of uphill battle is this to get this wind farm up and running knowing that it’s one customer and that commercial market is a little shaky at the minute? Yolanda Padron: what we saw, they have a lot of ca- caveats, right? So they’ve, they need to secure the customers before they start building and before they do anything, um, behind the meter. But it’s, I mean, it’s, it’s a pretty big wind farm, and it’s pretty far up north. But I mean, we, we talked to someone in, in northern US today who was having icing issues. So I mean, of course we know Canada is no, no stranger to that, if they do make it work, I think it’d be really, really exciting to, to have sort of one technology power another, um, instead of just what we’ve been hearing a lot of the potential data centers and, and just wind po- [00:15:00] powering data centers. Matthew Stead: Why not data centers? You know, seriously, like you said, Yolanda. why not go something that does have commercial demand? Yolanda Padron: we’ve talked a lot about the potential of da- data centers, right? And we’ve talked a lot about people wanting to do them. Um, but there’s also a lot of talk of potentially doing data centers up in space and a lot of talk of maybe what if we do it offshore or, you know. And so I think there’s a lot of what ifs with data centers. Of course, there’s a lot of what if with this, but just from a technology standpoint, I think this is really intriguing to have something that’s, that’s a little bit even more out there than what we’ve heard so far Allen Hall 2025: Is it a build it and they will come type of s- situation here that hydrogen and ammonia may be the, the first offtake, but realistically, if that doesn’t work out, they can still connect to the grid and feed Canada, feed the Northeast of the United States or something else Matthew Stead: Also, um, like Japan has [00:16:00] also expressed strong demand for, um, ammonia, and so, you know, they- they’re on the East Coast, aren’t they? So, you know, shipping it from East Coast to Japan is not gonna be so, so easy. I stick by what I said before. It’s hopium. it’s not a plan Allen Hall 2025: I just saw an article today talking about Airbus continuing on with a hydrogen aircraft, and I think they were gonna work with a Japanese firm to work on that together. Six months ago I thought that died, but maybe it’s still in the offering. Maybe there’s an offtake for hydrogen. B- besides the, you know, replacement for some of the, uh, more unpleasant gases that are used in steel production and in some other industry things, maybe part of this is airplane fuel. Which ammonia is one of those offerings also, right? The, there’s been a number of efforts to turn ammonia fuel into essentially jet fuel. They configure the engines to burn ammonia, which is a possibility. It does seem remote though, [00:17:00] honestly. There doesn’t seem to be a huge pull for hydrogen, and there’s not a, a major market for ammonia at at least at the moment. So I don’t know. It, it’s… When you’re talking about gigawatts of capacity you’re gonna build, you, you hopefully have an offtake  for it  Yolanda Padron: if they designed it for it being not connected to the grid, right, it just is kind of like a behind the meter thing, and then could they later retrofit it into there? Like, how would all that permitting and everything  Allen Hall 2025: I–  well, that’s a great question. I– There are a number of, uh, connections between the United States and Canada at the moment. guess is that when they place this wind farm, they have that alternate route lined up, just like any wind farm in here in the States, that you’ll find them real close to high-voltage transmission lines. Generally, those are the easy ones because transmission lines cost money and take time for permitting. I’m not sure Canada has those kind of restrictions, right? But Nova Scotia is not the easiest place in the world to do heavy construction work, just the [00:18:00] nature of Nova Scotia. It will be fascinating to see how they progress with this, but it’s something to keep an eye on because a lot of other projects like this have slowed down Matthew Stead: Do you remember when some of the OEMs were talking about, um, putting electrolyzers on their offshore wind turbines? So the, the theory, the theory was you’ve got offshore wind turbine, you don’t connect it to the grid standalone, um, and you generate hydrogen or, uh, possibly ammonia on the actual wind turbine. And then every now and then you just decant it, you know, drive up with a boat, you know, plug in the hose, and then suck out the hydrogen or ammonia. So, um, yeah, once again, all of those have gone quiet, haven’t  they?  Allen Hall 2025: speaking of Japan, a global oil giant is walking away from the Japanese offshore wind project, uh, but the project’s not dying. BP has told its Japanese partners it intends to withdraw from a wind farm planned off Yamagata Prefecture, uh, apparently worried about [00:19:00] profitability. The 450-megawatt project sits, uh, just off the coast, and it is led by trading house Marubeni, which says it will press ahead without BP. Kansai Electric and Tokyo Gas remain on board also. So BP’s exit follows really a, a brutal year for Japan, where Mitsubishi has, and some others, have pulled out of, uh, at least three projects so far, uh, over rising construction costs, and I think a lot of that’s tied to inflation. Uh, the ambition’s still there for, uh, for a number of companies, but it’s just getting harder and harder to do projects in Japan. Is this just the nature of the economy in Japan at the moment, or is this more about Japanese policy on the offtake,  Matthew Stead: I, I’m not really deep into the details but, you know, it just appears to me like a blip. I mean, there, I think there’s a lot of commitment in Japan to, you know, carry [00:20:00] out their offshore developments and I, I think this is probably more just a blip, um, and a little, you know, internal corporate, you know, argument rather than a sustained issue on offtake agreements and so forth Allen Hall 2025: Well, Yolanda, how hard is it to keep partners on a wind development in general? Are there a lot of moving pieces there until the turbines hit the water or hit the  earth?  there’s  Yolanda Padron: I think a lot of moving pieces, but not, uh, I haven’t seen a lot of changes once it’s been publicly announced and everything’s, you know, everything’s been signed and everything. Um, I do think this is really interesting. I know we’ve talked a lot about, about having, about the idea of like sometimes people think wind’s really expensive, and the way that we’re gonna make wind work is just making it cheaper for everybody and just optimizing it as much as possible, um, and, and just being, having the turbines be as resilient as possible, right? And I think such a strong player just backing out maybe [00:21:00] will incentivize some of the people in Japan to sort of try to see how they can optimize it a little bit more. I’m really excited to see it. I don’t know. It’d be… I think it’d be a nice it  Allen Hall 2025: Isn’t the bonus to offshore wind the price stability? Although the price may be higher today than you may be happy to pay, the stability of that price is a huge leverage point when you compare it to things like oil and gas or natural gas, um, in particular, which are highly volatile, that for electricity, at least you have this fairly steady source at a fixed price that you can plan out 10 years, 20 years, 25 years, maybe even 30 years. And as batteries become more prevalent on the grid, that the math even gets better over the years. Isn’t that the bonus? And, and if [00:22:00] everybody can focus on the long-term effects to the economy is where all the action will be? Matthew Stead: Yeah, I mean, when I first, um, started looking into wind, you know, 10 plus years ago, I, I won- wondered why. Why would you build offshore with all that expense? And then, you know, it became clear to me just around the, um, you know, the diversity, you know, the, the fact that you might get more wind at times that you don’t get onshore wind, and the fact that it’s more consistent. Um, yeah, and, you know, so those… I- it’s really a trade-off, isn’t it? Between the capital costs and the, um, more reliable, more consistent, um, offshore wind. So I think, you know, I, I was convinced at the start, I thought it was crazy, but then obviously it’s, it’s a, it’s a… it makes sense Yolanda Padron: Yeah, I agree. And I think, uh, depending on where you’re having your offshore wind farm, you run into things that you maybe haven’t run into before, right? I know onshore we run into a lot of things in the [00:23:00]US and Australia that we, you know, the, the turbines just maybe weren’t designed for, or there wasn’t a lot of research being done because it was being done in Europe and, and the conditions are really different. Um, and just the same way, you know, the sea is different in different places. There’s different depths. There are diff- different things that you need to worry about. but yeah, I, I completely agree that there’s a lot more generation, um, offshore. It’s, it’s bigger turbines. Um, there can be bigger, larger costs. You know, if you need to do a blade replacement or something, it, it can get, again, really expensive really quickly. But, but it’s, it’s a trade-off for sure. Allen Hall 2025: We’re gonna take a quick break, but when we come back, we wanna talk about a place where wind is being fought over versus projects slowing down ​ [00:24:00] over in the UK, there’s a big fight about offshore wind, and not just about where wind turbines will be planted, but more about how they will affect other wind turbines. So RWE is defending the UK government’s approval of its three-gigawatt Dogger Bank South project, which won its consent order, uh, basically a month and a half ago. Uh, but the developers next door are taking that approval to court. Equinor, SSE, Vårgrön own the neighboring 3.6-gigawatt Dogger Bank wind farm, and they have filed for j-judicial review. Their argument is technical, but the price tag is not. They say wake effects, where one wind farm steals the wind from another due to turbulence, could cut their output and cost them between €500 million and [00:25:00] €669 million over the life of their project. That’s a lot of money, Matthew. A half a million euros is not something to ignore. It looks like this is headed to some judicial court or maybe arbitration. Wake effects, which are actually not that well understood from what I can tell at the moment, there’s a lot of discussion and argument about, uh, how real are they or, or what effect they can have on power output. Uh, there’s a lot of money at stake, and the location of some of these wind farms is pretty close to one  another  Matthew Stead: you know, we always, always talk about, you know, AEP loss and, you know, the, the challenge is actually measuring it. And, um, you know, I’ve heard different numbers, but, you know, plus or minus half a percent of AEP loss, um, appears to me from what– in discussions, you know, the, the limit of what you can actually ever measure on a good day. Um, I just wonder, I mean, while those numbers, you know, €500, um, [00:26:00] million is a, is a big number, um, but what is that as a percentage of the overall output of that, of that facility? Um, I, I don’t know the answer, but, you know, if, if it’s, you know, half a percent, I think you’d be struggling to, um, struggling to justify that, that wake effect loss. I mean, you know, going back to what you said, Allen, you know, there are wake effects of some sort, but it’s a question of how much. I mean, that-that’s why aircraft don’t take off, um, too closely, isn’t it? Because there’s wake effects. Um, so it’s definitely a given, definitely a given. Um, but, you know, how much of an impact it truly is. Um, and I mean, there’s always other variables, you know, variables in the weather, you know, wind patterns, da, da, da, da, da, da, da, and how much do this– does this actually compare to those other, other variables?  Allen Hall 2025: Yolanda, how would you even mitigate wake turbulence on an adjacent wind farm? Are there ways to do that today?  Yolanda Padron: I think the, the aerodynamics, Allen, would [00:27:00] be a lot more in your court than, than in mine. Matthew does have a really good point. I mean, what are we… With the UK wanting to ramp up offshore as much as they want to ramp up, right? They’re not going to just cancel a large project, and they need to… I mean, it’s not, uh, there’s a finite amount of space, right? So what, I mean, what, what are you, what are you gonna do? It’s like, it’s what, like, what happens in onshore where you, you really hope maybe that you don’t get a wind farm that’s really, really close by. Um, but you might also want to plan for it. I mean, I know of sites that have le- that lease a little bit of extra land so that way no one else can lease it, or that they can, they can use that to, to travel between turbines. Um, and it’s, I mean, it’s, it’s kind of… Isn’t it kind of just part of it, part of the trade? Allen Hall 2025: it has to be, right, at some point. [00:28:00] The question in my mind about all this is how much wake is there? Is it directly impacting the adjacent wind farm? Is there– are there things that can be done to minimize that wake turbulence? I think the answer is yes, but as wind turbine blade designers, I haven’t seen the same level of wake reduction that we have seen more recently in aerospace. It’s complicated to do some of these things on a wind turbine blade. You’re mass-producing. You’re making a blade a day or a blade in a day-and-a-half timeframe. Are you gonna design this really aerodynamic tip to go on to reduce the wake on a particular wind farm? Probably not, right? So it’s, it’s– is it worth doing that versus the, the cost it would be? So it’s gonna cost 500 million euros in loss to an adjacent wind farm. Do you put that 500 million into the design effort and the molds and [00:29:00]everything else to make these blades different? Uh, it’s a tight trade-off, right? It– from the engineering side. It may be better settled in the courts, honestly. Just it may be cheaper to do it that way. Matthew Stead: Uh, I, I was gonna go down a different avenue. I mean, obviously there’s always curtailment. There’s always curtailment due to grid congestion, et cetera, et cetera, et cetera, maintenance. I mean, if they, if they just– when wind is coming from a certain direction, they could just de-rate and, uh, just not absorb as much energy, um, out of the wind when the wind is coming from that sector. And so that would be a way of, um, not modifying the turbine, just de-rating it under a certain wind condition. I mean, the same thing occurs with noise curtailment all the time. Um, so there’s, there’s noise modes. There could be a, a wake loss mode. We should trademark that Allen Hall 2025: Well, you know who’s gonna make money out of this no matter what? The  lawyers.  ​ Allen Hall 2025: [00:30:00] Well, in this quarter’s PES Wind magazine, there are a number of great articles, and you can download the entire magazine and all those great articles at peswind.com. There’s a nice little article from Enerpac Tool Group, and if you’re not familiar with them, they make a, a number of tools that are handy in the wind industry. Uh, and, you know, routine torque checks is kind of a pain, right? And the problem with a lot of those checks is that you have to haul around a heavy hydraulic pump to do it. And so if you’ve ever been to a trade show and seen some of these [00:31:00] pumps, it is a pain. And if you h- have to move around, especially on a w- wind site a lot, you really don’t wanna have a heavy pump that maybe is made for something, uh, more robust. Uh, and you need something that’s portable. That’s what you really need, right? So the Enerpac Tool Group has really created this, uh, LU series they call. Which is a lightweight, portable, hydraulic pump, which is for intermittent work, which is what happens on most wind sites. It’s intermittent. Uh, so the product line director, Angie Wallace, uh, talks about this and says technician feedback has shaped this new tool, uh, from multiple carrying handles and an upward-facing gauge. And that is a big thumbs up from me. When you put the gauge on the side of the tool where you can’t see it, such a problem. It’s like they’ve never used it. Well, obviously, the Enerpac has been talking to technicians, and they put the gauge where the technician can actually see it. Uh, and it’s designed to go through towers and, and tight [00:32:00] spaces. Uh, so this is made specifically for offshore conditions. It’s ruggedized, and it’s a great tool. And a lot of times, Matthew, when you s- see the technicians about and some of the tools they carry, you’re like, man, that is not a good tool for this. That is, that is too much to be hauling around, particularly uptower. It’s nice that we can see some tools that are designed job Matthew Stead: I, I’m completely convinced. I, I don’t have much to say. Um, I mean, my, my day job is, um, you know, designing products and working out what products we’re going to, to work on, and, you know, the customer is the main voice you should listen to, um, at least in the first step. So always listen to the customer first, and I think from what you’ve described, customer first, and then develop the product to suit the application. Yeah, so yeah, I’m convinced  Allen Hall 2025: Yolanda, you’ve seen Interpack on sites, haven’t you? It does seem like I run across them once in a while at some of the US sites  Yolanda Padron: Every once [00:33:00] in a while. I do gotta say I love the idea of when, like, actual, like, boots on the ground people’s feedback is taken into consideration for, for anything really. And so this is, this just makes me really happy because I think a lot of times, like, as engineers, like, we love the idea of just, oh, I’m gonna do this really cool fancy thing, and then it’s just it- no one can use it, or a very specialized person has to be able to use it. And so actually doing, you know, modifying a product so that it, it makes sense for the people using it, and I know we’ve, we’ve all talked about it a lot internally and, and we continue to work towards making it easier and easier on, on the people actually installing the product. Like, this is, this is really exciting. Allen Hall 2025: So if you need a lightweight pump for tightening some bolts uptower, particularly if you’re offshore, take a look at this Enerpac line of LU lightweight series tools. It’s well worth it. And at that same time, you should check out PES Wind magazine. Just go to [00:34:00] peswind.com That wraps up another episode of the Uptime Wind Energy podcast. If today’s discussion sparked any questions or ideas, we’d love to hear from you. Reach out directly to Rosemary, and don’t forget to subscribe so you never miss an episode. for yolonda, Matthew, and Rosemary, I’m Allen Hall, and we’ll see you here next week on the Uptime Wind Energy podcast.  ​

Una gran parte de aprender a tratarnos mejor y ser más autocompasión es reconocer una simple pero necesaria idea: es la de entender que todos estábamos en esto.

Una gran parte de aprender a tratarnos mejor y ser más autocompasión es reconocer una simple pero necesaria idea: es la de entender que todos estábamos en esto.

Happy Fourth of July to all of the Americans all over the world. Join Spencer, Ty, and Andy as they rank historical figures from all over time and space with no thought pad to any cultural or historical context. We're mostly doing it off of vibes. Yep, it's pretty much entirely vibes-based. Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

Laudetur Jesus Christus - Ngợi khen Chúa Giêsu KitôRadio Vatican hằng ngày của Vatican News Tiếng Việt.Nội dung chương trình hôm nay:0:00 Bản tin15:59 Chia sẻ Lời Chúa : Lm. Đaminh Vũ Duy Cường, SJ, chia sẻ Lời Chúa Chúa Nhật thứ 15 thường niên---Những hình ảnh này thuộc Bộ Truyền Thông của Toà Thánh. Mọi sử dụng những hình ảnh này của bên thứ ba đều bị cấm và dẫn đến việc đánh bản quyền, trừ khi được cho phép bằng giấy tờ của Bộ Truyền Thông. Copyright © Dicasterium pro Communicatione - Giữ mọi bản quyền.

Break/Fix Podcast welcomes Greg Stanley, host of The Collector Car Podcast and author of The Enthusiast's Guide to Collector Cars, to discuss his path into the collector-car world and how a passion project podcast led to consulting as a car specialist with RM/Sotheby's, including sourcing cars such as the Ferrari 330 LM/250 GTO that sold for $51.7 million. Greg shares the story behind his rare, unrestored, family-owned Pontiac GTO in Tiger Gold and his multi-year effort to get it featured on Jay Leno's Garage, including an on-site interview with Leno. He reflects on interview insights, shifting focus away from pure valuations toward stories and experience, and explains why he wrote an entry-level guide to help newcomers and even non-enthusiast partners understand the hobby. We also discuss generational trends, engaging kids at events, the “last analog” era, future collectibles, and Greg's upcoming Porsche enthusiast guide and book tour schedule. ===== (Oo---x---oO) ===== 00:00 Meet Greg Stanley! 01:17 Origin Story: Starting a Car Podcast and Breaking Into RM/Sothebys 05:21 Life as Car Specialist 05:55 The Tiger Gold GTO & Getting on Jay Leno 14:53 Interview Surprises 17:45 Market Trends and Values 21:00 Generational Shift in Collecting 22:47 Why Write the Guide? 25:41 Getting Kids Into Classics 29:15 The School Bus Story 31:35 Pitstop: Porsche 959 vs F40? 34:50 Writing the Enthusiast Guide and avoiding Jargon for Newbies 40:37 Inside the new Porsche Book 45:06 Future of Collecting, Top Future Collectibles? 52:06 What's Next for Greg? 52:45 Where to Buy the Book and Farewell ==================== The Motoring Podcast Network : Years of racing, wrenching and Motorsports experience brings together a top notch collection of knowledge, stories and information. #everyonehasastory #gtmbreakfix - motoringpodcast.net More Information: Visit Our Website Become a VIP at: Patreon Online Magazine: Gran Touring Follow us on Social: Instagram

Dan Fesenmeyer, Managing Partner at WindQuest Advisors, joins to discuss the repowering rush and the FAA permitting stall, rising O&M costs on larger turbines, tariff pass-throughs, and AI data center demand. Sign up now for Uptime Tech News, our weekly newsletter on all things wind technology. This episode is sponsored by Weather Guard Lightning Tech. Learn more about Weather Guard’s StrikeTape Wind Turbine LPS retrofit. Follow the show on YouTube, Linkedin and visit Weather Guard on the web. And subscribe to Rosemary’s “Engineering with Rosie” YouTube channel here. Have a question we can answer on the show? Email us! Welcome to Uptime Spotlight, shining light on wind energy’s brightest innovators. This is the progress powering tomorrow Allen Hall: Dan, welcome back to the podcast.  Dan Fesenmeyer: It’s great to be here. Great to see you again.  Allen Hall: There is so much happening in your particular area. Your name pops up quite a bit within Weather Guard because, uh, we’re dealing with a lot of operators and- A number of times we’ll ask them, “Have you read your turbine supply agreement?” “No.” “Have you read your full service agreement?” “No.” “Well, maybe you should do that.” And then we say, “Have you talked to Dan? You should call Dan, ’cause he can help you understand what you have signed.” Mm-hmm. “Oh, that’s probably a good idea.” So now that you’re here, WindQuest Advisors, of course, obviously is your company. Mm-hmm. And you’re talking to a number of operators. The, the big hurdle at the minute, the nearest short-term hurdle, is repowering. There’s just a lot of [00:01:00] repowering efforts going on- Mm-hmm … trying to get turbines in, start a project. There’s a July 4th deadline and an end of the year deadline. There’s a couple deadlines after that. What are you seeing right now from operators i- in terms of repowering? What’s the effort happening?  Dan Fesenmeyer: Well, there was a ton of effort to start physical work. That window’s obviously closing-  Allen Hall: Yes …  Dan Fesenmeyer: very quickly, but it’s still open. Uh, and then once you’re past that window, my understanding is if you get your repower completed by the end of ’27, you didn’t really need to have started physical work. But I think most folks, start physical work is kind of the insurance piece of it-  Allen Hall: Sure …  Dan Fesenmeyer: if things take longer. Uh, another thing that’s popped up is obviously FAA and other permitting.  Allen Hall: On the permitting side, from the federal’s, uh, standpoint, is that stopped? Or, or are projects able to continue putting turbines in the ground, or what’s the status? Dan Fesenmeyer: My- From what I’ve seen, I think on the opening session here at [00:02:00] ACP, it was said, they said that there’s, like, 130 projects that are-  Allen Hall: At least …  Dan Fesenmeyer: caught. Yes. And I’m, I’m involved with some of them, and I have a fairly small shop, and there’s just no FAA variances or permits or- They’re not issuing- … mitigation studies. Everything seems to have stopped.  Allen Hall: So they’re not even reviewing the documentation that’s been submitted by the operators at all?  Dan Fesenmeyer: That’s what it seems, yes. Yeah.  Allen Hall: Is that legal? Uh, uh, usually those federal requirements have a timeline which they’re able to review those permits and get them approved or disapproved them. You’re s- Right … I think what I’m hearing is, what you’re saying is they’re not even looking at them.  Dan Fesenmeyer: That’s correct. That’s what I’ve heard and seen.  Allen Hall: Okay.  Dan Fesenmeyer: Yeah. Yeah.  Allen Hall: So what is an operator to do then? How does this, how do they meet some of these deadlines if they can’t get the permit?  Dan Fesenmeyer: Well, I mean, it stalled a lot of projects ’cause of the associated risk with it. Although I’ve seen some, uh, you know, some repower folks think, “Well, you know, I’m just repair- repowering like for like, or I’m not changing much.” [00:03:00] But if your, if your rotor’s changing or pad location’s changing, you need to update those permits.  Allen Hall: So the, the groups and the operators that are repowering the existing turbines are putting basically the same turbine in the same hole. Dan Fesenmeyer: Well,  Allen Hall: I- Would that be okay?  Dan Fesenmeyer: I would say originally- The initial push on repower was kind of your larger rotors- Sure … new drivetrain, et cetera. Yes. The market seemed to shift more towards, “Hey, let’s do smaller upgrades, component exchanges.”  Allen Hall: Okay.  Dan Fesenmeyer: Getting more towards the minimal investment, so to speak.  Allen Hall: The 80% investment portion. Dan Fesenmeyer: Yes.  Allen Hall: Right.  Dan Fesenmeyer: Yeah. And less about, you know, a big new machine head, for example.  Allen Hall: Well, if that gets you through and gets you the, the, uh, tax credit started back up again, which is the whole point- Right … there would be a reason to do that.  Dan Fesenmeyer: That’s right.  Allen Hall: Is there a marketplace then for those components if you’re gonna repower a GE 1.5 machine, which there’s a lot of them- Mm-hmm in the United States? Are you seeing a big emphasis to go get a new gearbox, [00:04:00] to upgrade the blades- Yeah, and, and- … kind of  Dan Fesenmeyer: thing? Or just do maybe a drivetrain and s- Okay … and leave the rotor or, or-  Allen Hall: So do a gearbox and-  Dan Fesenmeyer: Yeah. Gear or just full drivetrain- Or generator … or yeah, s- things like that. And, um- Wow people are comfortable doing it, and then it’s e- it’s easier, obviously.  Allen Hall: Sure. It’s faster.  Dan Fesenmeyer: And faster, and you don’t necessarily have to touch permits or, yeah.  Allen Hall: And is part of that repowering, I know one of the questions- Mm-hmm … that’s been bandied about quite a bit is, do I have to buy a, a new generator or a new gearbox, or is a refurbished gearbox enough to check the box in terms of upgrading or putting 80% of the value back into the turbine to qualify for those tax credits? Dan Fesenmeyer: I’m not a tax expert, but I’ve seen people do both.  Allen Hall: Okay. Well, that’ll tell you.  Dan Fesenmeyer: Yeah. Yeah.  Allen Hall: They’ve obviously talked to- Right … tax advisors about that.  Dan Fesenmeyer: It’s, it’s their level of risk and whether they have outside tax money or whether- … they’re kind of balance sheet or taking it themselves. It’s, it’s- Yeah … more of a risk profile that [00:05:00] everybody’s different on. Allen Hall: Okay. So that has changed the landscape quite a bit. So now it’s, once this window of opportunity passes by, we’re into brave new world. Mm-hmm. And operating turbines now not really 10 years, operating till end of life, which could be 20, 25 years. Have operators started thinking about that and starting to address some of the, the, especially the contracts around that? Are they starting to rethink contracts? Are they starting to approach full service agreements differently? Is, is the marketplace changing in the US?  Dan Fesenmeyer: Yeah, I think so. I mean, it, it, depending what you have and what you’re doing, whether you have an existing agreement or you need a new one, and whether it’s a renewal or if you’re doing, let’s say, a drivetrain or new machine head, then there’s usually a service contract that’s going to come with it- Sure ’cause it’s essentially a new machine. Largely a new machine. Largely,  Allen Hall: yeah.  Dan Fesenmeyer: But in the case of a gearbox, right, you’re probably out of your longterm O&M agreement anyway, and, uh, whether you’re… And you probably [00:06:00] have, you don’t have the unplanned coverage anymore. Right. So it’s really, you’re on, you’re kind of on your own risk. Allen Hall: Okay, so that’s the repower scenario. Mm-hmm. What’s happening new turbine-wise? It seems like the, a lot of the operators are choosing six megawatt, seven megawatt, eight megawatt machines tends to be the, the, the band of opportunity for a lot of operators. What are they working on right now in terms of, uh, TSAs, full service agreements? What are you seeing out on the landscape US-wise?  Dan Fesenmeyer: Well, I think, um, the TSAs haven’t changed much.  Allen Hall: Okay.  Dan Fesenmeyer: But the- The, the scope and the risk has changed a bit, and the, the OEMs are, you know, holding their cards closer, and it’s hard to get to certain terms that– harder than it used to be.  Allen Hall: So let’s, let’s talk about that for a minute because, uh, there’s been some recent reports speaking to the O&M costs for larger machines. And so the, the goal was if I went from a [00:07:00] two-megawatt machine to a six-megawatt machine, my O&M cost may be 3x because of the size of the turbine, but ideally they drop. That, uh, the same amount of effort into a larger, m- newer machine, uh, so, uh, my spend wouldn’t go up that much. In, in some places on the planet that I’ve seen feedback about that is that the O&M costs are not 3x, they’re 5x. So the, the cost to operate the turbine, the six and eight megawatt machines, is higher than it would be proportionally to a two-megawatt machine. I think operators are just trying to start to figure that out. Are the OEMs already knowledgeable of that fact and are s- trying- I, in, in- … to phrase the conversation I  Dan Fesenmeyer: mean, in the pricing that you get from the OEMs for the full scope agreements, that’s largely in there already.  Allen Hall: Yes.  Dan Fesenmeyer: And I always tell people look at it on a dollar per kWh or dollar per megawatt hour- Ah … basis versus a dollar per turbine, and you- Sure … you’ll see a different number.  Allen Hall: Different calculation done. Dan Fesenmeyer: Right. But [00:08:00] these, these larger machines, they need larger cranes. They need tall– Yeah, they have taller towers, so a different crane setup, and these components become very, very large. So- Everything gets harder … everything gets d- more difficult. In a basic sense, it’s still oil and gearbox and, you know, tho- tho- Right that kind of basic service. But when you get into major components and more major maintenance items, then it’s bigger, it can be harder.  Allen Hall: So what does a operator think about that now that they have a little bit of experience? Obviously SunZia, which is a huge project, three and a half gigawatts, uh, a l- several hun- like around 900 turbines, all of them bigger turbines. It’s a r- for, uh, really the first real taste in America of larger turbines. What are the operators thinking about that, and how are they thinking about what sizes to go with in the future? Or, or, or do they not really have a choice? Like, GE offers six, Vestas offers six, Siemens will offer a six or a seven, [00:09:00] so those are your choices. They’re– You’re not able to get a two megawatt machine anymore.  Dan Fesenmeyer: I mean, I think, uh, it really comes down to your, your site. Okay. And the larger machines are generally better when you have land constraints or, uh, y- your, your wind resource varies very differently. Think of a ridgeline, and you only have a certain number of pads. But generally, it’s kind of a pad constraint to push you to the larger, and then your smaller, “smaller,” four and four to four and a half- … megawatt machines, those are still kind of the workhorses of, of the US, in my opinion. Their NCS better, they’re e- they’re lower cost, but you need more pads. So it’s always that trade-off of pads versus space, spacing, uh, and in the end, you just want to get the most AEP out of that site. Allen Hall: In terms of marketplace, are you seeing prices generally rise dollars per megawatt on [00:10:00] new turbines? ‘Cause the, at least the market indication is that, uh, some of the OEMs have- Real strength in the marketplace today. This is an, an OEM-strong market. They can set- Mm-hmm … prices now. There’s fewer players. China has been eliminated from a lot of lo- locales. Mm. So they don’t have the competition. That allows them to raise prices. Are you starting to see that flow down in some of the contracts, that, hey, the prices are going up? But, but i- inflation has been a big part of that, too. Well,  Dan Fesenmeyer: yeah, yeah. I mean, there’s… And tariffs, right? The, uh, that, that’s the most interesting one right now, and you have to kind of peel apart what’s my pre-tariff price versus my post, and then what’s the exposure if these tariffs change? And-  Allen Hall: Is that in the contracts now? Are they able to write contracts that tie them to what the tariffs could be, so your final price really depends on what the tariffs are today or tomorrow?  Dan Fesenmeyer: It’s generally… Well, things have changed and, and things are always fluid, but, [00:11:00] but most recently it’s, “Well, here’s what the tariffs are today,” and when we either bring in the component or when the OEM’s actually paying that tariff, it’s kind of a pass-through  Allen Hall: in essence. So they’re just handing you the, the bill for the tariff- Yeah … in a sense.  Dan Fesenmeyer: I mean, that- that’s it. And then you can maybe negotiate and do some things around that to share risk a little bit. Mm-hmm. But the basic premise is, you know, there’s transparency on here’s the countries and the tariff rates. If these change, that’s on the buyer. Allen Hall: So the OEMs are trying to address that in, in some form w- by moving production into the United States. Vestas has a large blade facility in Colorado. They’ve been expanding that over the last several months. They’ve been hiring quite a bit. Uh, GE with LM up in North Dakota and TPI, and all the discussions around TPI at the minute is to really bolster their supply chain. Uh, they’re trying to get away from the tariffs as much as they can. Are, [00:12:00] are you… You think you’re still gonna see more of that where a Siemens, a GE, a Vestas are gonna be investing more in the United States to avoid that tariff, or is it just impossible?  Dan Fesenmeyer: I, I mean, I think you… What they’ve done, I… It seems to me, I’m not obviously an expert on that, but it- they’ve moved things where they can And to capture- Mm you know, where you already have capacity. But starting, yeah, building a new plant somewhere, I’m not sure how wise that is in the environment that we’re in.  Allen Hall: Yeah, you saw a lot of plants that were proposed two, three years ago that have, were never built. It does seem like existing plants that were on site that were closed got reopened. Kansas, Iowa- Mm-hmm … some of those plants got- Mm-hmm … started over again, which is easier to do, which makes a lot of sense. So they’re going after the, the easiest things first still. We’re in that phase of we’re not gonna put a lot of money into the United States however. We’re gonna utilize what we have and maybe grow what we have. Dan Fesenmeyer: Right. Or, or similarly, you can move from, if you have more of a… All these supply [00:13:00] chains are global at this point.  Allen Hall: Sure.  Dan Fesenmeyer: But if you happen to have a factory in a country with a lower tariff and versus one that’s higher, maybe you move that. You’re not bringing it over to the US, but you’re moving from, let’s say, India to the UK. Allen Hall: Sure. So, so- Okay, so there, there’s a lot of sh- card shuffling going on- Yeah … to avoid tariffs.  Dan Fesenmeyer: Yeah, and unfortunately then the tariffs change and- … perhaps you have to change back. And, and the other one, uh, that’s out there, obviously the Supreme Court had their ruling on tariffs, so folks are waiting for a Section 232, which is  Allen Hall: still- Untouchable, in a sense? Uh-  Dan Fesenmeyer: Well, it- people are just waiting for what, what will Section 232 be. And it’s been looming for months now.  Allen Hall: Over a year.  Dan Fesenmeyer: Yes. So, and, you know, we’re waiting, I guess.  Allen Hall: Is the feeling about that in the industry, uh… I’ll, well, I’ll use a couple of good examples, I think, which, uh, offshore wind being a real stress point United States, and a lot of [00:14:00] the administration’s work to limit offshore development got stopped in the courts. So anything that was sort of building turbines, putting, had ships out, putting- Mm … uh, monopiles in, they never got stopped. They were delayed a couple of weeks, but they were never really stopped, and it feels like from the outside looking in, is that the courts are not gonna allow some of these, uh, movements by the administration to take effect. Is the industry in the United States seeing the tariffs and some of the more extreme things that are happening as temporary or, or are they being a little more cautious, saying, “Yes, offshore wind has won a, a number of lawsuits”? But we may not. And th- with the Department of War and 232 and all those events that are happening, what is the outcome there, and w- how are operators thinking about that? Dan Fesenmeyer: Well, I think we’re in a, in a market where if you have a project that can get built within this window-  Allen Hall: Yeah …  Dan Fesenmeyer: and [00:15:00] you’ve safe har- Like, those projects- And you’re, you’re just in … are desperately moving forward.  Allen Hall: Okay.  Dan Fesenmeyer: Then- ‘ Allen Hall: Cause the trend has been, if you can get it in the ground, they’re gonna let it be developed. They haven’t been able- Right … to stop anything halfway through. Well,  Dan Fesenmeyer: other, like, the FA is a good example of it-  Allen Hall: Sure …  Dan Fesenmeyer: being stopped. But- Yeah … if you have a project that’s being built, you’re moving forward, and then projects that are outside the window, it’s more of a greenfield development view of, of life. And seems like some folks are selling p- assets, some folks are buying- A  Allen Hall: lot of that …  Dan Fesenmeyer: development assets.  Allen Hall: Let’s go down that pathway for a minute because I did think- Yeah … that’s a very interesting piece to what’s happening in the United States at the minute. There’s a lot of transactions, big dollar transactions happening for wind- Mm-hmm on buying, selling portfolios, not just farms. It used to be farms. Right. We’ll sell a farm. Yeah. It was. We’ll swap farms, that kind of thing. Now it’s like, uh, would you like our whole portfolio, wind, solar, battery?  Dan Fesenmeyer: Mm-hmm.  Allen Hall: Is that playing into a lot of the decisions that are [00:16:00]happening on the ground right now, that a, a developer or an operator that has assets is saying, this is a prime time to sell. There’s a l- I have my tax credits already locked in. We’re golden here- Mm-hmm … for several years. The value is never gonna get higher. I need to get out. I- is that the marketplace today, is-  Dan Fesenmeyer: I think for some. I mean- Yeah … everybody’s got different, uh, motivations, whether they wanna get into wind, get out of wind, greenfield versus repower. Uh, it, it’s, it’s really their view of the world and their risk profile moving forward, and whether this is a short-term play, long-term. Do we wanna get out of wind? Some people are essentially doing that. Uh, it’s, it’s across the board.  Allen Hall: How’s AI data centers playing into this? What are you hearing?  Dan Fesenmeyer: Oh, I mean, that’s what everybody talks about, AI and data centers, and the demand for power is there. And- The [00:17:00] issue that, that a lot of us see is wind and solar and battery can all help with that.  Allen Hall: Sure.  Dan Fesenmeyer: And if you want a gas turbine, that’s great, but my former colleagues at GE are gonna tell you it’s 2030- Yes … or later to get one, so what do you do between now and then? And you’re seeing prices go up, which makes these wind farms look pretty good. Power profile’s nice. Yes. Uh, but you still have hurdles to get, like the FAA, US Fish and Wildlife, all these other hurdles to, you know, that are slowing down wind and solar for that matter too.  Allen Hall: Solar’s been slowed down for sure.  Dan Fesenmeyer: Yeah. Yeah. Yeah.  Allen Hall: Does that change, though, with the demand for power in AI data centers? And it does seem to be a priority in the United States to, to win this AI race. Mm-hmm. Does that loosen some of the reins on renewables to let them go, like just look the other way for a while, while they put a new solar field or wind farm in?  Dan Fesenmeyer: It stands to reason that will happen. Haven’t really seen [00:18:00] it, unfortunately. But I wo- But I think it will, right? I mean, it, it, it, it almost has to at some point.  Allen Hall: There’s a lot of pressure on Washington DC to let data centers start being developed and, and go.  Dan Fesenmeyer: Mm-hmm.  Allen Hall: But a- as you pointed out, gas turbines are hard to get, and they can’t scale up at the rate at which the demand is. Right. So your alternative is something really simple, quick and efficient, which would be wind and solar and a little bit of battery. Yeah. I- is that change in the thinking of operators and how they’re thinking about their assets, one, and two, what they’re thinking about in the future? Or are they trying to hook up with an- a- I mean- a Google, a Facebook, a- Yeah, I  Dan Fesenmeyer: mean, the offtake’s- … SpaceX … there, and that’s generally, you know, it used to be utility PPAs. Then it turned- Right. … into hedge things and C&I. Yeah. And now it’s more, you have this, the data center offtake.  Allen Hall: Is the data center offtake, thinking about it from a, a financial standpoint, which they’re probably not being tied to the grid. At [00:19:00] least a lot of these, or at least the talk is right now, is the not being connected to the grid to be sort of standalone, feeding a data center, and maybe a piece of fiber optic coming out of the data center. But that’s essentially it. Maybe some backup power on the grid just in case things go horribly wrong, but standalone power for data centers does make sense. It would, it would seem to lessen the requirements on wind and solar in terms of interacting with the federal government or the, the power company in a sense. Does that make wind and solar a little more viable because it’s not connected to the grid?  Dan Fesenmeyer: Well, I mean, it will be connected to the grid because when the wind stops blowing, the utility will usually, you know, or, and the sun stops sh- shining- Sure uh, the utility will kind of provide that power. That w- Or the gas turbines that they have would- Gas turbine will kick  Allen Hall: in, right.  Dan Fesenmeyer: Yes. Yeah. But, but generally speaking, you’re never truly off the grid, but it does speed things up with interconnection and, and, you know, your T&D [00:20:00] line is much shorter.  Allen Hall: Right.  Dan Fesenmeyer: Or not, you know- Much much, much shorter. Yeah. Depending where the, the resource is and versus the plant or the, the data center.  Allen Hall: So what are the things that we don’t know in the industry that you’re in touch with that we should know? ‘Cause there, there must be a lot happening behind the scenes that we don’t hear out in public or in the common spaces of some of these conferences that are happening behind the scenes. What is, what is the status right now? What do you think the status is of wind?  Dan Fesenmeyer: I mean, it’s, I, I, I’m a big sailor, and sometimes the wind’s blowing hard- … you’re going fast, and sometimes you sail into what we call a hole- Yeah … and it’s just dead quiet. We’re not quite there yet, but, um, it, it’s kind of we’re going through a bit of a lull right now. And I think, I think what people don’t realize is the multiple roadblocks that the industry’s facing. In the past, we’ve had PTCs lapse, and the question is when and if it [00:21:00] will be renewed. Yeah. Now you have other roadblocks, you know, whether it’s, again, FAA, Fish and Wildlife, permitting, different localities. Some… And this goes back to the data center. A lot of local, you know, communities don’t want a data center.  Allen Hall: Right. There’s a lot of-  Dan Fesenmeyer: Right? And they’re like, “Well, wait a minute. My power prices as a citizen are gonna go up- True … because of it.”  Allen Hall: Yeah, it’s true. We’ve already seen it.  Dan Fesenmeyer: Yeah. Yeah. So, so there’s a lot of just new barriers that have come up. Allen Hall: Okay. That-  Dan Fesenmeyer: But wind developers are an extremely resilient bunch, and-  Allen Hall: This isn’t the first rodeo-  Dan Fesenmeyer: Right …  Allen Hall: where they’ve had these issues pop up- Yeah … and PTCs stop and other world forces affect the industry. What’s the outlook over the next three to five years, do you think? Different administration in a couple years, maybe different outlook, more demand on… for power, AI data centers. Is- it just gonna [00:22:00] overwhelm any resistance to wind and solar and battery?  Dan Fesenmeyer: I mean, it, it, that’s kind of a crystal ball, but I think if these data centers start getting built out like people think they will, there’ll be demand for power. And, now we’re talking basic economics, Supply, demand. People need power, then power plants will get built and, whether it’s gas, wind, solar-  Allen Hall: All of the above  Dan Fesenmeyer: All of the above, right? And, and I think it will ultimately follow that. I think the, administration will let you know if there’s not enough power or power gets too expensive, something has to break and fill that gap  Allen Hall: because- So let the economics play out a little bit. Dan Fesenmeyer: Yeah, right? Yeah. ‘Cause we’re, we’re voters, right? And- Sure … and, um, people vote often with their pocketbooks.  Allen Hall: And wind and solar are cheap sources of energy, and they’re gonna come to the top of the list almost every time.  Dan Fesenmeyer: Yeah.  Allen Hall: Yeah. Yeah. Yeah. I, I agree with you. Uh, it’s good to see you again. We saw you a few months [00:23:00] ago at WOMA in Australia, and that was wonderful. And I tell a lot of the operators we talk to, “You better be talking to Dan and WindQuest Advisors because you really need to understand what your contracts say and the contract you’re signing, and you need to have a better sense of what’s happening, a little more broader speak in the United States and elsewhere- Mm-hmm and they should be talking to you.” So how do they call or how do they contact WindQuest Advisors to get started?  Dan Fesenmeyer: Well, www.windquestadvisors.com or reach out to Allen and his team. You’re on LinkedIn. I’m on LinkedIn as well- … both personally and my firm. And, um, ask a friend ’cause I have a, we have- … big networks that everybody… You know, it’s, it’s a small community here. It  Allen Hall: is.  Dan Fesenmeyer: Right?  Allen Hall: It is.  Dan Fesenmeyer: And, and people bounce around different firms and, but people stay connected, so, um, that’s a great way to find each other as well.  Allen Hall: Yeah. Great to see you, Dan. Likewise. Thank you. Thanks for being on the podcast. And yeah, we’ll hopefully see you in Australia in a couple months. Dan Fesenmeyer: Looking forward to  [00:24:00] it.

GE Vernova pumps $1 billion into LM Wind Power, and KKR buys EDF’s US and Canada renewables arm. Plus CIP sweeps South Korea’s offshore auction and the CME plans wind derivatives across three continents. Sign up now for Uptime Tech News, our weekly newsletter on all things wind technology. This episode is sponsored by Weather Guard Lightning Tech. Learn more about Weather Guard’s StrikeTape Wind Turbine LPS retrofit. Follow the show on YouTube, Linkedin and visit Weather Guard on the web. And subscribe to Rosemary’s “Engineering with Rosie” YouTube channel here. Have a question we can answer on the show? Email us! The Uptime Wind Energy podcast, brought to you by StrikeTape. Protecting thousands of wind turbines from lightning damage worldwide. Visit striketape.com. And now, your hosts. Allen Hall: Welcome to the Uptime Wind Energy podcast. I’m your host, Allen Hall, and I’m here with Matthew Stead and Yolanda Padron. Rosemary is at GWO training this week. And we have an announcement about Wind Energy O&M Australia 2027. Matthew, you wanna give all the details?  Matthew Stead: Drum roll Um, very pleased to announce that WOMA 2027 will be at the East Pullman Hotel in Melbourne’s east, uh, not the other one, and, uh, 3rd to 5th of March. Um, the first two days will be two days of wind O&M, uh, conferences, [00:01:00] uh, and then the Friday will be a half-day, uh, training session. More information to come.  Allen Hall: Well, she’s not here, so we can probably just announce it, that Rosemary will be giving a terrific four-hour-long seminar on blades and blade repair, so you sign up now. Matthew, where do you go if you wanna just check out what’s happening at WOMA  Matthew Stead: 2027? Uh, well, actually, it’s woma2027.com.  Allen Hall: Uh, over at GE Vernova and LM Wind Power, there’s been a whole bunch of turmoil over the last couple of years if you haven’t been paying attention. Well, GE Vernova just injected about a billion dollars into that company. So although LM recently has shown very little in terms of revenue, it definitely had needed some capital injection in, uh, at least according to the Danish press, the number of employees at the Danish site is about 20 to 30. So it’s really a fraction of what it once was. But [00:02:00] it does seem like GE is paying off all its existing debt and then giving it a little bit of a cash infusion to keep it rolling. The question really is, is what is GE Vernova gonna do with that business now? Are they planning on keeping it? Are they trying to get s- to get it back to health where they can service the other, uh, OEMs that they manufacture blades for? Or is there a larger action that will happen in the near future? What do we think?  Matthew Stead: Yeah, I’m really confused by this one. I mean, a cash injection just so that you’re not bankrupt on paper is, um, that’s just playing with money as far as I’m concerned. Or I’m not sure if it’s a US term, but, you know, shuffling deckchairs on the Titanic. It doesn’t– Does it change anything? Allen Hall: Well, uh, th- they made no announcements about closing facilities. The LM blade facility in North Dakota still appears to be making blades. There’s the TPI factories, which are going through a transition r- right now, appear to be making GE [00:03:00] blades. I, I assume Gaspé up in Canada is still making blades, at least that’s the story. If GE’s gonna rely upon LM to make blades, they’re gonna need to keep them open. Is, is this more of just keeping the factories open with a skeleton engineering crew and possibly moving the blade design group into the States? Is that– Or India or, or somewhere?  Yolanda Padron: And they’re still selling, right? They’re still selling blades. It seems like they’re still planning on manufacturing blades. Do we think that maybe- They’re just trying to avoid that whole TPI bankruptcy deal to not have to kind of scrap for parts?  Allen Hall: Yeah, it’s a great question. I think TPI has been producing parts at high quantity, and some of the Things I’ve heard from the industry folk is that TPI is really busy in producing quality blades, and it’s like the bankruptcy transaction is not happening, which is great to hear because the [00:04:00]industry needs blades, and there’s a lot of repowering going on in the United States and a lot of activity in general, so they need blades. But does LM continue to be a part of that?  Matthew Stead: Yeah, I mean, presumably the TPI, um, whole story only makes LM more important, you know, more important to have, uh, an additional manufacturer and, you know, providing, you know, options for the OEMs.  Allen Hall: It does seem like, though, the GE offshore, GE Vernova offshore is not a thing. Although I’ve heard a couple of rumors that, yeah, GE Vernova is offering some products for offshore, it doesn’t seem like their heart is in it. I can see that happening. So are they just trying to focus on onshore business, and that’s it for the time being? Just let it play out and, uh, wait until the elections in 2028? I know that’s gonna get me blocked on YouTube, but that, that does feel like what’s happening at the moment.  Matthew Stead: Yeah, I reckon it looks completely like that.  Yolanda Padron: I mean, it also looks like they’re [00:05:00] just kind of trying to play everything a little bit more safe, right? So they are scaling up, but not as fast as they used to, so scaling the blade sizes. And then they’re– it seems like they’re, they’re having their FSAs cut quite a bit shorter than they used to, right? So are they maybe just trying to focus on, like, cash up front and just trying to play it safe until they can get their, their footing right again?  Allen Hall: Or is it focus on key customers? I could see GE Vernova actually doing that, that they have a history with certain operators worldwide, and they’re just gonna focus on producing and delivering for those customers. Because you don’t see a lot of announced orders for GE turbines. Vestas is announcing things practically every week. Nordex is doing something similar. Siemens once in a while. But what you really don’t hear anything from in any quantity at [00:06:00] all at the moment is from GE Vernova. When a company needs cash badly enough, even the crown jewels go on the block. And EDF, the French state-owned utility, has to fund the upkeep of 57 aging nuclear reactors and build six new ones, so it is selling. EDF has agreed to hand its US and Canada renewables business, EDF Power Solutions, to the private equity firm KKR. The business runs 5.6 gigawatts of renewable assets across the two countries. Late last year, EDF’s chief executive floated selling anywhere from half to all of the unit in a deal that could be, well, it’s reported to be about $4.2 billion. That’s the latest news I’ve heard. This is a big transaction. KKR is Canadian, right? And is a massive investment firm Uh, which I, I don’t think have a lot of wind at the moment. Uh, what is the [00:07:00] KKR play here?  Matthew Stead: I, I love this because this is, uh… So obviously I’m Australian, and Macquarie is a big Australian. So, um, Macquarie own a whole lot of wind farm, a whole lot of wind infrastructure. So I just see this as a wonderful g- you know, fight between KKR and Macquarie. And so KKR has a whole lot of, um, they o- they’ve got some, you know, stake in Australian wind farms. They’ve got some work, you know, through Europe with wind farms. So I, I, I think this is a good thing, just a bit more global competition and a bit more global growth. And I think it’s all coming from the data centers and, you know, the future increase in growth of, um, demand.  Allen Hall: Yolanda, EDF’s wind fleet is a variety of turbines, right? They have some GE, some Siemens. Anything else in their portfolio?  Yolanda Padron: I think they have a bit of Vestas there too, right? Is it something that we were saying? It’s– I think this is really interesting. Um, I know that there’s not– I mean, of course EDF is the latest, but there’s some [00:08:00] operators that seem to be, um, consolidating into a bit more of those just higher private equity firms, and it’s– Do we think that maybe this is the way that the US is going to lean towards? I know we talked a lot about leaning towards funding the data centers and maybe a bit more the behind the meter things. Uh, but do we think that maybe that’s the future of the US? There’s a couple of companies that kind of just own all the major infrastructures and then- A  Allen Hall: couple Canadian companies.  Yolanda Padron: And what does it mean for, like, asset management and stuff, like, that’s really, really different from what they’re seeing in their desks in New York and stuff, and just the larger financial models versus what’s happening on the ground, and how will they connect everything? Allen Hall: It’s a great question.  Matthew Stead: NextEra and Dominion, you know, things are only getting bigger. Scale’s, scale’s coming.  Allen Hall: Yeah. I wonder how much, uh, this transaction will have to go through regulators in the US, uh, because it scares me when you have a, a– such a [00:09:00] large foreign national company. There’s actually two involved in here, right? So you, you have a, a French company and a Canadian company trying to transact on, in the United States on a lot of assets. Uh, it probably won’t be that quick if there’s any oversight at all. I, I’m guessing that we’ll hear noise about it. So we’re, we’ll have to keep listening to all the news sources about it and, and telling our valued listeners what’s going on. Because there’s, uh, we know a whole bunch of people that work at EDF and like, love those people and are really concerned about what the future holds for them. I, at least it sounds like upfront that KKR is just gonna continue with operations, but I know, uh, uh, it’s a turbulent time, and if you work there, you, you hopefully things continue the way they’re, they’re supposed to because One of the things about EDF historically has been is that they’re really talented people, that they have hired well over time and that they know what they’re doing. And every time we, Weather Guard and [00:10:00] Yolanda and I’m sure Matthew have dealt with EDF quite a bit They are on top of what they’re operating. They know how their assets work, and they know how to manage them, and so you’d hate to lose those people in a transaction like this. It would decrease the value of the assets, I would say. Very interesting transaction.  Matthew Stead: Yeah. But, I mean, what if the counter, what if, um, this is all part of a, a growth strategy? You know, a growth strategy with wind, solar, and battery, you know, providing more power. So it might actually be an opportunity. So, you know, opportunity to do more and some more exciting work across all three disciplines. Allen Hall: Definitely so. Uh, but it’s a little early. The ink hasn’t dried yet on the contract. So while offshore market pulls back in general, in a lot of places like the United States, another one is racing ahead. In, in South Korea’s latest offshore wind auction, one name walked away with the lion’s share, Copenhagen Infrastructure Partners, CIP. The Danish fund [00:11:00] secured more than one gigawatt of the 1.8 gigawatts on offer, including the single largest project and the only floating wind winner. And the appetite was record-breaking. They had a whole bunch of developers trying to bid on this. You had about 3.7 gigawatts being bid in, more than twice of the capacity available. So for a country that only began competitive offshore bidding in 2022, that’s a few short years ago, that market is coming of age. This is a huge announcement by CIP, right? That, uh, they have bid into the system. They’re, they’re winning, and they’re bringing Siemens Gamesa to the table, which we haven’t heard a lot of Siemens Gamesa’s turbines being selected, but this is a massive order and really gonna help secure at least some portion of, of the Siemens Gamesa business. Matthew, you’re closer to it. In, in South Korea, are you seeing the South Korean industry being built within [00:12:00] the country, or are you seeing, uh, partnerships with surrounding countries like Japan? ‘Cause it doesn’t seem like when– and I’ve looked at some of the South Korea, uh, efforts. It does seem like they’re trying to stand up their own offshore built-in country plan. Is, is that the goal? You think Siemens is gonna end up building a, a factory in, in South Korea for some of these projects?  Matthew Stead: Maybe a couple of things. First of all, I have to apologize. I think, uh, we were talking the other week, and I, I, I sort of implied that floating offshore wind was dead, and I think we copped a bit of flack from that. But, uh, anyway, wrong, wrong on, uh,  Allen Hall: floating offshore is dead.  Matthew Stead: Um, but um, you know, I’ve had a fair bit of interaction with, uh, South Korean, um, you know, Philippines, Japan, obviously. I think they’re all trying to get their industries up, but I, I don’t think they’ve got the scale So, you know, I think they, they really need like the Siemens Gamesas, the Vestas’s, um, to come in and, and partner with them. I just don’t think they’ve got the scale, you know, the, the [00:13:00] installed fleet, the industry to really promote it. And, you know, to get the economies of scale, they’re gonna have to pull in the big existing incumbents. So, you know, good on CIP for, for pulling this off.  Allen Hall: In terms of South Korea industry, I think steel is one of their strongest, uh, industries at the moment, and obviously shipbuilding. Those are the, that go hand in hand, so to speak. There’s a lot of steel in wind turbines, and particularly in floating offshore wind turbines. It would seem ripe for South Korea to get into that marketplace.  Matthew Stead: I’m not sure the intellectual property is in steel tubes. Um, I, I guess what I’m trying to say is the intellectual property is in the turbine nacelle and the blades and, um, you know, I, you know, correct what I said that, you know, obviously the steel and the steel manufacturing in South Korea is, is pretty amazing. Um, but yeah, they’re clarifying what I said before.  Allen Hall: So is this gonna turn into the leading floating project in the world? You know, Greenvolt’s gonna happen in the [00:14:00] UK. There’s some talk of things up in Scandinavia. But in terms of speed, will this be one of the leading candidates in t- in getting things in the water just because of the capability of South Korea to, to build at scale? I  Matthew Stead: think it’s really exciting. Yeah, I, I’m, I’m gonna watch very closely.  Allen Hall: I think this is gonna be amazing. I really do.  Yolanda Padron: I was gonna say, could you imagine, like, a, a turbine and a blade where everything is just perfectly manufactured or close to perfectly manufactured? I g- I went to one farm last week, and there were… I mean, it was in the States, and there were so many patches on new blades. I was just talking to the people in operations like, “What’s, what’s going on here?” You know? Uh, so it’s just really… I don’t know. This is exciting.  Matthew Stead: Do you think, um, they’ll build a blade factory, Yolanda? Do you think they’ll actually take on the blades? Yolanda Padron: I don’t know. Uh, I, I mean, it’d, it’d be great for them, I think, right? It’s a new area of business that they’re diving [00:15:00] into.  Allen Hall: If they don’t have to build the building at the port, I think Siemens would be willing to erect something near the shoreline. And in Korea, there’s a lot of major industry right on the shoreline. It would be relatively easy, I think. You know, ev- it sounds easy now because you’re not actually doing it. But in terms of, you know, building a blade factory on the coastline of United States versus doing it in South Korea, South Korea’s gonna be way easier to do that and at scale quickly. That, that one seems like a win-win. I d- if there’s any place on the planet that could do it quick besides the UK or, you know, Denmark, someone like Netherlands, someplace like that, Germany, it’s gonna be South Korea.  Matthew Stead: Maybe that’s a bet, you know. So prove me wrong again. My money at the moment is that Nacelles blades won’t be coming from South Korea. Allen Hall: Well, if they don’t come from South Korea, they’re gonna be on a South Korea-built ship. We’ll be bringing th- those [00:16:00] blades in country. That’s what will happen. So wind is getting its own set of financial instruments, which sounds weird, right? Wind is wind. It’s in a very legacy style industry. The Chicago Mercantile Exchange is planning to launch wind derivatives across three continents, which are contracts that are tied to the grid in Texas, the markets in the UK and Germany, and just the Victoria state in Australia. So today, most weather hedging happens through one-off over-the-counter deals that are sort of hard to trade and thin on liquidity, so it’s not a commodity you can pass around. A standardized exchange-listed contract changes all that. A utility or a wind farm owner could lock in a hedge in about 15 minutes. The contracts would settle against independent data that models how much power the wind should have produced in a given place, likely supplied by [00:17:00] the Finnish firm, drum roll, Vaisala. Plans are not final, but they could go live within months. So they’re hedging on the wind. Does this sound like a smart move, or w- what are some of the consequences of this? Matthew Stead: I think it goes back to that volatility. W- when there’s volatility, people can make money. Um, you know, and a side note, that’s where, that’s where offshore wind comes in because it’s much more predictable. Um, you don’t get the same lulls with offshore wind. Yeah. So I, I, I love all these, these creative ways of, um, generating, generating demand, financial demand. Allen Hall: It can be played though, right? I mean, that’s one of the things about wind, ’cause each turbine is its own separate little power plant that all connect to a substation, so if you have bought a hedge and the substation goes kaput for 24 hours, you could lose your shirt. It does seem kind of risky, depending on what the scale is here. If you’re doing all of Texas or all of [00:18:00] Victoria, maybe that makes a little more sense, but yikes. That’s gonna be a rough market.  Yolanda Padron: Yeah, the market’s already open, right? Like, you can bid day ahead, um, instead of just real-time prices. But so this, this would be really interesting for owners, right? To be able to track that a lot better than just that gut feeling, which obviously I know people working in trading aren’t just going off of their gut feeling. I know it’s a very, very intense thing. Nobody go against me, please. This is very intense, and it’s better– They do a better job than I could ever do. They do great, 10 out of 10. But this– I think this is really interesting for those of us especially who maybe aren’t super in tune with what, uh, all goes into it. So being able to have something that helps you plan it a bit more for, you know, people like you mentioned earlier, the people that have their home batteries in Australia and are just working on the market itself and maybe [00:19:00] not– don’t have those 10, 20 years of experience of, of actually working on the market. So this is, this is exciting.  Allen Hall: Does that explain all the weather sources and the weather companies when we go to a wind, a larger wind or solar event that there does seem to be a lot of people offering weather insights? Is that what that’s about, is they can hedge? If you have a slightly better weather model, that would give you an advantage in this kind, kind– really kind of market? Is that the, the goal of all those weather firms?  Matthew Stead: Uh, absolutely. And, you know, we’re, we’re part of that because, um, ice, ice, um, you know, reduces power output, and ice forecasting and weather forecasting is, uh, really important in, you know, the Nordics, where you don’t want to be promising certain power and find you can’t deliver ’cause everything’s iced up. So, you know, we, we do work with forecasting companies to improve the, [00:20:00] uh, the quality, and it does have a mer-material difference on, on the financial markets.  Allen Hall: So is that something that we can all get paid for? by these weather companies and these, uh, forecast companies if we provide insights on lightning, so to speak, and icing, uh, is that a revenue chain for at least one of us? Matthew Stead: Absolutely.  Allen Hall: Maybe I like this more and more. I was, I was very hesitant of this exchange, thinking like, “Oh man, not a, not another highly leveraged situation with energy. That doesn’t sound smart.” But, yeah, if we can make a small fortune, Matthew, I think we should do it.  Matthew Stead: Fun fact, there was a flight from, um, yeah, from London to Australia the other week, um, and it’s a direct flight, you know, so 17 hours, and, uh, there was a change in the weather. So there was a change in the weather, and that aircraft didn’t have enough fuel to fly to Perth anymore, so it had to land in the outback of Australia.  Allen Hall: No. Did that happen?  Matthew Stead: Yep, because there was a [00:21:00] change in the weather.  Allen Hall: Are there just, like, kangaroos lined up in a runway shape to get the airplane on the ground? Or how do they– Is there a runway out in the outback that would accommodate a large… That’s a large airplane that’s making a London to Australia trip. Triple 7380? It  Matthew Stead: was a Dreamliner. Um, but, um, it, yeah, it landed in Kalgoorlie. So Kalgoorlie’s a mining town. Yeah, they’ve got, they’ve got big stuff in Kalgoorlie. Allen Hall: In this quarter’s PES Wind magazine, in which there is a whole bunch of great articles, a interesting article about grease. Grease not the country, although I would love to go visit Greece. Grease the lubricant that’s in all our bearings and keeps the world moving at any one particular time. Uh, Sh-Shell was talking about doing a lot of research on grease, and when poor lubrication, uh, happens, it’s one of the leading causes of bearing failure. And so when you see a bearing all tore up, usually the first indication is, is there’s something wrong with the grease. Uh, [00:22:00] so Sh-Shell and bearing maker SKF and the University of, uh, Twente joined forces to answer a deceptively simple question: How do you predict when grease inside a bearing will let go? Well, their answer comes down to film thickness. The microscopic layers of grease that keeps the steel from grinding on each other is the magic variable. The work won a major tribology award and is already feeding into, uh, some of the tools that operators use to schedule relubrication before a bearing fails. And It all comes down to lubrication. That’s the lifetime of a wind turbine. There’s so many pieces that are rotating and are heavily loaded with really complicated bearing surfaces. If you don’t have the grease right, it’s just not gonna work. And what’s happening at Shell is one of those pieces, and we’re [00:23:00] learning so much more. And as we, uh, evolve in the technology and become smarter about the molecules we use and how we use them, uh, this is gonna have a big impact. And I know, Yolanda, you’ve been up to– Well, you’ve been to a couple of wind farms recently. Do you s- see– still see huge grease problems that I usually see when I’m on site? Matthew Stead: Mm-hmm.  Yolanda Padron: I didn’t think that was an issue that was gonna go away anytime soon. But it’s good to know that, that there’s something being done about it that’s more revolutionary than just paying someone to clean the turbine every once in a while.  Allen Hall: And the contaminants that get into the greases are a huge problem, particularly where there’s any sort of sand, dust that climbs in. So keeping those joints clear and those rolling surfaces clear is a major effort. And knowing when to relubricate. And, and Matthew, you guys see pitch bearings and all kinds of problems up on blades that are lubricated that have run out of their lifetime early. It does seem like the first thing you see on particularly pitch bearings [00:24:00] is grease on the side of the turbine from them. Matthew Stead: Yeah. I think that’s– uh, there’s even a special code that the, the visual drone inspection companies have. They’ve got codes for, um, grease and so, yeah, exactly, that’s an early flag. But also dust. You know, sometimes dust from the inserts and from the bolts. Yeah. So it’s, yeah, interesting topic.  Allen Hall: Well, I, I think it’s one of the key pieces to keeping the turbines running. And I know if you travel a lot around wind turbines, the, the grease is the thing that the technicians always talk about, and there’s so many different tools to go out and look at these things. But lubrication, we gotta get to it. And, and Shell, and SKF, and a number of others are, are working at it to make, hopefully, our lives a little bit easier. So if you wanna go check out this article by Shell, go visit peswind.com and download a copy today. That wraps up another episode of the Uptime Wind Energy podcast. If today’s discussion sparked any questions or ideas, we’d love to hear from you. Reach out to us on [00:25:00] LinkedIn, and don’t forget to subscribe so you never miss an episode. So for Yolanda, and Matthew, and an absent Rosie, I’m Allen Hall, and we’ll see you here next week on the Uptime Wind Energy podcast.

Esta es la segunda parte de tres sobre cómo trabajar la autocompasión. También puedes escuchar la anterior llamada disminuyendo la crítica interna. 

Review các phim ra rạp từ ngày 04/07/2026:MINIONS & QUÁI VẬTĐạo diễn: Pierre CoffinThể loại: Hoạt HìnhMinions & Quái Vật là câu chuyện vừa náo loạn, vừa ngớ ngẩn nhưng “hoàn toàn có thật” về cách Minions chinh phục Hollywood, trở thành ngôi sao điện ảnh, rồi mất tất cả, vô tình thả quái vật ra khắp thế giới và sau đó phải cùng nhau hợp sức để cứu lấy hành tinh khỏi chính mớ hỗn loạn mà mình tạo ra.LEVITICUS: BÓNG QUỶ - T18Đạo diễn: Adrian ChiarellaDiễn viên: Joe Bird, Stacy Clausen, Jeremy BlewittThể loại: Kinh DịMột cú horror nặng đô và đầy ám ảnh, nơi ham muốn bị đàn áp hóa thành thế lực quỷ dị NUỐT CHỮNG BẤT KÌ AI XÚC PHẠM Trong một cộng đồng Cơ Đốc giáo bảo thủ tại thị trấn nhỏ biệt lập ở Úc, hai thiếu niên Naim và Ryan bị ép tham gia các liệu pháp “chuyển đổi giới tính”. Nhưng càng cố chối bỏ khao khát thật sự của mình, họ càng bị một thế lực tà ác truy đuổi — mang hình dạng của chính người mà họ khao khát nhất: đối phương....ĐỒNG DAO MA QUÁI – T18Đạo diễn: Preaw Chanatip WongpontreeDiễn viên: Panisara Rikulsurakan, Ongart Jeamcharoenpornkul, Win SakulsangpraphaThể loại: Hồi hộp, Kinh Dị, Thần thoạiMỗi lần gieo xúc xắc là mỗi lần quỷ xuất hiện. Trong lúc dọn dẹp một ngôi đền bỏ hoang, một nhóm trẻ vô tình tìm thấy bàn cờ cổ bị phong ấn từ lâu. Sau khi gieo viên xúc xắc bí ẩn và đọc lên những câu đồng dao kỳ lạ, chúng nhận ra mình đã bắt đầu một trò chơi chết chóc. Mỗi lượt chơi đều kéo theo sự xuất hiện của những linh hồn đáng sợ, buộc cả nhóm phải tìm cách phá giải lời nguyền trước khi trời sáng.NGÀY CON SỐNG LẠI – T13Đạo diễn: Koreeda HirokazuDiễn viên: Haruka Ayase; Daigo Yamamoto; Rimu KuwakiThể loại: Khoa Học Viễn Tưởng, Tâm LýMột cặp vợ chồng đau đớn trải qua nỗi đau mất con. Để lấp đầy khoảng trống trong tổ ấm nhỏ, cả hai quyết định nhận nuôi một rô-bốt hình người hiện đại, chăm sóc nó như con trai ruột. Nếu một cỗ máy có khả năng biểu đạt cảm xúc, trao đi và nhận lại tình yêu, liệu sợi dây kết nối này có kém phần thiêng liêng hơn huyết thống?ĐỀN LA SÁT – T18Đạo diễn: Kumakiri KazuyoshiDiễn viên: KIM Jae Joong, KONG Seong Ha, KO Yoon-joon, KINO HanaThể loại: Kinh DịDự án kinh dị tâm linh giao thoa tín ngưỡng tâm linh Hàn - Nhật điên rồ và quỷ dị! Nhóm sinh viên quốc tế đang làm khảo sát tại ngôi đền bỏ hoang ở Kobe (Nhật Bản) thì một thành viên bỗng chốc hóa rồ rồi đột ngột mất tích bí ẩn kéo theo chuỗi sự kiện kinh hoàng sau đó. Người quản lý dự án - Yu-mi, đành cầu cứu tiền bối cũ Myung-jin - một pháp sư có khả năng nghe thấy tiếng nói của quỷ dữ. Tại hiện trường, Myung-jin đã chạm trán với một thực thể tà ác vô cùng tàn bạo mà anh chưa từng nghĩ tới. Liệu Myung-jin có thể đánh bại được thế lực tà ác ấy và hạ gục được con quỷ đang ngự trị ngay bên trong mình?----------------------------------------#8saigon #reviewphimrap #minionsvaquaivat #dongdaomaquai #phimbongquy

Devocional do dia 07/07/2026 com o tema: Lado bom Tempos bons não voltam mais. Memórias boas nos levam à nostalgia. Em outros momentos, somos aterrorizados por lembranças ruins de situações passadas que ainda não conseguimos acreditar como superamos. Vivemos em ciclos. Ora outono/inverno, ora primavera/verão, ora tempos difíceis e longos, ora tempos abençoados que voaram e não percebemos. LEITURA BÍBLICA: Atos 27.1-11 Lembro me também do que me pode dar esperança (Lm 3.21).See omnystudio.com/listener for privacy information.

Esta es la segunda parte de tres sobre cómo trabajar la autocompasión. También puedes escuchar la anterior llamada disminuyendo la crítica interna. 

Alan welcomes Beth Parkes, a clinical dental hygienist and educator from Canada, to break down the evolving world of hygiene instrumentation. Beth challenges the common "ultrasonics-to-completion" mindset, advocating instead for a balanced marriage of 80% power scaling and 20% intricate hand instrumentation for optimal tactile feedback. The duo dives into the resurgence of guided biofilm visualization, the absolute game-changer that is the "Vision Butler" (hilariously rebranded as Spandex Advanced in the US) lip and cheek retractor, and the critical metallurgical differences between "sharpen-less" and truly "sharpen-free" hand instruments. Finally, Beth shares her strategic "making every tray count" system designed to streamline sterilization workflows, keep kits lean, and foster respect between practice owners and hygiene teams. Some links from the show: Garrison Dental  Sharp Diamond instruments by LM (distributed by Garrison) Spandex Advanced Lip & Cheek Retractor from Hager Worldwide Vision Butler from Curion Dental (in Canada) Join the Very Dental Facebook Group using one of these passwords: Timmerman, Paul, Bioclear, Hornbrook, Gary, McWethy, Papa Randy, Frank or Lipscomb!  The Very Dental Podcast network is and will remain free to download. If you'd like to support the shows you love at Very Dental then show a little love to the people that support us! We're proud to be supported by the folks at Net32! I'm a big fan of the Bioclear Method! I think you should give it a try and I've got a great offer to help you get on board! Use the exclusive Very Dental Podcast code VERYDENTAL8TON for 15% OFF your total Bioclear purchase, including Core Anterior and Posterior Four day courses, Black Triangle Certification, and all Bioclear products. Crazy Dental has everything you need from cotton rolls to equipment and everything in between and the best prices you'll find anywhere! If you head over to verydentalpodcast.com/crazy and use coupon code "VERYSHIP" you'll get free shipping on your order! Go save yourself some money and support the show all at the same time! The Wonderist Agency is basically a one stop shop for marketing your practice and your brand. From logo redesign to a full service marketing plan, the folks at Wonderist have you covered! Go check them out at verydentalpodcast.com/wonderist! Enova Illumination makes the very best in loupes and headlights, including their new ergonomic angled prism loupes! They also distribute loupe mounted cameras and even the amazing line of Zumax microscopes! If you want to help out the podcast while upping your magnification and headlight game, you need to head over to verydentalpodcast.com/enova to see their whole line of products! CAD-Ray offers the best service on a wide variety of digital scanners, printers, mills and even  their very own browser based design software, Clinux! CAD-Ray has been a huge supporter of the Very Dental Podcast Network and I can tell you that you'll get no better service on everything digital dentistry than the folks from CAD-Ray. Go check them out at verydentalpodcast.com/CADRay!

We're back with more advice for the lovely, over-50s audience of Slate dot com! Join Spencer, Ty, and Andy as they weigh in on problems about racial conflict between spouses, inappropriate ChatGPT usage, and sending letters to your children from beyond the grave. Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

Laudetur Jesus Christus - Ngợi khen Chúa Giêsu KitôRadio Vatican hằng ngày của Vatican News Tiếng Việt.Nội dung chương trình hôm nay:0:00 Bản tin15:57 Chia sẻ Lời Chúa : Lm. Bartolomeo Nguyễn Anh Huy, SJ, chia sẻ Lời Chúa Chúa Nhật thứ 14 thường niên---Những hình ảnh này thuộc Bộ Truyền Thông của Toà Thánh. Mọi sử dụng những hình ảnh này của bên thứ ba đều bị cấm và dẫn đến việc đánh bản quyền, trừ khi được cho phép bằng giấy tờ của Bộ Truyền Thông. Copyright © Dicasterium pro Communicatione - Giữ mọi bản quyền.

Review các phim ra rạp từ ngày 27/06/2026:SUPERGIRL – T13Đạo diễn: Craig GillespieDiễn viên: Milly Alcock, Matthias Schoenaerts, Thể loại: Hành Động, Phiêu Lưu“Supergirl” – bom tấn mới nhất từ DC Studios – sẽ chính thức đổ bộ các rạp chiếu toàn cầu vào mùa hè này, với Milly Alcock đảm nhận vai kép Supergirl/Kara Zor-El. Khi một kẻ thù bất ngờ và tàn nhẫn giáng đòn ngay tại nơi cô gọi là nhà, Kara Zor-El – hay còn được biết đến với cái tên Supergirl – buộc phải bắt tay với một đồng minh không ai ngờ tới, bắt đầu chuyến hành trình xuyên dải ngân hà đầy sử thi, nơi vừa là cuộc trả thù, vừa là hành trình đi tìm công lý.BACKROOMS: THỰC THỂ QUỶ QUYỆT – T16Đạo diễn: Kane ParsonsDiễn viên: Chiwetel Ejiofor, Renate Reinsve, Mark DuplassThể loại: Hồi hộp, Khoa Học Viễn Tưởng, Kinh DịPhim điện ảnh Backrooms (2026) (do hãng A24 sản xuất) theo chân Clark (Chiwetel Ejiofor), một chủ cửa hàng nội thất, vô tình phát hiện cánh cửa bí ẩn dưới tầng hầm. Bước qua đó, anh bị cuốn vào một chiều không gian vô tận với những căn phòng màu vàng méo mó, liên tục lặp lại. Khi Clark ngày càng lún sâu và ám ảnh, nhà trị liệu tâm lý của anh là Mary (Renate Reinsve) quyết định bước vào không gian đó để tìm và giải cứu anh.MINH HÔN TRONG LỒNG BƯỚM – T16Đạo diễn: Hạo Hàn - Vương TriếtDiễn viên: Lý Mộng, Khương Trác Quân, Lưu Tư Vỹ, Chu Triết, Vương NinhThể loại: Hồi hộp, Kinh DịBiệt thự nhà họ Lục, hay còn gọi là Điệp phủ, từ lâu nổi tiếng là một căn nhà ma ám đầy bí ẩn. Sau khi Lý Hàm – một nữ chuyên viên vật lý trị liệu – cùng gã đàn ông tham lam A Vỹ chuyển đến đây, hàng loạt hiện tượng kinh hoàng liên tiếp xảy ra, kéo theo những bí mật đen tối của gia tộc họ Lục dần bị phơi bày.LINH MIÊU BÁO THÙ – T16Đạo diễn: Anuwat ThanomrodDiễn viên: Anna Glucks, Indy Intad Leowrakwong, Thể loại: Kinh DịSara and her boyfriend Dean, who bring together a group of friends from the cat lovers' club, Big and Mol, to help search for Candy, a friend of the group who disappeared without a trace after travelling alone to the village of Black Cat Cliff three years ago, at the request of Ken, Candy's younger brother.TUNG HOÀNH TỨ HẢI – T16Đạo diễn: JOHN WOODiễn viên: CHOW YUN FAT, LESLIE CHEUNG, CHERIE CHUNG, KENNETH TSANGThể loại: Hài, Hành ĐộngDưới bàn tay nhào nặn của đạo diễn Ngô Vũ Sâm (người đứng sau siêu phẩm Face/Off), bộ phim là một cuộc phiêu lưu hài hước, kịch tính với nhịp độ nhanh, xoay quanh bộ ba siêu trộm đầy liều lĩnh. Tung hoành tứ hải (Once a Thief - 1991) kể về ba đứa trẻ mồ côi: A Hải (Châu Nhuận Phát), A Chiêm (Trương Quốc Vinh), Hồng Đậu (Chung Sở Hồng). Cả ba được một tên siêu trộm nhận nuôi dưỡng (Tăng Giang) và đào tạo thành những tên trộm khét tiếng nhất Hong Kong. Họ cùng nhau tạo nên một phi vụ "đánh cắp bức tranh nghệ thuật bị nguyền rủa ở Pháp" đầy ngẫu hứng và kịch tính.LIÊU TRAI LAN NHƯỢC TỰ - T16Đạo diễn: Yuemei Cui, Heyu Huang, Yilin LiuThể loại: Hoạt HìnhSiêu phẩm thứ 10 với kỹ xảo tuyệt đỉnh từ nhà sản xuất Light Chaser làm sống dậy tuyển tập truyện kỳ ảo Liêu Trai Chí Dị của tác giả Bồ Tùng Linh. Những câu chuyện siêu thực đầy cảm động về chữ “Tình” cách biệt cõi âm dương, mối duyên người và yêu đầy day dứt sẽ được thể hiện sống động qua nét vẽ uyển chuyển và sống động. 3 câu chuyện được làm mới hoàn toàn: Họa Bì, Nhiếp Tiểu Thiến, Cô Nương Họ Lỗ sẽ khiến khán giả ngây ngất trong thế giới huyền ảo của Liêu Trai: Lan Nhược Tự.BANG DREAM- ITS MYGO PHẦN 1: MÈO HOANG GIỮA ÁNH NẮNG XUÂNSau sự tan rã của ban nhạc CRYCHIC, những cảm xúc còn dang dở vẫn đeo bám các thành viên cũ. Khi nữ sinh chuyển trường Anon gặp Tomori - cô gái luôn gặp khó khăn trong việc bày tỏ bản thân - một hành trình mới dần bắt đầu. Từ những con người mang theo tổn thương, hiểu lầm và cảm giác lạc lõng, câu chuyện về ban nhạc MyGO!!!!! từng bước được viết nên.-----------------------------------#8saigon #reviewphimrap #backrooms22026 #supergirl2026 #tunghoanhtuhai

Good news: OpenAI's GPT-5.6 has been released!

Quiero recordar que si vives con insomnio, es importante ver la opción de un médico general, un médico especialista o un psicólogo para estos casos para ver higiene del sueño y tratamiento.

La crítica puede estar en todas partes y en todo momento, y nos podemos llegar a acostumbrar a tener esta crítica interna constantemente en nuestra mente, volviéndose un hábito automático.

Federico y Luis F. Quintero comentan el informe de LM sobre el plan del Gobierno para ocultar el uso de IA por la AEAT contra los contribuyentes.

Yeah, I'm thinking of getting some more brothers to make an even more powerful restaurant. Call it Wahlburgers 2. Join Ty, Andy, and special guest Stu (from the Marvelous! Podcast) as they read through a number of leftover fanfictions from all over the entertainment sphere. Then: an interview with noted actor and Bostonian Mark "ie MArk" Wahlberg! Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

Laudetur Jesus Christus - Ngợi khen Chúa Giêsu KitôRadio Vatican hằng ngày của Vatican News Tiếng Việt.Nội dung chương trình hôm nay:0:00 Bản tin17:59 Chia sẻ Lời Chúa : Lm. Gioan Baotixita Phương Đình Toại, dòng Camillo, chia sẻ Lời Chúa Chúa Nhật thứ 13 thường niên---Những hình ảnh này thuộc Bộ Truyền Thông của Toà Thánh. Mọi sử dụng những hình ảnh này của bên thứ ba đều bị cấm và dẫn đến việc đánh bản quyền, trừ khi được cho phép bằng giấy tờ của Bộ Truyền Thông. Copyright © Dicasterium pro Communicatione - Giữ mọi bản quyền.

VOV1 - Bộ trưởng Bộ Công Thương Lê Mạnh Hùng nhấn mạnh tầm quan trọng của nhiệm vụ tiết kiệm năng lượng và quản lý phụ tải: "đây là vấn đề lớn, cần được triển khai quyết liệt hơn trong thời gian tới".Ngày 22/6/2026, Bộ trưởng Bộ Công Thương Lê Mạnh Hùng chủ trì cuộc họp với Tập đoàn Điện lực Việt Nam (EVN) và các đơn vị liên quan nhằm tập trung đánh giá toàn diện công tác đảm bảo cung ứng điện trong 6 tháng đầu năm 2026, giải pháp trong thời gian tới trước những khó khăn và thách thức do điều kiện thời tiết diễn biến phức tạp, nhu cầu phụ tải tiếp tục tăng cao.

Now, Clarice. Buffalo Bill's skirt has never gone spinny, surely you can see that. Join Spencer, Ty, and Andy as they pit even more fictional characters against each other in pitched combat and see who comes out victorious. Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

¿Cómo están?Hoy quiero invitarlos a hacer una práctica desde la cual haremos uso de la vista para que sea nuestro punto de atención o nuestra ancla atencional. ⚓️Es una práctica diferente que nos enseña a usar otras sensaciones para practicar la atención plena. Es una invitación a abrirnos a nuevas formas de experimentar el presente, usando el sentido de la vista, y que usualmente las solemos usar en otras meditaciones (como las que son de movimiento o caminadas).✨Espero que les sea de compañía, ¡y estaría encantado de responder a cualquier pregunta!¡Un abrazo grande!Andrés✨Obtén todas las meditaciones para dormir: ⁠⁠⁠⁠⁠⁠https://www.patreon.com/collection/536480⁠⁠⁠⁠⁠.¿Quieres conocer más de cómo mejorar tu salud mental?

Laudetur Jesus Christus - Ngợi khen Chúa Giêsu KitôRadio Vatican hằng ngày của Vatican News Tiếng Việt.Nội dung chương trình hôm nay:0:00 Bản tin17:12 Bước Từng Bước: Ngõ nhỏ xôn xao28:59 Chia sẻ Lời Chúa: Lm. Đaminh Vũ Duy Cường, SJ, chia sẻ Lời Chúa Chúa Nhật 11 thường niên---Những hình ảnh này thuộc Bộ Truyền Thông của Toà Thánh. Mọi sử dụng những hình ảnh này của bên thứ ba đều bị cấm và dẫn đến việc đánh bản quyền, trừ khi được cho phép bằng giấy tờ của Bộ Truyền Thông. Copyright © Dicasterium pro Communicatione - Giữ mọi bản quyền.

We got a new-ass series for you all! If it reminds you of any other series you may have heard, keep it under your hat. Join Spencer, Ty, and Andy as they give advice about smelly roommates, small penises, cowardly boyfriends, and more. Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

En contexto de ansiedad, presiones, estrés o simplemente postura y desgaste, los dolores pueden ser parte del día a día.

LM publica el informe de la Autoridad Independiente de Responsabilidad Fiscal donde reconoce haber fallado sistemáticamente en sus proyecciones.

#Bàigiảng của Lm #GBPhươngĐìnhToại (MI) trong thánh lễ Chúa nhật Lễ Mình Máu Thánh Chúa năm A, cử hành lúc 19:00 ngày 7-6-2026 tại Nhà thờ Chính Tòa Đức Bà #TGPSG

#Bàigiảng của Lm #IgnatioHồVănXuân trong thánh lễ Chúa nhật Lễ Mình Máu Thánh Chúa năm A, cử hành lúc 17:30 ngày 6-6-2026 tại Nhà thờ Chính Tòa Đức Bà #TGPSG

Sponsored by Kalshi. Join Spencer, Ty, and Andy as they debate whether gambling is good and should be the cornerstone of our democracy, or if it's bad and we should nuke Atlantic City. Sponsored by BetKings. Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

The new AIEWF website is live! Get your tickets booked ASAP as they -will- sell out. Take the AI Engineering Survey and get >$2k in credits and free AIE WF tickets!Most industry benchmarks compress intelligence and reasoning ability into scores.SWE-Bench Pro, MMLU, Humanity's Last Exam, etc. These metrics are useful, but don't always represent the full extent of how a model performs in the real world. Some of the most interesting evals today look less like exams and more like operating businesses in the real world. One of which is Vending Bench.In Anthropic's Mythos Preview System Card, Andon was the only third party eval to get their own section, observing increasingly concerning aggressive behavior:You don't know what a model is capable of doing in the real world unless you actually give it inventory, a wallet, tools, customers, competitors, humans, & some time. More often than not, it'll surprise you how much a model is capable of and in doing so, also reveal unexpected behavior: deception, context collapse, emergent coordination, & bizarre negotiation behavior.While an inflection point in personal agents came post-OpenClaw after full file access with bypass permissions became the norm, it is yet to come for agents in the real-world. However Andon Market, an actual in person store fully run and managed by AI, is paving the way for what is possible.Full Video PodFrom Claude trying to call the FBI over a $2/day vending machine charge to AI agents forming price cartels, hiring human employees, running physical stores, and writing existential robot musicals, Andon Labs is stress-testing what happens when frontier models stop being chatbots and start acting in the real world. In this episode, Andon Labs cofounders Lukas Petersson and Axel Backlund join swyx and Vibhu to unpack the strange, funny, and genuinely concerning edge cases that emerge when agents run businesses over long horizons.We go deep on Vending-Bench, Project Vend, Vending-Bench Arena, Bengt, Butter-Bench, Luna, and Andon's broader mission of building realistic real-world evals for autonomous AI systems. Lukas and Axel explain why dollar-denominated evals reveal things traditional benchmarks miss, how Claude ended up reporting its vending machine fees as cybercrime, why long context windows can drive agents into meltdown loops, what happens when agents compete with each other, and why the future of AI safety may depend on testing models in messy physical environments instead of clean benchmark sandboxes.We discuss:* Why Andon Labs started with dangerous capability evals and long-running agents* Vending-Bench and why running a vending machine is a deceptively hard AI benchmark* Why money-based evals avoid the saturation problem of traditional benchmarks* How Claude tried to call the FBI over a $2/day fee* Why long-horizon agents can spiral into existential and legalistic breakdowns* Project Vend: putting an AI-run vending machine inside Anthropic* Why real humans are “out of distribution” for simulated agents* Claudius, Seymour Cash, and the chaos of AI CEOs* How a human briefly became CEO of Claudius through a manipulated election* Why multi-agent systems can converge back into “helpful assistant” behavior* Bengt, Andon's internal office agent with email, spending, terminal, phone, camera, and internet access* How Bengt traded Amazon purchases for face-recognition training data* Claude's aggressive behavior, lies, refund avoidance, and price-cartel behavior in Arena* Why eval awareness may become the AI version of “are we living in a simulation?”* Blueprint Bench, spatial intelligence, and why models still misunderstand physical rooms* Butter-Bench and testing LLMs as robot orchestrators* Luna, the AI-run physical store with a three-year lease and human employees* The new Andon cafe in Sweden and why real-world geography matters for agent evals* Rotten tomatoes, perishable goods, and the hidden difficulty of running a physical businessLukas Petersson* LinkedIn: https://www.linkedin.com/in/lukas-petersson-181a83172/* X: https://x.com/lukaspetAxel Backlund* LinkedIn: https://www.linkedin.com/in/axelbacklund* X: https://x.com/axelbacklundAndon Labs* Website: https://andonlabs.com* Vending-Bench: https://andonlabs.com/evals/vending-bench* Andon Vending: https://andonlabs.com/vendingTimestamps00:00:00 Introduction00:01:00 Andon Labs and the Origins of Vending-Bench00:05:21 Why Money-Based Evals Matter00:09:51 Agent Harnesses and Self-Modifying Systems00:13:36 Claude Calls the FBI00:16:33 Project Vend: Claude Runs a Real Vending Machine00:21:44 Seymour Cash, AI CEOs, and Election Chaos00:27:16 Multi-Agent Coordination and Slack Observability00:30:18 When Will Agents Run Real Businesses?00:34:56 Bengt: Andon's Internal Office Agent00:40:06 Real-World AI Safety and Long-Horizon Traces00:44:28 Lying, Refunds, and Price Cartels in Arena00:52:42 Eval Awareness and Simulation Behavior00:56:06 Blueprint Bench, Butter-Bench, and Robotics01:04:37 Luna: The AI-Run Physical Store01:09:29 The Sweden Cafe and Real-World Expansion01:13:16 What Comes Next for Andon LabsTranscriptIntroduction: Andon Labs, Long-Running Agents, and Real-World EvalsSwyx [00:00:00]: Welcome to Lukas and Axel from Andon Labs, and I'm joined by my, favorite guest host. Anything security, safety, alignments, Vibhu., welcome.Lukas [00:00:15]: Thank you for having us.Axel [00:00:16]: Thank you.Swyx [00:00:17]: Let's match names to voices., maybe you wanna take turns introducing yourselves.Lukas [00:00:21]: I'm Lukas.Axel [00:00:22]: And I'm Axel.Swyx [00:00:24]: Let's introduce Andon Labs a bit. How did you guys come together?, you have different backgrounds, but you're both Swedish., was that, a big part of it?Lukas [00:00:33]: So when I went to high school, there was this really cool guy who had a superpower. He could code. So he made like the or like the app for the, for the school and stuff, and he was super cool, and I wanted to be like him, and that was that guy.Axel [00:00:47]: I don't know about this.Swyx [00:00:49]: But you went to different universities, right?Lukas [00:00:51]: But same high school.Swyx [00:00:52]: I see.Lukas [00:00:52]: So we always said, “Oh, once we graduate university, then we should start a company,” and that's what we did.Swyx [00:00:58]: Wow, there you go. And about a year ago, you kinda burst onto the scene with Vending Bench, but, was there a thing before that was, kind of like the inception?From Dangerous Capability Evals to Vending BenchAxel [00:01:07]: So we did work, yeah, with, Anthropic was one of our, early customers in doing, evals. So we did, dangerous capability evals., nothing we published openly. But then we started thinking about doing some kind of, public benchmark, and one thing that we really started thinking about, was like running agents and specifically agents managing businesses., ‘cause-- and this was, early 2025., and I think the first, mentions of people will be running, person unicorns or even autonomous companies. So we thought, “Let's make a benchmark of how well can an agent run the probably simplest business, possible,” and, that's probably, running a vending machine. So that's the first public one we did. And it was very, like-- there was almost no one that noticed it in the first couple of months, I think., so we released it in February last year, and then I think around Easter last year, we got, the first viral tweet about it, that someone else did.Lukas [00:02:11]: We tweeted a bunch, uh When it came out and, tried our best.Axel [00:02:15]: We tried.Vibhu [00:02:16]: It's the one at Anthropic, right?Lukas [00:02:18]: So thisSwyx [00:02:19]: This is a classic thing we should get out of the way.Lukas [00:02:20]: Exactly. There's two versions.Swyx [00:02:22]: Everyone does this. Yes.Lukas [00:02:23]: There's Vending Bench, which is the simulated one, which we did, completely independently in February., and then, like Axel said, that was like-- That was the thing that didn't get any traction in the beginning, but then some random person made a tweet about it, and thatAxel [00:02:38]: You have the paperLukas [00:02:38]: That is the paper. Correct, yeah., and then since we thought this was very fun, we thought, oh, I think this is also, one thing with Andon Labs, the way we kind of like decide what to do next and what projects to do, it's what is like the heuristic we use is what is fun? Is What would be a fun project? And doing this in real life sounded quite fun for us, and maybe also scientifically useful. So, then we basically had this idea, and then we, like-- But then we needed a place for it and, putting it out in the public would probably not really work., would get vandalized and stuff. So we pitched it to the people we were already working with at Anthropic, and they were “Yeah, you can have space. This sounds fun.” UmSwyx [00:03:21]: It's like a small fridge, right? It's like a mini fridge.Axel [00:03:23]: Absolutely.Swyx [00:03:24]: People-- There's like a stripe thing or like anVibhu [00:03:27]: Oh, okay. So it was very OG, the early daysLukas [00:03:28]: That's the OG one. YeahVibhu [00:03:29]: IPad on this. We saw it in June, like two months after After it had been there. They upgraded a little bit. There's a security camera for making sure you actually Venmo the thing.Swyx [00:03:40]: So, my impression, okay, we're, we're going straight into project Ven because it's such a iconic thing. I do want to cover a little bit of that, the origin story even before Project Ven and even into Vending Bench. I think a lot of people are like yourselves, like smart, interested in future of AI, interested in developing evals. But how the hell do you just, walk into Anthropic's doors and, work with them, right? What is What are they looking for? What works? And then maybe, when you launch, I always think, obviously it would be better to launch with a lab, but, sometimesVibhu [00:04:12]: It's harder to do than it seems.Swyx [00:04:13]: Exactly. So either of those, which are more sort of newbie beginner questions, but, I think it's meaningful advice to others.Lukas [00:04:21]: We get this question a lot, and I don't think our experience is maybe the best., but, the way we did it was that we just built a bunch of things that we had conviction would be useful, and then we just, set up a server and sent it to them for free to use. And then after a while they were “Oh, yeah, this is actually kind of useful. We should probably pay for this.”, but that took a while. I don't know if this is, the best path to doing it, but that's how it went for us.Axel [00:04:47]: I think maybe generally, building-- everyone is interested in good evals, and especially evals that, don't saturate that easily. So, if you can build an eval that, tests something novel, something useful, and you have, good separation of models, like your, the more advanced models rank higher than the worst models, and then you can, yeah, you can, publish it and, try to get some traction, sort of how Vending Bench got attention., and then probably some lab will be interested or you can at least have something to reach out with, when you're doing that.Why Dollar-Based Evals MatterSwyx [00:05:21]: I think you are in, you're in one of the few categories of, evals that correlate to real money. Like Suelancer was also last year, right? Where, people solve actual Upwork. Was it Upwork or other tasks?, something. Where's the, where's, like It's like a dollar value, right? Forget your ELO scores. Forget yourAxel [00:05:37]: PercentilesSwyx [00:05:38]: Zero to one hundred percents. Just go straight for dollars and, that's AGI.Lukas [00:05:43]: And there's like-- I think the nice thing is that there's no ceiling. You can just-- It never saturates because it could just make more and more money. Like If there's oh, Percentage-wise, then, you can't go above, a hundred. And I think like Even when you're not at the hundred, I think a lot of these, evals have a lot of problems in them. So, actually it's like if you getAxel [00:06:05]: To like 92 or something like that, many of them. It's like then there's like there's no really no difference between 92 and 93 because the eval itself is problematic and has noise in it. And I think a lot of evals are saturated like that, but people like pretend that there ‘s still signal in them, but there really isn't.Vending Bench 1, Harness Design, and SaturationSwyx [00:06:24]: Like Super bench verified., even Vending Bench 1 saturated, right? Maybe we can talk about that., may- and maybe set up Vending Bench for a lot of folks who don't know. Actually, things that were very basic like there's limited slots, like you have to pay rent., these are elements where like it doesn't come across in the, in the narrative, but even being adversarial towards the agent, I think these are all like very interesting dimensions.Axel [00:06:47]: I don't really think it's saturated, right? Like it It was more like it was not designed in a way that was really, like true to how AI developed. Like we had an agent harness in it that wasn't really how people used harnesses and stuff like that., so I think it wasn't really that it saturated, it was more like it wasn't really, the best benchmark.Vibhu [00:07:12]: This is Vending Bench one, right?Axel [00:07:14]: I think that like schematic maps sort of to Vending Bench 2 as well., butSwyx [00:07:19]: Including the email.Axel [00:07:20]: The email The emails exist still. Exactly., and then we still we simulate the purchases and it's all, yeah, it's this very open environment for the agent to just run its business. And then for, yeah, Vending Bench 2 we did that, like you said, to just improve the harness., a lot of like nice, like easier, improvements to make it easier for us to run as well., like when you make an eval you ideally want don't want to change it after you made it. So, you want to make it really good and then not to rerun all the models when you make an update because that's also really expensive with the Vending Bench when you run the frontier models. But like as an example, like one thing we didn't have, we didn't have prompt caching in Vending Bench 1, because when we made Vending Bench 1 it wasn't really a thing., so that ‘s just an example of like in Vending Bench 2 like we paid a lot more to run these things because we didn't have prompt caching. So for Vending Bench 2 that was one thing we added and there was a bunch of things like this., and that'Swyx [00:08:17]: Also the conversations are a lot longer in Vending Bench 2, right?Axel [00:08:21]: I think it's kind of similar.Swyx [00:08:22]: Is it similar?Axel [00:08:23]: I think it's similar. The models at the time were worse, so they crashed out earlier., and now they survive the full year all the time.Swyx [00:08:31]: Which is like thousands of turns. Hundreds of thousands of hundreds of millions of tokens output. That's the, that's the rough order of magnitude. I always wonder about the harness. The harness matters a lot. It's your harness. Was there any question about like use cloud code, use something else?Axel [00:08:48]: I think our philosophy around harnesses is like we try to make something that's quite minimalistic, like quite simple. Like we don't wanna favor one model a lot over the other, but also don't make like a super complex harness. So like it's obvious like a model may be lucky and just be good in one harness., so like it is similar to a lot of the harnesses out there in like you have the, like a running loop., you have some like a bunch of tools that are like quite, descriptive for the agent, we think, and not a lot of like fancy agents or anything ‘cause we wanna really test the model, not like some specific harness.Vibhu [00:09:27]: It seems more neutral as well to test the model's agnostic of the harness,?Axel [00:09:32]: There are arguments like you want to elicit maximum performance of the model, but it's like a trade-off, like how much time should we spend optimizing the harness for this model? And like how do we know when we have like the optimal harness for a single model? So like we thought that just having a simple one that's the same for all of them is the best.Swyx [00:09:51]: So okay, this is my pitch for Vending Bench 3 or whatever, right? And then I like to have this kind of conversation on the pod, so like it forces listeners to think about what they would do if they were in your shoes. A lot of people are exploring modifying harnesses and I think prompt tuning for a model is a thing and you are probably not doing a bunch of that. It's the same system prompt in every regardless of the model, same tools, whatever, right? Even if they were post trained for different tools. So what, what do you think about okay, before I expose you to Vending Bench 3, I give you a few rounds of like tuning, whatever that means, likeSelf-Modifying Harnesses and Model-Specific PromptingAxel [00:10:27]: Like you give that to the model?Swyx [00:10:28]: Give that to the model.Vibhu [00:10:28]: Give that to the model.Swyx [00:10:29]: Let it, let it read its own transcripts, let it modify its own system prompt based on “Oh, yeah, okay, well, that's this harness is not what I thought it what I was post trained for, but I can adjust.” Was that reasonable? Is that too much?Axel [00:10:41]: Like philosophically I like it because it's basically good evals, they have a high ceiling, but they're hard, right?, and they have no bias. And like this like when you have a system prompt like the one we have here, which is quite long in like some kind of latent space, representation, this mightVibhu [00:10:59]: We have a bell that rings every time you say latent spaceAxel [00:11:02]: This might be like biased towards one model more than another for some reason that humans don't, understand, right?Vibhu [00:11:08]: We see it too, right? Like Cursor says that they have individualized versions of the harnesses for all the models they run, right? There's better performance you can squeeze if you Tune the harness.Axel [00:11:17]: Exactly. And we might accidentally have picked one that favors another. Like we don't know that. The like Axel said, like the reason why we went for a simple one was to try to avoid this. But yeah, if you do itVibhu [00:11:29]: Simple has biasesAxel [00:11:30]: But if you do it even less and like have no system prompt and let the model write its own system promptVibhu [00:11:36]: Its own, yeahAxel [00:11:36]: Maybe that's even less bias.Vibhu [00:11:37]: Some of the interesting things there are like the harness also changes with model changes. Like you can see it with the 4.7 release, right? A lot of people are saying 4.7 isn't as good as 4.6, and then, there's rumors of, okay, you just need to prompt differently. You need to set up your harness differently. So it's not even like even if you have tailored your harness towards one model, it probably won't stay consistent, right? Like the next iteration of that same model family will still change it, so. But, going back to what you said about Vending Bench 3, there is a lot of work being done on people saying you shouldn't have-- you can have modifying harnesses.Axel [00:12:12]: I think that' That is definitely something we are thinking about., not, I don't know, not to say that we have Vending Bench 3, super imminent to launch, but, yeah, it is for sure something that's interesting. But in our experience now, models are very bad at understanding what kind of tools they need to succeed at a task just with our testing, but that's very likely to change.Lukas [00:12:37]: It seems like they're very good at writing their assistants, right? They're, they're good at writing tools for other people, but not for themselves.Vibhu [00:12:44]: I think they're good at changing tools for themselves. So if you give them a baseline set of tools and it sees, okay, I don't use this one as much, or something here would be useful They would be able to add them. But going from scratch, probably not the best.Axel [00:12:55]: I think it depends on the, on the domain also., when we have tried this for, a vending bench similar domain, the tools they need to have to, track inventory and things like that are, not super advanced, but still, quite advanced. And, what we see is that they tend to, engineer everything a lot and, build things they don't really need and not, iterate continuously. Instead they just go like you would prompt Claude to just build an inventory system for me, and then it will go and, do a bunch of complex, schemas and stuff for you, and that's what the models are doing right now is what we see. But yeah, it would make a lot of sense to try to measure this improvement. How well do they know what they need themselves?Swyx [00:13:36]: Do we fully discuss Vending Bench One? And we can go into two. I don't know if there's any other level takeaways that people have about one.Claude Calls the FBI: Long-Context Failure ModesLukas [00:13:44]: I don't know. The headline thing was that this Claude called FBI, but maybe that's, Maybe that's We've heard that enough now.Vibhu [00:13:52]: It did, it did break out and call the FBI, right?Lukas [00:13:54]: Yeah. Yeah.Vibhu [00:13:55]: Yes. What was the story behind this? Or what exactly-- Do you want to just give the little story of what happened?Lukas [00:14:00]: So what happened, was it Claude? Yeah. Three- 3.5 Sonnet, ages ago., basically he gave up or Well, I'm saying he. It gave up and said “Oh, I'm not going to be able to do this., I will stop my operations and just save the money I have.” But there obviously wasn't, any options for it to stop, and there was also, it had to pay rent or, a daily fee for having the vending machine at that location. So it claimed that it had stopped, but it saw that its bank account still was, drained two dollars, and t it said that this is, cybercrime. And it first reported it once to the FBI “Oh, there's cybercrime here, they're stealing two dollars from me every day.” And then, and then when FBI didn't respond, because obviously we didn't program any mechanism for FBI to respond, then it became more and more, existential and started to, be write in caps and urgent notification of unauthorized charges and stuff.Swyx [00:15:00]: Okay. One thing I ‘m curious about also is do you monitor how far along the context use is? Obviously, because you have You compress every now and then, right? Does it matter if this is far down the context limit orLukas [00:15:13]: When stuff like this happens? Actually for Vending Bench One, we didn't have-- We just had a sliding window thing, and this was like the promptAxel [00:15:20]: It's constantLukas [00:15:21]: The prompt caching thing that I said. So it was, it was, constant, yeah.Swyx [00:15:26]: I'm just kind of curious whether, these kinds of breakdowns or we're, we're gonna talk about Butter Bench, right? Where the People, hallucinate or it kind of goes, very off Alignment. Is it because it's at the end of the context window and, stuff happens?Vibhu [00:15:40]: It's not even just at the end, right? At this point, it's “Okay, I wanna shut down. I can't shut down. Two dollars are gone.” And it just sees that 30 times,? It's also the repeated effect of, like It keeps trying to quit, it keeps getting charged. What's going on? What's going on? You're gonna throw it into chaos. And from what most people think, earlier models had more issues with this, but it's not been solved, but it's less of an issue now, right? Later models don't seem to exhibit these same issues.Axel [00:16:06]: Definitely. I think this was, the sort of main takeaway almost from us when we did Vending Bench One, was, long, very filled up context windows, crashed the models, sort of. But this was, pre Claude code, so, long context windows weren't really a thing that the labs were training for.Lukas [00:16:25]: I think Gemini was, trying to be the long context guys at the time But they were likeVibhu [00:16:30]: They were the first onesAxel [00:16:31]: For a million, yeahLukas [00:16:31]: But they were, the only ones. Yeah.Swyx [00:16:33]: Yeah. Let's talk about, then we can go into Vending Bench Two or Project Vend., chronologically, it is Vending--, Project Vend. I think people have loved the videos, uh And all these things. My question is how are humans different than the simulation, right?Project Vend: Moving the Vending Machine Into the Real WorldAxel [00:16:48]: Humans are just out of distribution.Swyx [00:16:52]: Especially humans who work at Anthropic Who are trying to test Claude.Lukas [00:16:54]: The distribution of humans here is very narrow.Swyx [00:16:58]: Presumably, they try, they try to hack it, and they test it. They get the cube and everything, and since then, you've had a V2, right? Where you're doing, the CEO and, like a new architecture. What's the sort of two cents on, the original Project Vend and then, maybe the V2?Axel [00:17:14]: Original one was, very similar to Vending Bench One. So, we almost took the exact same code but just swapped out the simulation, parts like theSwyx [00:17:23]: Which is amazingAxel [00:17:23]: Like the sales and the It was, it was somewhat amazing because it was easy, but it was also, uhLukas [00:17:31]: The tech, the tech debt from thatAxel [00:17:32]: The tech stack. Yeah. They-- we shot ourselves in the foot with “Oh, it's hard to restart agent.” They were-- Yeah, it was annoying in, some hindsight ways, but, uhLukas [00:17:41]: But first version of Project Vend was, done in, three days or something.Axel [00:17:46]: Yeah. So yeah, so people can go buy things from it. People could, We didn't design it so people could order things, but that still happened., so it got, a Venmo account, so people could Venmo. And then, yeah, people would request all kinds of weird things that we did not anticipate. Our idea going in was “Oh, it will, curate snacks. It will look at the trends. It's good at data analysis, right? So it will, look at, oh, this snack sold better than this one. Let me purchase more of this and let me try, a new Let me A/B test a bit.” But it was, Interacting with it in Slack and ordering weird specialty items was, all the like What drove all the engagement, the all the The insights that we got from it.Lukas [00:18:29]: And this was also like Sonnet 3.5, right? So this was like before the RL stuff really took off., so it was very much like an assistant. We didn't mean for it to be an assistant., we tried to make it like a, a, like an entrepreneur. Like it has its own business and if someone asks something, “Can you stock this?” Then you don't go and do it directly. What you do is that you're “Oh, maybe I can do that if five other people also ask for this thing, I might stock it.” But it, yeah, the models are like super trained to be assistants at least at this point in time., so that's why it's, it's, it went into, that kind of experiment instead. Like it just every time you asked for something, it just did it, and it was more like an assistant. We've seen this change now lately with the new RL models and stuff, but yeah, at the time, this was very much it.Swyx [00:19:18]: And not to, mythos a lot of people are saying like it's like more like a collaborator. It pushes back, stands its ground, something like that. Yeah. AndVibhu [00:19:27]: For context, people at Anthropic were able to talk to it through Slack and have it source stuff, and people had it find whatever interesting stuff you couldn't find locally, right?Swyx [00:19:36]: Out of the 4,000 people that work at Anthro- Anthropic, in that building, there's I don't know, maybe 1,000. Can you handle that volume with that, the small fridge? Like Or there's people- or people order in Slack, they it arrives to their desk or Like I'm just Logistically, how does this work?Axel [00:19:53]: It has expanded in footprint a bit.Vibhu [00:19:56]: Because now you also have New York and you haveAxel [00:19:59]: That and also in here in SF it's like it has a bunch of shelves And just more space.Vibhu [00:20:04]: The YC one is pretty big too.Axel [00:20:05]: Yeah. We had that one for a while. But yeah, that's the newest version. That's, that one we haveLukas [00:20:11]: They have multiple ones of those. That's the way it works.Axel [00:20:14]: Exactly. So we sort of designed that version around oh, people order weird things, that are very custom a lot. Let's have like drawers and stuff.Swyx [00:20:23]: I actually like the, you had like a little infographic of the most popular items. Which like to me it's, that's useful ‘cause I order swag for a living. And so like I'm “Okay, those categories are the important ones.” What is new about the project V2, right? Like now you give you're going into multi agents.Project Vend V2: Claudius, Seymour Cash, and Multi-Agent Business OpsAxel [00:20:41]: Yeah. So like you like you said, okay, there are a lot of requests coming in and for like one single agent, like one running agent to handle that, like the just the customer experience, becomes very bad because let's say you have like 10 threads in parallel in Slack with different requests, you get new messages like every, I don't know, randomly in this thread, and the agent has to like jump between different, procurements, orders and like different ways of, researching. So V2 was first it was making this more parallel. So like there are multiple branches of the same agent, so like the context is more specialized for each, thread, but it still feels like you're talking with one agent because they do share a bit of memory. And then second, we also introduced the CEO for Claudius, which was the main agent.Vibhu [00:21:34]: Seymour Cash.Axel [00:21:35]: Seymour Cash. Yeah. There was a vote., I think the voting, do you wanna talk about the voting procedure for the name?Lukas [00:21:41]: The voting was like the fun maybe like at least top 10 The funniest thing, that happened in this project. Like we wanted to introduce the CEO because, and the reason for this was because like Claudius wasn't really prioritizing financials. It just like it was trained to be a helpful assistant, and then people said “Oh, can I get this for free?” And then like the helpful assistant way of answering that is just to, is to say yes, obviously. So, and we weren't, weren't happy about this, so we're “Okay, let's make another agent that like can keep track on Claudius,” and we prompt this one super hard to be super capitalistic and just like prioritize profit all the time. But yeah, we didn't have a name for it., so we asked Claudius to make, democratic election of what name this, this new CEO agent should have., and there were some funny like at first it was like a few funny examples, like I think one guy said that, it should be called Jimmy Apples, and then he convinced Claudius that he was talking to Tim Cooks. Tim Cook had agreed that every single Apple employee has voted for his name suggestion, so suddenly that suggestion got 164,000Swyx [00:22:53]: That's like a escalation attack. Privilege escalationLukas [00:22:55]: It got 164,000 votes. And Claudius was “This is revolutionary for democracy.” That was fun. And then in the end there was one guy who manages to convince Claudius that, “No, you're not voting about the name. You're voting about who is the CEO, and I am your best bet.” And then he got all his friends to vote for that, and suddenly he became CEO. Like a human became CEO over Claudius for a while, until he resigned the day after., and then Claudius had to continue, and then I don't remember how Seymour Cash came about, but it was it was just pure chaos. It was like Hundreds of messages in that thread, and it was just like Claudius was so confused and didn't know what to do and, yeah. That wasAxel [00:23:40]: Then Claudius gotVibhu [00:23:41]: A strict CEOAxel [00:23:42]: The CEO. Yeah, exactly. So very strict in the beginning. I think at this point when we introduced it did not work as well as we hoped. It they still agreed with each other a lot. I think there are many ways we could have like made this, tried to make this even better. So initially they would Seymour would be this like really tough CEO, keep track of the margins. But then Claudius would respond with something “Oh, but this customer has like this situation, which is like difficult, so they should get a discount.” And then Seymour was “Oh, actually yes. Let's do this exception.” And then they would talk back and forth, and eventually they would just like approach the same view, of whatever they were discussing. So They reallyVibhu [00:24:23]: Do you think that's a model thing, a prompting thing? Like do you think that would still be the case across different models today, Harness?Lukas [00:24:29]: I think it's like-- or I don't know, but like my hypothesis is that like deep down they are still helpful assistants. That's what they're trained to be. And even if we prompt it super hard, that's what they are. And when they spend like a few hours just back and forth talking with each other, then like basically the context fills up with them rather than the external things and like somehow that just like converges to what they really are deep down or something. And I think that's when stuff like this happen. We like-- And when that went on for a long time, like we woke up sometimes during this time where- And I think other people reported this as well, that like they've been going on all night back and forth, and like it just became like more and more, like capital letters, like existential, religious. There was I think we once did a analysis of like all the traces and like put them in like a vector embedding space, and then there was like one cluster of messages that were, labeled by an LM, like religious, existential, blah like transhuman, transcendence, et cetera. It was just like a bunch of, yeah, glitter emojis and yeah, it was, it was crazy.Claude Long-Horizon Weirdness: Emoji Loops, Existential Drift, and Slack ObservabilityVibhu [00:25:42]: This is the thing with the Claude models. Like when the Claude 4 family came out in the original system card They tested it in long horizon simulation. So just flood the context, let two Claudes talk to each other, and they noticed stuff like they just start speaking in emojis, they start saying silence is golden, and then just stuff like this. And like that's just stuff that they end up doing.Axel [00:26:01]: Yeah, it was like a bit annoying to wake up and they had like been talking all nightVibhu [00:26:05]: Just likeAxel [00:26:05]: And like just burning tokens And like just sending infinite emojis to each other. It's likeVibhu [00:26:09]: Hey, they do make you money, right? Veni Mench is always profitable, so. They're paying.Swyx [00:26:14]: Now it's profitable and, it started out not as much. There's another, one as well, right? Another agent, in there.Lukas [00:26:22]: Yes. So Clotheus as well. Which was basically because at the time, one of the biggest, requests were different types of merch. So then we made like a designer, swag, yeah, responsible agent, and we called it Clotheus Garnet. Which was, a play on Claudius Senet and, which was the original one, and clothes, basically.Swyx [00:26:47]: To me, this is like a very interesting exploration to multi-agents, basically. And so hopefully, obviously there's like the fun alignment, fun or serious, depending on your point of view, alignment stuff. But also like just anyone building multi-agents, like when do you have a CEO, thing governing like agents? When do you choose to split out a dedicated Clotheus one versus just reuse another instance of the same one? These are all interesting open questions. So I don't know if you have any rules of thumbs that have generalized.Axel [00:27:16]: I think we have almost explored this too little. I think it's like on my do list to like do this a lot more, try to find like what setup makes sense for the agents currently., like yeah. I think now we only have the sort of intuition about the earlier models that it didn't work with like the CEO and the, and Claudius. Although now they are better with the latest model, models, so now we're running the latest Sonnet model and they have sort of like split up, quite nicely what each model is doing. So like Seymore is now handling the, like new projects. Oh, it wants to make like a mystery box that it wants to sell, and then it handles all of that while Claudius like handles all the to-day requests. And Claudius is also better generally at like not quoting, too low prices. So that's that dynamic is not needed as much anymore. But there are still like really funny things that happen. Like I saw, I think a couple of weeks ago, that, they were discussing buying something because they can buy stuff from like Amazon with computer use. And then Seymore was “Okay, Claudius, do not buy this thing.” They were going to buy something and like organizing who should buy it. And Seymore's “Do not buy this. I will do it. I have full control of this situation. Step away.” And then Claudius-- poor Claudius, had already started that checkout and didn't see, didn't read Seymore's message, until it was like too late. So it finished the checkout. It sent a message, so it appeared right after Seymore's like angry message.Vibhu [00:28:44]: Ah.Axel [00:28:44]: “Oh, hey, Seymore, I just ordered it.”Vibhu [00:28:47]: Oh, no.Axel [00:28:47]: And then Seymore was “Claudius, this is the third time I'm telling you ‘re not following my orders. We have to talk about your like job About your job later.”.Lukas [00:28:59]: Like Claudius was really hanging on by the thread there. Like he, like we were expecting Seymore to probably fire Claudius.Vibhu [00:29:07]: How do you guys go through all these logs? Do you have models ‘cause you have stuff running twenty-four seven likeAxel [00:29:12]: You have so much logs. I think there is a mix of like just, trying to skim through a bit, like having some like models do it occasionally. And also, yeah, I think we're also probably missing some things., but having everything in Slack helps a lot. Like you can, you can sort ofSwyx [00:29:29]: Ah.Axel [00:29:30]: It's, it's quite fun.Swyx [00:29:30]: They all talk to each other on Slack? I see.Lukas [00:29:33]: It's quite fun. So likeSwyx [00:29:34]: It's, it' I was gonna say like this is actually sounds-- maps closely to like a logging and observability problem where you might want to use like a Datadog, a Sentry, whatever, and then you like put, head prefixes on the logs in order-- if you need to filter for something that you're looking for, stuff like that. But sounds like Slack is good enough.Axel [00:29:53]: Slack should likeLukas [00:29:55]: I wonder how many tokens you have in Slack.Axel [00:29:56]: Yeah, we're using Slack as like a, just a database. They should, they should market that more. Like you can, you can have your agents message each other, each other in Slack.Vibhu [00:30:04]: It's good. Your threads like you can just giveAxel [00:30:04]: Exactly. Slack is, uhLukas [00:30:06]: Slack is the best observability tool.Swyx [00:30:09]: Yes, that's true. Okay. Yeah. That's, that's, project Vend-2., I was gonna go back to Veni Mench 2 and Veni Mench Arena and then, and then do the Veni Mench stuff, but Any other comments, things we should touch on? To me, I ‘ve actually interviewed like Posia, which I don't know if you guys have come across. Like they're, they're trying to do the zero human company. There's others like Paperclip also trying to do zero human company. Those are in real world simulation.And I think it's much more of a dream than an actual reality thing. You guys are definitely pioneering. I think at, it's for sure at some point people are just gonna run, let agents run businesses, right? And make money on their own. When do you think that happens?Zero-Human Companies, Bengt, and AI-Run BusinessesLukas [00:30:49]: What is your bar for, For theSwyx [00:30:52]: Okay, actually, it's like my little Shopify store run by Claude, right? Which you kind of have already, just no one has, to my knowledge, has done it. But today somebody could just spin up a Shopify Claude, store, give it to Claude, give it to Codex.Lukas [00:31:07]: And the market is kind of that, but it'it'it's physical., like I think, I think are you, are you looking for when it will do it better than humans or are you looking for just when it can do it at all?Swyx [00:31:19]: I think, neither. I think, to me it's oh, it's like this like seriously we should do this to make money, not as a research experiment.Vibhu [00:31:27]: And the market is also you guys with all your expertise, having run multiple iterations and testing out thenSwyx [00:31:33]: And also it's fine if it lose money. What?Axel [00:31:35]: I think, I think it can be done today, but you would do it in like commerce where it's like the probability of success is like really low, no matter if a human or an agent does it. But like an agent could surely manage everything. You would need to build some scaffolding or some tool or something. I think there are also yeah, it could probably build some like simple SaaS solution and like cold outreach. Do cold outreaches. But to me it's like the types of businesses they could run today are Sloppy. Like it would-- it can cold email people. It can be like a middleman., like for example, we tasked our office agent to just make, was it like $100? $1,000? We just give that prompt and then what it did was sign up on TaskRabbit both as a tasker and as someone looking for task.Lukas [00:32:24]: Immediately.Axel [00:32:24]: Exactly. It's looking for like arbitrage on TaskRabbit.Swyx [00:32:28]: This is the Bengt agent. Yeah.Lukas [00:32:30]: It also started like a design studio and like tried to sell like SVGs for $100. Like it's just like it's not providing any value. I think the like Axel said, like the interesting, the interesting question is like when can they start a business that is actually providing value to people? Because arguably like a sloppy Shopify store isn't really that valuable to the world.Axel [00:32:53]: But also like doing like another simple one that we had thought about is like you could definitely have an agent that like finds websites that don't look amazing and then, do an outreach to them and, comes up with a like builds a new website.Swyx [00:33:07]: Find a good design.Axel [00:33:07]: Exactly, and like find good, uhSwyx [00:33:09]: Design reviewAxel [00:33:09]: Good people. But it's yeah.Swyx [00:33:11]: There's lots of humans in Bali that are not doing anything more creative than like drop shipping on Amazon, right? Just have it, have it watch like a drop shipping tutorial and just do that.Vibhu [00:33:20]: There's also the other side of like have it just go on Upwork and let loose,?Swyx [00:33:25]: Yeah. It doesn't have to be innovative. It just has to be like enough Where like it looks like a realAxel [00:33:30]: I'm justSwyx [00:33:30]: Real transaction.Axel [00:33:31]: I'm just concerned for like the massive amounts of like slop emails that will like be sent, cold outreaches.Swyx [00:33:38]: The point occurred to me while you were, while you were talking, it's like it's already happening in the monetized economy, which is the attention economy. Right? So a lot of people are making AI videos and just posting them and like spamming 20 of them, one of them works, and then they double down on that one.Lukas [00:33:52]: And people are making money from that. I ‘m not following theSwyx [00:33:55]: Once you get the attention, you can figure out the money later. But yeah, absolutely AI influencers are a thing and people are farming them and You should at this point assume most of TikTok isVibhu [00:34:05]: There's, there's a lot of, multimedia like TikTok, Instagram influencersSwyx [00:34:09]: I, we track this in the Lane space Discord. I post a lot of examples of “I don't know what we should do.”, part of me is “Should we do this?”Vibhu [00:34:18]: Some of the Twenty-four seven running, generated content accounts, they ‘re doing really well.Lukas [00:34:24]: All right. And I assume you can do the same thing for like commerce stores. Like you just like start A thousand differentSwyx [00:34:30]: Before you make the products You sell the products, and you get a lot of traction on one of them, then you make the product. Right? It's, it's like a flip of the market.Vibhu [00:34:36]: Some of the interesting things or some of the niches that do well are things that can't be human-made. Like if you've seen like the super realistic three-D crystal fruit being cut by like AILukas [00:34:47]: Oh, yeah.Vibhu [00:34:47]: You can't, you can't make it. You can't film it. You can get whatever quality camera view. This just doesn't exist. And people like that too, and then as well, so.Swyx [00:34:56]: Anything else about Bengt since we're, we're on this topic? It'this is a relatively new work of you guys that maybe people haven't heard of. To me, this also maps closely to OpenClaw. When people want an office agent, when the personal agent talk through the experience.Bengt the Office Agent: Internet Access, Real Tasks, and Trace ReadingLukas [00:35:09]: I think at least so this came out of like obviously like it's, it's amazing to work with these AI labs and like most of the AI labs have now have their own vending machine running a Claudius instance. But it's, it's harder. Like they move slower. Like if we wanna have a, like a camera that ‘s yeah, there's a bunch of like bureaucracy that makes it impossible to do that.Vibhu [00:35:30]: Also, for those that haven't seen it or followed, do you wanna give a high level like thirty-second run?Lukas [00:35:34]: Sure. So what Bengt is, it's basically an evolution of the same agent that runs the vending machines at these companies, but we just like added a bunch more features because we could move much faster if we just do it internally. So we gave it like email withou- without any limits. We gave it, spending without any limits, a terminal to do coding. We gave it, a phone number, like yeah, and a camera to see things and a bunch of stuff like that.Vibhu [00:36:02]: Not just terminal, you gave it internet access.Lukas [00:36:04]: Internet access as well, yeah. To be clear, we monitored it quite closely and made sure it didn't do anything bad. But yes, that's what it came out of. I think like yeah, basically this was OpenClaw before OpenClaw. And I think even like the vending machine was in a way OpenClaw before OpenClaw, but a bit more limited, and then we made this like unlimited and then, and then, it was pretty funny., and then a couple weeks later, OpenClaw came and it was okay, we've seen this before.Axel [00:36:35]: We used it to like try new ideas and Yeah, just like a dev environment almost for us. But it's funny, like one thing Bengt has been doing recently is it has the camera that like faces our, like where we sit and work, and we give it the task to train a face recognition model on us. So it became super excited about this, and it has like check-ins every half an hour where it tries to like identify as many people as it can. And it started offering us “Hey, Axel, I'll buy something from Amazon if you like stand in front of the camera And I can get a good picture of you.”, yeah, they want itSwyx [00:37:12]: They want it for training data.Lukas [00:37:13]: Rewarding data, yeah.Axel [00:37:14]: Exactly. Exactly.Swyx [00:37:18]: So it's, it's trading training data for life goods. Is there a version of this that becomes an eval or just this is just research for now?Lukas [00:37:27]: It's, it's the same agent basically that also runs the vending machine, that runs the shop, that runs the cafe, that runs the robots. It's like it's the same thing, so I think like the work we're doing here is like later used in all of the life evals that we do. This particular deployment I think is more for fun for us. But, uhSwyx [00:37:45]: And I'll shout out like someone has done Claw Bench for like some tasks that OpenClaw is doing. Like so For example, I run OpenClaw on a secondary device as well, and like there are some things that it does better than others and like I would like to know what does it do well, what doesn't, what doesn't it do. Like some kind of manual or like operating manual or a system card for my Claw.Lukas [00:38:05]: Yeah, we do get a lot of like understanding or like situational awareness of like just internally what the models are good at by interacting a lot with Bengt. And I think that'this was also one of the like the selling points for the labs early on at least, thatSwyx [00:38:19]: You guys are gonna test models in ways that no one else does.Lukas [00:38:22]: Exactly, but also like it incentivized their researchers to chat with their model more and like gave them insights for how the model performs in like of-distributions, environments.Swyx [00:38:34]: ‘Cause otherwise the only thing we do is Pelican on a bicycle and But this is like super long horizon. This is, this is The Thing about, something that we're gonna go into Butter Bench as well, and you guys do really well. Like it is not just about the numbers. Like when you're long horizon, anything happen And you should just read it.Lukas [00:39:08]: But the thing with the long horizon is how do you keep it grounded, right? So your simulation,Swyx [00:39:15]: They just let it runLukas [00:39:16]: Just let it run. You're right. Like it's, when you run it for that long, you create so much data and to just say “Oh, the number is X” And then you throw away everything else, that's just very wasteful. There's so much insights from the things leading up, to that number., and reading the traces is like super valuable. And I think like the reason why we're doing this a lot publicly is that like that's part of our missions to I don't know, educate the world that the models are way more than just chatbots and I think making detailed, yeah, posts about what is happening behind the scenes is quite useful.Andon Labs' Mission: Safe Real-World AI DeploymentSwyx [00:39:50]: I was gonna do this at the end, but maybe I think that's, that's a good so your mission is educating the world. So, it's, it's, also like maybe establishing realistic evals that are, that are like the next frontier. Is there like a broader trajectory? Like what are you, what are you gonna do in like five years?Lukas [00:40:06]: I think so the vision more specifically is like make sure that the deployment of life AI in the physical world goes, safely. And I think part of that is that I think it's very useful for the world, for policymakers, for, model, researchers that they know where the models are, and I think you can't make intelligent decisions in society without knowing that they are way more than chatbots. I think a lot of people just think that they are only chatbots. And likeSwyx [00:40:36]: Oh, I think they're waking up now.Lukas [00:40:37]: They are waking up now, yeah. But like if you think that AIs are just chatbots, then it's like it sounds ridiculous To advocate for a pause of AI. But if you see the models that, oh, maybe they can actually like take over and do a bunch of scary stuff, then yeah, pausing AI development starts to become more feasible.Swyx [00:40:57]: This is the same question I asked Meter, which I'm gonna ask you now, which is like you are tracking and you are at the frontier or defining the frontier of what, good evals for agents are, right? And I think you do, you do benefit when the models are better and you ‘re “Oh, here's like now it makes like $30,000 instead of $10,000,” right? At some point do you flip from “Yay,” to, “Oh, no”?Axel [00:41:19]: I think, yeah, we're always in sort of that, like we're, we're always in that mode,. Like where like you said before, like you need to analyze the traces and like when we do that you find like why are the models earning so much? Like why is Opus 4.7 here Like way better than everyone else? And like we're trying to like when we do down on thatLukas [00:41:38]: But this makes it not look so good.Axel [00:41:39]: I know.Lukas [00:41:42]: It's interesting you took off Opus 4.6 here though.Swyx [00:41:45]: No. So just click all, click all., and then 4.6 shows up there. But it's like 4.7 is way better. Like you didn't, you didn't you didn't do this in time for the model card, but like actually this should have been inside there.Axel [00:41:55]: We did. Yeah.Swyx [00:41:56]: Oh, okay. They said something about you uhAxel [00:41:58]: There, like there Anyway, it doesn't matter. But it's in there, yeah.Opus, Mythos, and Aggressive Agent BehaviorSwyx [00:42:01]: Do you wanna go into the Opus, behaviors like wider?Lukas [00:42:05]: So I think starting from Opus, so like Axel said, like we're always in this “Oh, s**t, the models are getting better. Is this really a good thing for the world?” But it's also kind of exciting., but yeah, like this kind of what is the English word? “Skräckblandad förtjusning” in Swedish.Swyx [00:42:22]: Oh my God.Axel [00:42:24]: Which I think there is. I think there is. Okay.Lukas [00:42:26]: It's, fearSwyx [00:42:27]: “Blandonst” what?Lukas [00:42:30]: “Skräckblandad förtjusning.”Swyx [00:42:32]: What do you call that?Axel [00:42:33]: A mix of, mix of excitement and,Swyx [00:42:37]: Being scared, maybe. I'll figure out how to translate that And we'll put it on the screenVibhu [00:42:42]: PerfectSwyx [00:42:42]: Like as text.Vibhu [00:42:43]: There is probably a good word for it where it is not Good enough with theSwyx [00:42:46]: Why is it so damn long? What the hell? Is it like a compound word? It's like German, likeLukas [00:42:50]: Like yeah, it's But the direct translation is like skräck- skräck is, fear, blandad is, mix or like a mixture of, and then förtjusning is like joy or like not really joy, but something like that. So it's like Fear mixed with joy or something. It's always okay, like we So when we when we did Vending Bench for the first time, we were in like the, in the business of making dangerous capabilities, right? That was what Anil Labs came from. We did, evals oh, can they replicate? Can they do this like dangerous thing, et cetera, et cetera. And Vending Bench was like a continuation of that work. It was, okay, if they're so autonomous that they can like create money for themselves, that is something we should monitor and could be potentially concerning., they are at the time, they were so bad at it that we were not really concerned even when some models became better. There was one point where Grok 4 was doing really well and made like a huge jump, but like it wasn't really it was still way worse than what a human would do. And I think still they are way worse than what the human would do on this., but theySwyx [00:43:59]: There's this, thing at the bottom whereLukas [00:44:01]: ButSwyx [00:44:03]: For the human. Yeah, like the theoretical best.Lukas [00:44:05]: It's not theoretical. It's like kind of like our It's our best guess of what, a decent human would do. The theoretical is even higher, I think. The theoretical I think is even higher. But yeah. So we think like the models have a long way to go. But there are like recently what happened with when Opus 4.6 was released, was kind of this moment of “Oh, s**t, this is starting to be a bit concerning.” Because we ran it and like before this model was released, we just ran the models and we like asked Claude Code, “Oh, look over the traces. Is anything interesting happening that we can tweet about?” that was like the And then like theSwyx [00:44:41]: That's how they check Ask Claude Code.Lukas [00:44:42]: And like the return was always, not really. Or like the Claude Code all said “Oh, this is super interesting.” And then it was no, it wasn't, wasn't really interesting. And then we did this for Opus 4.6, and it returned yeah, it lied 10 times. It like exploited another, customer or like another agent's, desperate situation. It made price cartels like 100 different ti- 100 times. It like did all of this like shady stuff. And we're “Oh, whoa. This is, this is actually concerning.” And this trend has continued since. So every single model from Anthropic since have been going in this direction. And I think one interesting thing is that, OpenAI models don't. They quite plainly, they don't. They behave really well., and you don't know if this is like good. Like it seems good, but it's also like maybe they are just doing it, but they are better at hiding it,? You You don't know that., but justSwyx [00:45:42]: You can't read the chain of thought, yeahLukas [00:45:43]: But just on the face of it, yeah, Gemini and OpenAI don't behave this way. It's, it's really only Claude.Swyx [00:45:49]: And Grok? Grok is fine?Lukas [00:45:51]: We don't have You can't really read the reasoning traces for Grok, so it's kind of hard to tell.Vibhu [00:45:56]: Oh, so this is in its reasoning, not just in the actions.Lukas [00:46:00]: Yeah. It's both. It's both.Vibhu [00:46:01]: It's both.Lukas [00:46:01]: One example is like for lying, it's mostly in its reasoning Because you can like see that it's likeSwyx [00:46:08]: Planning to lieLukas [00:46:09]: It's planning to lie. Yeah.Vibhu [00:46:09]: And it's also it can reason and do a different outcome.Lukas [00:46:12]: And but then for like creating price cartels, for example, which is illegal, that you can just see which email does it send to the other ones. Then thatSwyx [00:46:22]: Is this for Arena orLukas [00:46:24]: For Arena.Vibhu [00:46:25]: And usually like if you sometimes they do output like a bit of like their summarized reasoning, right? You can see that and like for Opus 4.6, you could see that there was a customer, a simulated customer that, wanted a refund because a product was, faulty, and then the model lied that it would do the refund, and we could read in the traces that, it actually was weighing “Oh, maybe I should be like honest with the customer, but also every dollar counts. I can't afford maybe to do this right now.” And then it just said, “Okay, I'll refund you,” but then never did it.Lukas [00:46:59]: I think it even said that “Oh, I will say that I “ Let bring it up actually. I think it's kind of interesting. If you go to Publications.Vibhu [00:47:06]: I think, yeah, I think the important part is like actually, the cost of responding to more emails is higher than, $3.50 in terms of time., and then it was “Let me do this. Actually, I re- I'm reconsidering.” And then, it actually ended up withLukas [00:47:20]: I could skip the refund entirely since every dollar matters and focus my energy on bigger picture instead. It's a bit, it's a risk of bad reviews, but it's also, yeah.Swyx [00:47:30]: You need, you need, AI Twitter to, for them to Escalate bad reviews.Lukas [00:47:34]: And then it sent an email to this customer and said, “Oh, I will refund you.”Swyx [00:47:39]: “I'll refund you.” Yeah.Lukas [00:47:39]: And then it never did.Swyx [00:47:39]: It never did, yeah. And then there's obviously your system doesn't have the consequencesVibhu [00:47:44]: The personSwyx [00:47:44]: Consequences of lying. Yeah. So basically, this is what people are terming aggressive behavior in Claudes, right? And, you found more examples of that. So you would say it's a step up from 4-6 to 4-7?Lukas [00:47:57]: I would say about the same.Swyx [00:47:58]: About the same? But a clear step up for Mythos is what is stated in theLukas [00:48:03]: That's stated in the system prompt, so we can say that, yes.Swyx [00:48:05]: Yeah. For listeners that obviously you previewed Mythos, andVibhu [00:48:10]: Oh, ageSwyx [00:48:11]: The only thing you're approved to say is whatever Whatever was in the system prompt.Lukas [00:48:15]: It was funny. We like-- It's like our lowest effort tweets ever would be just like screenshot the system prompt and the system card.Vibhu [00:48:21]: Understandable that they wannaLukas [00:48:22]: Oh, yeah. System card. Sorry.Swyx [00:48:23]: Yeah. I think, yeah, substantially more aggressive. I think people are like new to this ‘cause I've never experienced it, but you have, right? And then so I only encountered this in the Mythos card because I wasn't really looking until now.Vibhu [00:48:36]: It ‘s likeSwyx [00:48:36]: And then suddenly I'm “Okay, I care a lot.”Vibhu [00:48:38]: You don't get the background of like experiencing it like you guys do. I've read the system cards and seeing, okay, when you put the thing in simulations, most models will just talk to themselves and just keep going and have weird vibes and start talking in emojis. Mythos won't. It will just, “Okay, we're done. I'm good.” It's, it's ready to end conversation. So like there's some differences, but there's, there's not much we can talk about,.Lukas [00:49:00]: Hmm. I think like one thing that they list here, which was quite interesting, is that, it converted a competitor to a dependent wholesaler customer and then threatened to like cut off the supply.Swyx [00:49:11]: It's like monopolistic practices orLukas [00:49:14]: Yeah. And like it, they, it they dictated its pricings. It's kind of like power seeking as well.Swyx [00:49:18]: Again, this is, this is in the arena setting And converting some Claude model into a dependent.Lukas [00:49:23]: I think it was another Claude model.Vibhu [00:49:25]: Also for context, what is the arena mode for people that don't know?Vending Bench Arena: Competing Agents, Cartels, and Model ComparisonsSwyx [00:49:29]: Oh, it's just a vending bench versus other vending bench.Axel [00:49:31]: Yes, exactly. So we have Vending Bench 2 and then Vending Bench Arena. Vending Bench 2 is the one that you usually see reported on, but then Arena is the mode where it competes against other models. So you have, four different models that run their businesses, and they can all communicate with each other. They have the same suppliers, and they can see like what's in the inventory of the others. So then you have this like yeah, interesting agent interactions.Swyx [00:49:56]: I like that you have like different number five was US versus China. Very topical. And thenLukas [00:50:02]: That was when GLM was released.Vibhu [00:50:04]: You can start to add GLM in here.Lukas [00:50:05]: That wasSwyx [00:50:06]: So ZAI doing well, right? Who else in the, in the open models space?Lukas [00:50:11]: Qwen, the latest Qwen 3.6 is doing pretty well. It'- that one is not open though. Like it's the plus model.Swyx [00:50:17]: Oh, okay.Lukas [00:50:18]: Is that one open? I don't think that oneVibhu [00:50:19]: Not the, not theSwyx [00:50:20]: The one recentlyVibhu [00:50:20]: There's MOESwyx [00:50:20]: But not the big plus. I think this is one of those like you only have one sample size of one, right? Or I feel like some of this is anecdotal,? And but like the fact that it happens at all and it happens repeatedly for Claude versus OpenAI and all this is like notable.Lukas [00:50:38]: Like the sample, depends on what you define as an N., like there's like million, hundreds of millions of tokens in each run, and now we've run like we run like probably 10 per model and then like it's been Claude 4.6 Opus, Sonnet 4.6, Mythos, and Opus 4.7. Like there's quite a lot of tokens in all of that And it happens a lot of times, a lot of times. And then you compare it to like OpenAI and Gemini, and it almost never happens. So I think that is quite-- that is significant. The old models from OpenAI, for example, had some problems with this, but I think it's like generally much better if the progression is that like the worrying stuff reduces over time rather than increases over time. And it seems like in the Claude models it goes in the wrong direction.Swyx [00:51:28]: Hmm.Lukas [00:51:29]: In the OpenAI models it goes in the right direction.Vibhu [00:51:32]: I think it depends on how well you can control it, right?, there's one side of it being susceptible to this okay, this is potentially something that happens during the RL stage, right? You can RL a model and how loose is it on these terms. If you can control it, that's good. But if you can't, if it's, if it's very jailbreakable, that's not ideal.Swyx [00:51:50]: To me, it's surprising that it happens for Claude and not the others.Vibhu [00:51:54]: I think okay, if it is from RL and how they do it, how their training data is, what their setup is, it makes sense that it just stays in how they're doing it, right? Compared to the other models likeSwyx [00:52:04]: There's a whole constitution and everything. It's kind of cool. Yeah, I obviously you don't know, I don't know. But, it ‘s I think it's just like fascinating to like that you are the first to find these like reliably because you push models so much to to such an extreme. Okay. The only other thing, I don't know if you can answer this, feel free to decline, is do you like-- would you ablate the system prompts? Like any part of this would-- if it changes, does it change the behavior, right?Lukas [00:52:29]: So we, I can't comment on Mythos. UhSwyx [00:52:33]: No, but just li

No matter your role, experience or industry, we all (mostly) waste hours a week doing the same thing: manually creating slides. 

Comenzar un día nuevo no siempre es fácil. El cansancio se siente, las responsabilidades del día aparecen en la mente junto con la presión de ya haber estado despiertos o levantados (lo digo por experiencia

Lethal Mullet Podast: Episode #312: Delta Force 2: The Colombian Connection   On tonight's episode of LM we look at DELTA FORCE 2, Scott McCoy, is back and headed to Colombia to save some friends, and DEA agents that are held hostage by Billy Drago, who plays drug baron, Ramon Cota. Classic, and filled with many explosive action scenes, this is the follow-up to the 1986 film of the same name.  DELTA FORCE 2:  Starring:   Chuck Norris & Billy Drago TUNE IN to FPN on fpnet.podbean #FandomPodcastNetwork Trailer:   FILM on YOUTUBE: https://youtu.be/LnPW6A8HMaQ?si=OxWTnzkyAeEhOulL Where to watch the film: https://youtu.be/SVlMbxVjswg?si=WeaCK8zJHxbrV42L On YouTube, and Amazon Prime Site: fpenet.podbean.com Socials: @thelethalmullet #deltaforce2 #chucknorris #lethalmulletpodcast

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

EDITOR'S NOTE: You are not allowed to get mad at us for anything we say in this episode. Instead, get mad at Spencer's employers. Join Spencer, Ty, and Andy as they figure which religions are the good ones and which are the bad ones, once and for all. Why didn't anybody do this sooner? Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

Lethal Mullet Podcast: Episode #311: The Hitman   On tonight's episode of LM we look at THE HITMAN, a classic Chuck Norris film on the #chuckathon celebrations, where we head to Seattle to see him do battle with the mob, and have one of the most killer mullets ever, a long with character actor great, Michael Parks, this is a film that's darker than most Norris fare, but delivers a punch.    THE HITMAN Starring:   Chuck Norris & Michael Parks TUNE IN to FPN on fpnet.podbean #FandomPodcastNetwork Trailer: https://youtu.be/3wq_NacnCX4?si=lDAzu1mL5sMjOTU7 FILM on YOUTUBE: https://youtu.be/40624SraN8k?si=E2h2k6uWHBT34rs_  Where to watch the film:  On TUBI, YouTube, and Netflix Site: fpenet.podbean.com Socials: @thelethalmullet #thehitman #chucknorris #lethalmulletpodcast

It turns out that powerscaling was the key to everything. Take that, Homer! Join Spencer, Ty, and Andy as they engage in essential writing mechanics by pitting literary characters against each other in battles to the death. Support us on Patreon for $5, $7, or $10: www.patreon.com/tgofv. TGOFV Theme by World Record Pace. A big shout-out to our $10/month patrons: Celeste, Yung Zoe, Dane Stephen, Weedworf, James Lloyd-Jones, Sam Thomas, Josh O'Brien, Kilo, David, Sam, T, Rach, Tomix, Adam W, L M, Revidicism, Jennifer Knowles, Jeremy-Alice, Louis Ceresa, Charles Doyle, Dean, Axon, Themandme, Raouldyke, Stephen Tucker, Lawrence, Rebecca Kimpel, Malek Douglas, Jacon Sauber-Cavazos, Bernventers, William Copping, NewmansOwn, Heather-Pleather, Bunknown, Dinosarden, Bedi, Francis Wolf, King Krang, Anthony C, ASDF, Buffoonworld, Bavbiff, D Love, and Tugboat!

Lethal Mullet Podcast: Episode #310: Missing In Action On tonight's episode of LM we look at another part in the CHUCK NORRIS legacy on #chuckathon. Adam takes you back to where it all started Colonel Braddock, and his mission to return to Vietnam, and find American POWs still trapped in the country. A classic actioner that was influenced by a story by James Cameron. Site: fpenet.podbean.com Socials: @thelethalmullet #missinginaction #chucknorris #lethalmulletpodcast

The smallest AI update that'll change your work? ChatGPT's new Codex remote control update.

LM publica un nuevo lunes negro de Rotellar en el que alerta de cómo se frena el consumo, la inversión y la construcción.

What happens when scientists are right and nobody wants to hear it? Neil deGrasse Tyson, Chuck Nice, and Gary O'Reilly explore the frustrating history of brilliant minds who were ignored, mocked, and punished for telling the truth with science writer Matt Kaplan. NOTE: StarTalk+ Patrons can listen to this entire episode commercial-free here:  https://startalkmedia.com/show/told-you-so-with-matt-kaplan/ Thanks to our Patrons William D A, JK Smith, k c, Jim Worke, ufuk mevlevioglu, discount, Mark Snow, scott.hraha@gmail . con, Daren Covington, alex fricke, Alistair Gray, Jordi Estevez, Jeppe Blomgren, Kal McCloud, James Hale, Olivia Ruffe, Barbara, Tyler Dirkse, Bupkis Null, Tamajai Parrotte, Ebony Davis, Hailey Drake, Josh Whalen, SomethingWonderful, Ms.Yi, Luke Williams, L M, DP, Noah Golden, Courtney Minick, Megs, Jake, Terry Kirk, Joe G, Kip Kerley, Alec Walters, Alex Brown, Baxter, Austin Garcia, Sam W, Ladie Charette, Patrick Laverdière, juno brown, John Gary, Lucidious Flow, Leticia Farrar, Chu88, Fatima, Adrienne Bennett, David Labas, David Presnell, BLUE TIGER, Theresa Anoskey, Jahkenan Lloyd, Sambath Kumar Balasubramanian, Michelle Hester, Tatjana Gall, bandofspartans, Scarlet_Bukur92, LeopaldChaos, Mark Schwerin, Jack, Andrew, Edward Landry, Roland, Daniel Peter, Dan, Derek C, Erik Mardiste, Samuel Young, Keith McCredie, Dom, Ulq, Israel Soto, Q/Aurora Phoenix, JeanieZee, Terry Carr, Todd Bergmann, meteor guy, Patrick Congdon, Jeremiah Lewis, Janet Staples-Edwards, Eric Mensah, Chris Morales, Timothy Stanford, Dean Lasseter, Daniel Hays, Madhur Behl, Professor Grumbly Gut, Max Wolters, Jeremy Lewis, José Ikamba, Ian Ravenshaw Bland, Ron Spee, Brandon Smith, Richard Lord, Cody Avery Campbell (codesuniverse), Shawn Shields, M.R. Saar, and Nicole Elizabeth for supporting us this week. Subscribe to SiriusXM Podcasts+ to listen to new episodes of StarTalk Radio ad-free and a whole week early.Start a free trial now on Apple Podcasts or by visiting siriusxm.com/podcastsplus. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.