[0:00] [Chamath] Alright, everybody. Welcome back to the number one podcast in the world. It's the all in podcast episode two seventy four. Saks is out today, but we're very lucky to have Gavin Baker from a treaties management joining us, the spicy takes must flow. Welcome back to the program. [0:19] [Gavin Baker] Bestie Gavin. Thanks for having me. [0:21] [Chamath] Always love it. It's been a huge week in tech. We can start with the SpaceX and OpenAI IPOs. We've got Andre Corpathy joining Anthropic, NVIDIA crushing it. So many different places to go, but I think we'll start with Andre Carpathy joining Anthropic. Carpathy is only thirty nine years old. He's already a legend in the tech industry if you don't know him. I believe he's also coming to [0:42] [David Sacks] liquidity. Yetchumah? He's gonna keynote on Monday morning. Oh, fantastic. Oh, to know Tuesday Tuesday Tuesday day two. [0:48] [Chamath] Think he's Keynote. Okay. As is Gavin. Gavin will be there. Founding [0:52] [David Sacks] member Gavin anchoring [0:54] [Chamath] day two as well. Excellent. Yeah. This is Gavin's [0:56] [David Sacks] second appearance. Hi. Look at those two bookmarks, Andre Carpathy and Gavin Baker. [1:00] [Jason] Oh, yeah. [1:00] [Chamath] You know? Liquidity pulls in the stars. Obviously, Andre was a founding member of OpenAI. He led the self [1:08] [David Sacks] driving team. Also, hold on. Gavin is gonna help us judge the best ideas section as well. Excellent. I don't know if you know that, Gavin, but you're [1:15] [Friedberg] a judge. [1:16] [Chamath] You're gonna [1:16] [Gavin Baker] be judged. I'm up for anything, man. [1:18] [Chamath] I'm easy. Yes. Kirpathy also coined the term vibe coding. He recently built auto research. I think we talked about that here a bit. That's an open source training tool. It helps AI models improve themselves by running five minute experiments. That got over eighty two thousand stars in GitHub. He did that, like, as a weekend experiment, and all these civilians started building their own, recursive LLMs. Really inspiring. And the Andre Carpathy skills is a tool based on his set of principles for clawed code, and somebody just released that. And so that's just pretty crazy when you think about it. He's gonna be in charge of a new pre training team at Anthropic. The focus obviously being recursive self improvement. In other words, they're [2:01] [David Sacks] going to have Claude improve itself, [2:02] [Chamath] and they've already talked a little bit about AI improving AI over at Anthropic. Shmok, what's your take on this? Is this, super important in twenty twenty six? Obviously, Carpathia is super well respected. He's obviously, you know, one of the true talents in the space, but, hey, we're in a we're in a different inning than we were, say, ten years ago when he was at Tesla or five years ago when he cofounded OpenAI. [2:26] [David Sacks] You know what's interesting? If go back to, like, Google, the culture of Google, which they got right was the singular technical talents there. They were singled out and they were called Google Fellows. I don't know if guys remember this, Yes. Hamid Singhal, Shrida Rameswamy, Jeff Dean. These guys are stars. And what's interesting is if you track what folks, particularly Jeff Dean, guess, now because the other two aren't there anymore, but what they did inside of Google, it's like wave upon wave. They were at the foot of those waves. What's interesting about Andre is he's been at the wave upon wave of AI. He was probably the first person that really commercialized the Richard Sutton bitter lesson essay when he was leading FSD at Tesla, which was really about the brute force computation. And I remember him telling me this story. I don't know if he said this publicly or not, but where he spent a portion of his time, I wanna say a quarter of his time, labeling data. Could you imagine, twenty sixteen, like, hand labeling video data from Tesla's? So he did that, then he's cofounder of OpenAI. He's a star, and he's an exceptional human being, and he's super curious. And then what he's done as a kind of a free agent is also quite impressive. So I think that this is a really important deal. I think he's one of these really curious people that can be sent off, they'll just go invent new things. And I think this idea of recursive self learning puts these models on a combination of overdrive and autopilot. And so if you put those two things together, I think that you start to you can potentially live out this idea that there's an order of magnitude improvement on a yearly basis. So, like, this new form of Moore's law. So then the model quality just goes absolutely parabolically just like [4:21] [Chamath] this straight up. I think that bunch of compute at the problem, and these things learn really quick, I think is the high order bit there. Gavin, what are you what's your take on Anthropic's recent success and their massive [4:34] [Gavin Baker] hiring binge? The success is extraordinary. It's undeniable. I think the fact that they are now they were EBIT positive per The Wall Street Journal in the most recent quarter is a really important fact for kind of the whole AI narrative because now there's you know, you could talk about circular funding. You could talk about ROI, we could go look at the ROIC of the hyperscalers. But if OpenAI and Anthropic are at, call it, a hundred billion dollars of ARR now with eighty percent -ish gross margins on inference, like, the returns are there. And then if we add in and they're growing really fast. If we in Gemini, we add in Cursor, we in XAI, we add an open source. You know, it's it's not hard to see two hundred, three hundred, four hundred billion dollars of ARR at the end of this year at high margin. Across all [5:29] [Chamath] of the Across all of language models. You're talking specifically about the private language model companies, maybe not [5:36] [Gavin Baker] Google, which is [5:36] [Chamath] No. Was Google. [5:37] [Friedberg] You're [5:38] [Gavin Baker] including Google. Okay. But I was excluding, you know, a lot of the returns to this GPU spend have come from, you know, better recommender systems at Facebook and Google, Amazon, better ad targeting, better ad [5:51] [Jason] measurement. [5:52] [Gavin Baker] Sure. So excluding that and just narrowing it to LLMs, which The tokens. Strict as possible definition. And it seems like there's going to be a really strong ROI this year, even excluding what are still some of the most economically important and profitable use cases for GPUs and AI infrastructure. I do think what Carpathy is working on, recursive self improvement is really important, and unlocking that in continual learning, you know, maybe the two final frontiers for for AI. And just the idea of rehearsal self improvement that the model, while it is training, you know, during a forward pass has input into its training or another model has input into the training, I think that could be really powerful. And I think Chamath's statistics of you know, ten x ing every year, you know, might seem conservative if that comes to pass. And then, of course, continue learning is the holy grail where the model learns from experiences the way humans do. Yeah. And that's something we haven't unlocked yet. And those those those two combined, I think, would they might pull the future forward in a very [7:13] [Chamath] real way. Yeah. And we have right now. Anthropic has a decent lead on everybody else, whether it's three months or six months. Obviously, they're probably six, twelve months ahead of open source. Maybe they're three, six, nine months ahead of their contemporaries, but they have a lead. You put Carpathy in there, Friedberg. Now you have Carpathy. He does recursive. And at some point, and it may have even occurred at Anthropic, the AI is going to be improving the language model more than the humans in the loop are doing it. Obviously, they're orchestrating at Fr