Raschka & Lambert on Lex Fridman: Why Post-Training Won 2025

Lex Fridman 0:00 The following is a conversation all about the state-of-the-art in artificial intelligence, including some of the exciting technical breakthroughs and developments in AI that happened over the past year, and some of the interesting things we think might happen this upcoming year. At times, it does get super technical, but we do try to make sure that it remains accessible to folks outside the field without ever dumbing it down. It is a great honor and pleasure to be able to do this kind of episode with two of my favorite people in the AI community, Sebastian Raschka and Nathan Lambert. They are both widely respected machine learning researchers and engineers who also happen to be great communicators, educators, writers and Twitterers, X posters. Sebastian is the author of two books I highly recommend for beginners and experts alike. First is Build a Large Language Model From Scratch and Build a Reasoning Model From Scratch. I truly believe in the machine learning computer science world, the best way to learn and understand something is to build it yourself from scratch. Nathan is the post-training lead at the Allen Institute for AI and author of the definitive book on reinforcement learning from human feedback. Both of them have great X accounts, great Substacks. Sebastian has courses on YouTube, Nathan has a podcast, and everyone should absolutely follow all of those. This is the Lex Fridman Podcast. To support it, please check out our sponsors in the description, where you can also find links to contact me, ask questions, give feedback, and so on. And now, dear friends, here's Sebastian Raschka and Nathan Lambert.

Lex Fridman 1:57 So I think, uh, one useful lens to look at all of this through is the DeepSeek, so-called DeepSeek moment. This happened about a year ago, in January 2025, when the open-weight Chinese company, DeepSeek, released DeepSeek-R1 that, uh, I think it's fair to say, surprised everyone with, uh, near or at state-of-the-art performance, with allegedly much less compute for much cheaper. And from then to today, the AI competition has gotten insane, both on the research level and the product level. It's just been accelerating. Let's discuss all of this today, and maybe let's start with some spicy questions if we can. [chuckles] Uh, who is winning at the international level? Would you say it's the set of companies in China or the set of companies in the United States? And Sebastian, Nathan, it's good to see you guys. Uh, so Sebastian, who do you think is winning?

Sebastian Raschka 2:53 Um, so winning is [chuckles] a very broad, uh, you know, term. I, I would say you mentioned the DeepSeek moment, and I do think DeepSeek is definitely winning the hearts of the people who work on open-weight models because they share these as open models. Um, winning, I think, has multiple timescales to it. We have today, we have next year, we have in ten years. One thing I know for sure is that, um, I don't think nowadays, 2026, that there will be any company who is, let's say, having access to a technology that no other company has access to. And that is mainly because researchers are frequently changing jobs, changing labs, they, uh, rotate. So I don't think there will be a clear winner in terms of technology access. However, I do think there will be, uh, the differentiating factor will be budget and hardware constraints. So I don't think the ideas will be proprietary, but the way or the resources that are needed to implement them. And so I don't see currently a take-it-all scenario where a winner takes it all. I, I can't see that at the moment.

Lex Fridman 3:59 Uh, Nathan, what do you think?

Nathan Lambert 4:00 You see the labs put different energy into what they're trying to do, and I think to demarcate the point in time when we're recording this, um, the hype over Anthropic's Claude Opus 4.5 model has been absolutely insane. Which is just-- I mean, I've used it and built stuff in the last few weeks, and it's, uh, it's almost gotten to the point where it feels like a bit of a meme in terms of the hype. And it's kind of funny because this is very organic, and then if we go back a few months ago, we can get the release date, and the notes is Gemini 3 from Google got released, and it seemed like the marketing and just, like, wow factor of that release was super high. But then at the end of November, Claude Opus 4.5 was released, and the hype has been growing. But Gemini 3 was before this, and it kind of feels like people don't really talk about it as much, even though when it came out, everybody was like, "This is, um, Gemini's moment to retake kind of Google's structural advantages in AI." And Gemini 3 is a fantastic model, and I still use it. It's just kind of differentiation is lower. And I agree with Sebastian, what you're saying with all these, like, the idea space is very fluid, but, um, culturally, Anthropic w-- is known for betting very hard on code, which this Claude code thing is working out for them right now. So I think that even if the ideas flow pretty freely, so much of this is bottlenecked by human effort and kind of culture of organizations, where Anthropic seems to at least be presenting as the least chaotic. It's, is, is a bit of an advantage, and if they can keep doing that for a while... But on the other side of things, there's a lot of ominous technology from China, where there's way many more labs than DeepSeek. So DeepSeek kicked off a movement within China, I say kind of similar to how ChatGPT kicked off a movement in the US where everything had a chatbot. There's now tons of tech companies in China that are releasing very strong frontier open-weight models, to the point where I would say that DeepSeek is kind of losing its crown as the preeminent open model maker in China, and the likes of, um, Z.AI with their GLM models, MiniMax's models, um, Kimi Moonshot, e-especially in the last few months, have shone more brightly. The new DeepSeek models are still very strong, but that's kind of a-- it, it could look back as a big narrative point where in 2025, DeepSeek came, and then all-- and it kind of provided this platform for way more Chinese companies that are releasing these fantastic models to kind of have this new type of operation. So these models from these Chinese companies are open weights, and depending on this trajectory of business models that these American companies are doing-... could be at risk, but currently, a lot of people are paying for AI software in the US, and historically, in China and other parts of the world, people don't pay a lot for software.

Lex Fridman 6:38 So some of these models, like DeepSeek, uh, have the love of the people because they are open weight. Uh, how long do you think, uh, the Chinese companies keep releasing open-weight models?

Nathan Lambert 6:47 I would say for a few years. I think that, like in the US, there's not a clear business model for it. I have been writing about open models for a while, and these Chinese companies have realized it, so I get inbound from some of them. And they're smart and realize the same constraints, which is that a lot of US comp-- tech companies and other IT companies won't pay for a API subscription to Chinese companies for security concerns. This has been a long-standing, um, habit in tech, and the people at these companies then see open-weight models as an ability to influence and take part of a huge growing AI expenditure market in the US, and they're very realistic about this, and it's working for them. And I think that the government will see that that is building a lot of influence internationally in terms of uptake of the technology. So there's gonna be a lot of incentives to keep it going, but building these models and doing the research is very expensive. So at some point, I expect consolidation, but I don't expect that to be a story of 2026, where there will be more open model builders throughout 2026 than there were in 2025, and a lot of the notable ones will be in China.

Lex Fridman 7:50 You were gonna say something?

Sebastian Raschka 7:51 Um, yes. You mentioned DeepSeek losing its crown. I do think to some extent, yes, but we also have to consider, though, they are still, I would say, slightly ahead, and the other ones-- it's not that DeepSeek got worse, it's just, like, the other ones are using the ideas from Deep, DeepSeek. For example, you mentioned Kimi, same architecture. They're training it, and then again, we have this leapfrogging, where they might be, at some point in time, a bit better because they have the more recent model. And I think this comes back to, um, the, the fact that there won't be a clear winner. It's-- well, it will just be like, like that. One person releases something, the other one comes in, and the, the recent-- the most recent model is probably always the best model.

Nathan Lambert 8:30 Yeah. We'll also see that Chinese companies have different incentives. So, like, DeepSeek is very secretive, where some of these start-ups are, like the MiniMaxes and Z.AI's of the world, those two literally have filed IPO paperwork, and they're trying to get Western mindshare and do a lot of outreach there. So I don't know if these incentives will kind of change the model development, 'cause DeepSeek famously is built by a hedge fund-

Sebastian Raschka 8:53 Mm-hmm

Nathan Lambert 8:53 ... High-Flyer Capital, and we don't know exactly what they [chuckles] like, we don't know what they use the models for or if they care about this.

Lex Fridman 8:59 They're secret in terms of communication. They're not secret in terms of the technical reports-

Nathan Lambert 9:03 Mm

Lex Fridman 9:03 ... that describe how their models work. They're still open on that front.

Nathan Lambert 9:06 Mm.

Lex Fridman 9:06 And we should also say, on the Opus- four five hype, there's the layer of, uh, something being the darling of the X echo chamber, on Twitter echo chamber, and the actual amount of people that are using the model. I think it's probably fair to say that ChatGPT and Gemini are focused on the broad user base that just want to solve problems in their daily lives, and that user base is gigantic. So the hype about the coding may not be representative of the actual use.

Sebastian Raschka 9:39 I would say also, um, a lot of the usage patterns are, like you said, name recognition, brand, uh, and, and stuff, but also muscle memory almost, where, um, you know, like, ChatGPT has been around for a long time. People just got used to using it, and it's kind of like almost like a flywheel. They recommend it to other users and that stuff. One interesting point is also the customization of, uh, LLMs. For example, ChatGPT has a memory feature, right? And so you may have a subscription, and you use it for personal stuff, but I don't know if you want to use that same thing at work, you know, because that's a boundary between private and work. If you're working at a company, they might not allow that, or you may not want that. And I think that's also an interesting point, where you might have multiple subscriptions. One, one is just clean code. It keeps-- mm, has nothing of your personal images that you... or hobby projects in there. It's just, like, the work thing, and then the other one is your personal thing. So I think that's also something where two different use cases, and it doesn't mean you only have to have one. It's, it's-- I think the future is also multiple ones.

Lex Fridman 10:38 What model do you think won twenty twenty-five, and what model do you think is gonna win 'twenty-six?

Nathan Lambert 10:43 I think in the context of a consumer chat bots, is a question of, are you willing to bet on Gemini over ChatGPT?

Lex Fridman 10:50 Mm-hmm.

Nathan Lambert 10:50 Which I would say in my gut feels like a bit of a risky bet, because OpenAI has been the incumbent, and there's so many benefits to that in tech. But I think the momentum, if you look at twenty twenty-five, was on Gemini's side, but they were starting from such a low point [chuckles] , I think, um, RIP Bard, and these earlier attempts, uh, uh, of getting started. I think huge credit for them for powering through the organizational cr- chaos to make that happen. But also, it's hard to bet against Chat-- to OpenAI because they always come off as, as so chaotic, but they're very good at landing things.

Lex Fridman 11:26 Mm.

Nathan Lambert 11:26 And I think, like, personally, I have very mixed reviews of GPT 5, but it had to have saved them so much money, with the headline feature being a router, where most users are no longer charging, like, charging their GPU costs as much. So I think it's very hard to dissociate the things that I like out of models versus the things that are gonna actually be a general public differentiator.

Lex Fridman 11:50 What do you think about twenty twenty-six? Who's gonna win?

Nathan Lambert 11:52 I'll say something, even though it's risky. I will say that I think Gemini will continue to take progress on ChatGPT. I think Google's scale, when both of these are operating at such extreme scales, and, like, G- Google has the ability to separate that research and product a bit better, where you hear so much about OpenAI being chaotic operationally and chasing the high-impact thing, which is a very start-up culture. And then on the software and enterprise side, I think Anthropic will have continued to success, as they've again and again been set up for that. And o- obviously, Google's Cloud has a, a lot of offerings, but I think this kind of like Gemini name brand is important for them to build, and, and Google's Cloud will continue to do well. But that's kind of a c- more complex thing to explain in the ecosystem because that's competing with the likes of Azure and AWS rather than on the model provider side.

Lex Fridman 12:40 So in infrastructure, you think TPUs give an advantage?

Nathan Lambert 12:45 ... largely because the margin on NVIDIA chips is insane, and Google can develop everything from top to bottom to fit their stack and not have to pay this margin, and they've had a head start in building data centers. So all of these things that have both high lead times and very hard margins on high costs, Google has a, just kind of a historical advantage there. And, uh, if there's gonna be a new paradigm, it's most likely to come from OpenAI, where they're kind of-- their research division, again and again, has kind of shown this ability to land a new research idea or a product. I think, like, deep research, Sora, o1 thinking models, like, all these definitional things have come from OpenAI, and that's gotta be one of their top traits as an organization. So it's kind of hard to bet against that, but I think a lot of this year will be about scale and optimizing what could be described as low-hanging fruit in models.

Lex Fridman 13:37 And clearly, there's a trade-off between intelligence and speed. This is what ChatGPT Five was trying to solve behind the scenes. It's like, do people actually want intelligence, the broad public, or do they want speed?

Sebastian Raschka 13:52 I think it's a nice variety, actually, or the option to, uh, have a toggle there. I mean, first, for my personal usage, most of the time when I look something up, I use ChatGPT to ask a quick question, get the information I wanted fast. For, you know, most daily tasks, I use the quick model. Nowadays, I think the auto mode is pretty good, where you don't have to specifically say thinking or, you know, non-thinking and stuff. Then again, I also sometimes want the pro mode. Very often, what I do is when I have something written, I put it into a ChatGPT and say, "Hey, do a very thorough check. Is-- are all my references correct? Are all my thoughts correct? Uh, did I make any formatting mistakes, and are the figure numbers wrong?" Or something like that. And I don't need that right away. It's something, okay, I finish my stuff, maybe have dinner, let it run, come back, and go through this [chuckles] and I think-- see, this is where I think it's important to have this option. I would go crazy if for each query, I would have to wait thirty minutes or ten minutes even. [laughing]

Nathan Lambert 14:46 That's me.

Sebastian Raschka 14:47 Yeah. [laughing]

Nathan Lambert 14:48 Um, I'm, like, sitting over here losing my mind that you use the router and the non-thinking model. I'm like, oh, how do you-- how do you live with-- how do you live with that?

Lex Fridman 14:55 Yeah.

Nathan Lambert 14:55 It's, like, my reaction. I'm-- been heavily on ChatGPT for a while. Um, never touched Five non-thinking. I find it, the, its tone and then its propensity of errors. It just, like, has a higher likelihood of errors. Some of this is from back when OpenAI released o3, which was the first model to do this deep search and find many sources and integrate them for you. So I became habituated with that. So I will only use GPT 5.2 thinking or pro when I'm finding any sort of information query for work, whether that's a paper or some code reference that I found, and it's just like I, I will regularly have, like, five pro queries going simultaneously, each looking for one specific paper or feedback on an equation or something.

Sebastian Raschka 15:38 I have a fun example of where I just needed the answer as fast as possible. For this podcast, before [chuckles] I was going on the trip, um, I have, like, a local GPU running at home, and I wanted to run a long, uh, RL experiment. And usually, I also unplug things because you never know if you're not at home, you don't want to have things plugged in, and I accidentally unplugged the, the GPU. [chuckles] It was like my wife was already in the car, and it's like, "Oh, dang!" And then basically, I wanted as fast as possible a Bash script that runs my different, uh, experiments and the evaluation. And I did something I know. I learned how to use the Bash, uh, interface, or Bash terminal, but i- in that moment, I just needed, like, ten seconds. Give me the command.

Lex Fridman 16:18 This is a hilarious situation, but yeah. So what did you use?

Sebastian Raschka 16:21 Uh, so I did the non-thinking fastest model. It gave me the Bash, uh, command. I, uh, to chain different, uh, scripts to each other, and then the thing is, like, you have the T thing where you want to route this to a log file. Top of my head, I was just, like, in a hurry. I could have thought about it myself.

Lex Fridman 16:37 By the way, I don't know if this is a representative case, wife waiting in the car.

Sebastian Raschka 16:40 [laughing]

Lex Fridman 16:40 You have to run, you unplug the GPU.

Sebastian Raschka 16:42 [laughing]

Lex Fridman 16:42 You have to generate a Bash script.

Sebastian Raschka 16:44 Yeah.

Lex Fridman 16:44 This sounds like a movie-

Nathan Lambert 16:45 I, I use-

Lex Fridman 16:45 ... like Mission Impossible.

Nathan Lambert 16:46 I use Gemini for that. So I use thinking for all the information stuff, and then Gemini for fast things or stuff that I could sometimes Google, which is, like, it's good at explaining things, and I trust that it has this kind of background of knowledge, and it's simple. And the Gemini app has gotten a lot better-

Lex Fridman 17:01 Mm

Nathan Lambert 17:01 ... and it's good for that sort of things. And then for code and any sort of philosophical discussion, I use Claude Opus 4.5, also always with extended thinking. E- extended thinking and inference time scaling is just a way to make the models, um, marginally smarter, and I will always edge on that side when the progress is very high because you don't know when that'll unlock a new use case. And then sometimes use Grok for, um, real-time information or finding something on AI Twitter that I knew I saw, and I need to dig up, and I just fixated on. Although, when Grok 4 came out, the Grok 4, what is super heavy, which was, like, their pro variant, was actually very good, and I was pretty impressed with it. And then I just kind of, like, muscle memory, lost track of it with having the ChatGPT app open. So I use many different things.

Lex Fridman 17:45 Yeah, I actually do, uh, do use Grok 4 heavy for debugging. For, like, hardcore debugging that the other ones can't solve, I find that it's the best at... And I-- it, it's interesting because you say ChatGPT is the best interface. Uh, for me, for that same reason, but this could be just momentum-

Nathan Lambert 18:04 Mm-hmm. Yep

Lex Fridman 18:04 ... uh, Gemini is the better interface for me, I think because I fell in love with their best needle in the haystack. If I ever put something that has a lot of context, but I'm looking for very specific kinds of information, make sure it tracks all of it, I find at least, uh, that Gemini, for me, has been, uh, the best. So it's funny with some of these models, if they win your heart over-

Nathan Lambert 18:28 Mm-hmm

Lex Fridman 18:28 ... for one particular feature, at one, on a one particular day, for that particular query, that prompt, you're like: This model is better. And so you'll just stick with it for a bit until it does something really dumb. There's, like, a threshold effect.

Nathan Lambert 18:43 Mm.

Lex Fridman 18:43 Some smart thing, and then you fall in love with it, and then it does some dumb thing, and you're like, "You know what? I'm gonna switch and try Claude or ChatGPT," and all that kind of stuff.

Sebastian Raschka 18:51 ... this is exactly like you use it until it breaks, until you have a problem, and then, then you ch- uh, change, uh, the LLM. And I think it, it's the same how we use anything, like our favorite text editor, um, operating systems, or the browser. I mean, there are so many browser options: Safari, Firefox, Chrome. All the- relatively similar, but then there are ex- edge cases, maybe extensions you want to use, and then you switch. But I don't think there is anyone who types the same thing, like the website, into different browsers and compares them. You only do that when the website doesn't render, if something breaks, I think. So that's, that's a good point. I think you use it until it breaks, and then you explore other options, I think.

Nathan Lambert 19:28 On the long context thing, I was also a Gemini user for this, but the GPT 5.2 release blog had, like, crazy long context scores, where a lot of people were like: "Did they just figure out some algorithmic change?" It went from, like, thirty percent to, like, seventy percent or something in this minor model update. So it's also very hard to keep track of all of these things, but now I'm m- look more favorably at GPT 5.2's long context. So it's just kind of like, how do I actually get to testing this? Just never-ending battle.

Lex Fridman 19:57 Well, it's interesting that none of us talked about the Chinese models from a user usage perspective. What does that say? Does that mean the Chinese models are not as good, or does that mean we're just very biased, uh, and US-focused?

Sebastian Raschka 20:11 I do think that that's c- currently the discrepancy between just the model and the platform. So I, I think the open models, they are more known for the open weights, not their platform yet.

Nathan Lambert 20:21 There are also a lot of companies that are willing to sell you the open model inference at a very low cost. I think, like, OpenRouter, it's easy to do the look at multi-model things. You could run DeepSeek on Perplexity. I think all of us sitting here are like, we use OpenAI GPT 5 Pro very consistently. We're all willing to pay for the marginal intelligence gain. And anyone that's... Like, the-- these models from the US are better, and they-- in terms of the outputs, I think the, the question is, will they stay better for this year and for years going? But it's like, so long as they're better, I'm gonna pay for it to use them. I think there's also analysis that shows that, like, the way that the Chinese models are served, this, you could argue, due to export controls or not, is that they use fewer GPUs per replica, which makes them slower and have different errors, and it's like speed and intelligence. If these things are in your favor as a user, I think in the US, a lot of users will go for this, and I think that that is one thing that will spur these Chinese companies to want to compete in other ways, whether it's, like, s- free or substantially lower costs, or it'll breed creativity in terms of offerings, which is good for the ecosystem. But I just think of the simple thing is the US models are currently better, and we use them, and I tried Chinese mo- I tried these other open models, and I'm like, "Fun, but not gonna..." I don't go back to it.

Lex Fridman 21:38 Uh, we didn't really mention programming. That's another use case that a lot of people deeply care about. So I use basically half-and-half Cursor and Claude Code because they're-- I find them to be, like, fundamentally different experience and both useful. Uh, what do you guys... You program quite a bit-

Sebastian Raschka 21:56 Mm-hmm.

Lex Fridman 21:56 -so what, what do you use? What's the current vibe?

Sebastian Raschka 21:59 So I use the Codex plugin for VS Code. Uh, you know, it's very convenient. It's just like a plugin, and then it's a chat interface that has access to your repository. I know that Claude Code is, I think, a bit different. It is a bit more agentic. It touches more things. It does a whole project for you. I'm not quite there yet, where I'm comfortable with that because, ah, maybe I'm a control freak, but I still [chuckles] would like to see a bit what's going on, and Codex is kind of like, right now, for me, like, the sweet spot where it is helping me, but it is not taking completely over.

Lex Fridman 22:29 I should mention, one of the reasons I do use Claude Code is to build the skill of programming with English. I mean, the experience is fundamentally different. You're, as opposed to micromanaging the details of the process of the generation of the code and, uh, looking at the diff, which you can in Cursor, i- if that's the IDE you use, and, and in changing, altering, looking and reading the code and understanding the code deeply as you progress, versus just kind of like thinking in this design space and just guiding it at this, uh, macro level, which I think, uh, is another way of thinking about the programming process. Also, we should say that Claude Code, it just seems to be somehow a better utilization of Claude Opus 4.5.

Nathan Lambert 23:18 It's a good side by side for people to do. So you can have Claude Code open, you can have Cursor open, and you can have VS Code open, and you can select the same models on all of them-

Lex Fridman 23:26 Mm-hmm

Nathan Lambert 23:26 ... and ask questions, and it's very interesting, like the, like, Claude Code is way better in that domain.

Lex Fridman 23:31 Yeah.

Nathan Lambert 23:31 It's remarkable.

Lex Fridman 23:32 All right, we should say that both of you are, are legit on multiple fronts: researchers, programmers, educators, tweeters. [laughing] And on the book front, too. So Nathan, at some point soon, hopefully, has an RLHF book coming out.

Nathan Lambert 23:50 It's available for pre-order, and there's a full digital preprint, just making it pretty and better organized for the physical thing, which is a lot of why I do it, because it's fun to create things that you think are excellent in the physical form when so much of our life is digital.

Lex Fridman 24:05 I should say, going to Perplexity here, Sebastian Raschka is a machine learning researcher and author known for several influential books. A couple of them that I wanted to mention, which is a book I highly recommend, Build a Large Language Model From Scratch, and the new one, Build a Reasoning Model From Scratch. So I'm really excited about that. Building stuff from scratch is one of the most powerful ways of learning.

Sebastian Raschka 24:27 Honestly, building an LLM from scratch is a lot of fun. It's also a lot of r- to learn, and like you said, it's probably the best way to learn how something really works. 'Cause you can look at figures, but figures can have mistakes. You can look at con- uh, concepts, explanations, but you might-... misunderstand them, but if you see the co- there is code and the code works, you know it's correct. I mean, there's no misunderstanding. It's like it's precise, otherwise it wouldn't work. And I think that's like kind of like the beauty behind coding. It is kind of like it doesn't lie, it's math, basically. So even though with math, I think you can have mistakes in a book, you would never notice. Because you're not running the math when you are reading the book, you can't verify this, and with code, what's, what's nice is you can verify it.

Lex Fridman 25:09 Yeah, I agree with you about the LLM From Scratch book. It's nice to tune out everything else, the internet and so on, and just focus on the book. But, you know, I read, uh, several, like, you know, uh, history books. It's just less l- lonely somehow. It's really more fun. Like, uh, for example, on the programming front, I think it's genuinely more fun to program with an LLM.

Sebastian Raschka 25:31 Mm.

Lex Fridman 25:31 And I think it's genuinely more fun to read with an LLM.

Sebastian Raschka 25:35 Mm-hmm.

Lex Fridman 25:36 But you're right, like, this distraction should be minimized. So it's, uh, you use the LLM to basically enrich the experience, maybe add more context.

Sebastian Raschka 25:48 Mm.

Lex Fridman 25:48 Maybe the... I just, the rate of aha moments for me in a small scale is really high with LLMs.

Sebastian Raschka 25:54 Hundred percent. I would s- I also want to correct myself. I'm not suggesting not to use LLMs. Uh, I suggest doing it in multiple passes, like one pass, just offline focus mode, and then after that, uh, I mean, I also take notes, but I, I try to resist the urge to i- immediately look things up. I, I do a second pass. It's just like, for me, more structured this way, and I get le- I mean, sometimes things are answered in the chapter, but sometimes also it just helps to let it sink in and think about it. Other people have different preferences. I would highly recommend using LLMs when reading books. For me, it's just, it's not the first thing to do, it's like the second pass.

Lex Fridman 26:30 By way of recommendation, as you said, I do the opposite. I like to use the LLM at the beginning-

Sebastian Raschka 26:34 Mm

Lex Fridman 26:35 ... to lay out the full context of, like, what is this world that I'm now stepping into? But I try to avoid clicking out of the LLM into the world of, like, Twitter and blogs, and because then you're now down this rabbit hole, you're reading somebody's opinion. There's a flame war about a particular topic, and all of a sudden, you're no longer-- you're now in the, in the realm of the internet and Reddit and so on. But if you're purely l- letting the LLM give you the context of why this matters, what are the big picture ideas, uh, but sometimes books themselves are good at doing that, but not always, so.

Nathan Lambert 27:12 This is why I like the ChatGPT app. It gives the AI a home in your computer when you are fo- you can focus on it, rather than just being another tab in my mess of internet options. And I think Claude Code, and these particularly, does a good job of making that a joy, where it seems very engaging as a product designed to be an interface that your AI will then go out into the world. And it's something that is very kind of intangible between it and Codex, is that it just feels kind of warm and engaging, where Codex can often be as good from OpenAI, but it just kind of like feels a little bit rougher on the edges. Whereas, like, Claude Code is, makes it fun to build things, particularly from scratch, where you just don't-- like, you don't have to care, but you trust that it'll make something. Like, obviously, this is good for websites and kind of refreshing tooling and stuff like this, which I use it for, or data analysis. So I, my, my blog, we scrape Hugging Face, so we keep the download numbers for every dataset and model over time now, so we have them. And it's like Claude was just like: "Yeah, I've made use of that data, no problem." And I was like: "That would've taken me days." And I was like, then I have enough situational awareness to be like, okay, these trends obviously make sense, and you can check things. 'Cause that's just the kind of wonderful interface where you can have an intermediary and not have to do the kind of awful low-level work that you would have to do to maintain different web projects and do this stuff.

Lex Fridman 28:29 All right, so we just talked about a bunch of the closed-weight models. Let's talk about the open ones. Uh, so tell me about the landscape of open LLM models. Which are interesting ones, which stand out to you, and why? We already mentioned DeepSeek.

Nathan Lambert 28:44 Do you want to see how many we can name off the top of our head?

Lex Fridman 28:47 Yeah, yeah, without looking at notes.

Nathan Lambert 28:48 DeepSeek, Qimi, Minimax, Z.ai, Ant Ling. [chuckles] We're just going Chinese. [laughing]

Sebastian Raschka 28:57 [chuckles] Let's throw in Mistral AI, Gemma, um-

Nathan Lambert 29:00 Yeah

Sebastian Raschka 29:01 ... GPT OSS, the open-source model by, uh, ChatGPT. Actually, NVIDIA Nemotron had a-- or NVIDIA had a very cool one, a Nemotron-3. Um, there, there's a lot of stuff, especially at the end of the year. Qwen, one, maybe the one-

Nathan Lambert 29:12 Oh, yeah, Qwen was the na- the obvious name I was missing. I was trying to get through the... You can get at least ten Chinese and at least [chuckles] ten Western. I think that, I mean, OpenAI released their first open model-

Sebastian Raschka 29:21 Mm, a long time

Nathan Lambert 29:22 ... since GPT-2. That was-- when I, when I meant, talked, when I was writing about OpenAI's open model release, they're all like: "Don't forget about GPT-2," which I thought was really funny [chuckles] because it's just such a different time. But GPT OSS is actually a very strong model and does some things that the other models don't do very well. And I think that selfishly, I'll promote a bunch of, like, Western companies. So both in the US and Europe have these, like, fully open models. So I work at Allen Institute for AI, where we've been building Olmo, which releases data and code and all of this, and now we have actual competition for people that are trying to release everything so that other people can train these models. So there's the Institute for Foundation Models/LM 360, which is, like, had their K2 models of various types. Apertess is a Swiss research consortium. Hugging Face, um, has SmolLM, which is very popular, um, and NVIDIA's Nemotron has started releasing data as well. And then Stanford's Marin Community Project, which is kind of making it so there's a pipeline for people to open a GitHub issue and implement a new idea, and then have it run in a stable language modeling stack. So this space, that list was way smaller in 2024-

Lex Fridman 30:31 Mm

Nathan Lambert 30:31 ... so I think it was, like, just Ai2. So that's a great thing for more people to get involved in to understand language models, which doesn't really have a, like a Chinese company that is, has an analog. While I'm talking, I'll say that-... the Chinese open language models tend to be much bigger, and that gives them this higher peak performance as MoEs, where a lot of these things that we like a lot, whether it was Gemma, um, and Nemotron, have tended to be smaller models from the US, which is, which is starting to change from the US, US and Europe. Uh, Mistral Large 3 came out, which was a giant MoE model, very similar to DeepSeek architecture in December. And then a start-up, RCAI, and both Nemotron have-- Nemotron, it's NVIDIA, have teased MoE models of this way bigger than one hundred billion parameters-

Lex Fridman 31:16 Mm-hmm.

Nathan Lambert 31:16 -like this four hundred billion parameter range coming in this like Q1 2026 timeline. So I think this kind of balance is set to change this year in terms of what people are using the Chinese versus US open models for, which will be-- which I'm personally quite gonna be very excited to watch.

Lex Fridman 31:32 First of all, huge props for being [chuckles] able to name so many of these.

Nathan Lambert 31:36 [chuckles]

Lex Fridman 31:36 Did you actually name Llama?

Nathan Lambert 31:38 Um, no.

Lex Fridman 31:39 I feel like- [chuckles]

Nathan Lambert 31:40 RIP.

Sebastian Raschka 31:41 This was not on purpose. [chuckles]

Lex Fridman 31:43 RIP, Llama.

Nathan Lambert 31:44 Mm-hmm.

Lex Fridman 31:45 All right. Can you mention what are some interesting models that stand out? So you mentioned Qwen 3 is, is, is obviously a standout.

Sebastian Raschka 31:51 So I would say the year is almost bookended by both, uh, DeepSeek version three and R1, and then on the other hand, in December, uh, DeepSeek version three point two. Because what I like about those is they always have an interesting architecture tweak-

Lex Fridman 32:03 Mm-hmm

Sebastian Raschka 32:03 ... that others don't have. But otherwise, if you want to go with, um, you know, like the familiar but really good performance, Qwen 3, and like, um, Nathan said, also GPD OSS. And I think GPD OSS, what's interesting about it is kind of like the first public or like open-weight model that was really trained with tool use in mind, which I do think is kind of a little bit of a paradigm shift, where the ecosystem was not quite ready for it. So with tool use, I mean that the LLM is able to do a web search, to call a Python interpreter.

Lex Fridman 32:32 Mm-hmm.

Sebastian Raschka 32:33 And I do think this, uh, it's a standout because I think it's a huge unlock, because, um, one of the most, uh, common complaints about LLMs are, for example, hallucinations, right?

Lex Fridman 32:43 Mm-hmm.

Sebastian Raschka 32:43 And so, in my opinion, one of the best ways to solve, uh, hallucinations is to not try to always remember information or make things up. For math, why not use a calculator app or Python?

Lex Fridman 32:54 Mm-hmm.

Sebastian Raschka 32:54 If I ask the LLM, "Who won the, I don't know, Soccer World Cup in nineteen ninety-eight?" Uh, instead of just trying to memorize, it could go do a, a search. I think mostly it's usually still a Google search, so ChatGPT and GPD OSS, they would do a tool call to Google, maybe find the FIFA website, find, okay, it was France. It would get you that information re- reliably instead of just trying to memorize it. So I think it's a huge unlock, uh, which I think right now is not fully utilized yet by the open-source, open-weight ecosystem. A lot of people don't use tool call modes because I think it's, first, it's a trust thing. You don't want to run this on your computer where it has access to tools, could wipe your hard drive or whatever. So you want to maybe cont- containerize that. Um, but I do think, you know, that, that is, like, a really important step, um, for the upcoming years to have this, uh, ability. Yeah.

Lex Fridman 33:44 So, uh, a few quick things. First of all, thank you for defining what you mean by tool use. I think that's a great thing to do in general for the concepts we're talking about. Even things as sort of well-established as MoEs-

Sebastian Raschka 33:56 Mm-hmm

Lex Fridman 33:57 ... uh, you have to say that means Mixture of Experts.

Sebastian Raschka 34:00 Mm-hmm.

Lex Fridman 34:00 You kind of have to build up an intuition for people, what that means, how it's actually utilized, what are the different flavors. So what does it mean that there's just such explosion of open models? What's your intuition?

Nathan Lambert 34:13 If you're releasing an open model, you want people to use it, is the first and foremost thing, and then after that comes things like transparency and trust. I think when you look at China, the biggest reason is that they want people around the world to use these models, and I think a lot of people will not. If you look outside of the US, a lot of people will not pay for software, but they might have computing resources where you can put a model on it and run it. I think there can also be data that you don't want to send to the cloud. So this-- the, the number one thing is getting people to use models, use AI or use your AI, that might not be able to do it without having access to the model.

Lex Fridman 34:45 I guess we should state explicitly, so we've been talking about these Chinese models and open-weight models. Oftentimes, the way they're run is locally, so it's not like you're sending your data to China or to whoever developed, uh... to Silicon Valley, whoever developed the model.

Nathan Lambert 35:04 A lot of American start-ups make money by hosting-

Lex Fridman 35:06 Mm-hmm

Nathan Lambert 35:07 ... these models from China and selling them, selling token-- It's called like, selling tokens, which means somebody will call the model to do some se- some piece of work.

Lex Fridman 35:15 Mm-hmm.

Nathan Lambert 35:15 I think the other reason is for US companies, like, chat-- OpenAI is so GPU-deprived. [chuckles] Like, they're so-- they're at the limits of the GPUs. Whenever they make a release, they're always talking about, like, "Our GPUs are hurting." And I think there's like, [chuckles] like in one of these, like, GPT OSS release sessions, Sam Altman said, like, "Oh, we're releasing this because we can use your GPUs. We don't have to use-- [chuckles] we don't have to use our GPUs, and OpenAI can still get distribution out of this," which is a- another very real thing, because this doesn't cost them though, anything.

Sebastian Raschka 35:43 And for the user, I think also, I mean, there are users who just use the model locally, how they would use, uh, ChatGPT, but also for companies, uh, I think it's a huge unlock to have these models because you can customize them, you can train them, you can, uh, add post-training, con- uh, add more data, like specialize them into, let's say, law, medical models, whatever you have. And the appeal, you mentioned Llama, the appeal of the open-weight models from China is that the open-weight models are also-- the licenses are even friendlier. I think they are just unrestricted open-source licenses, where if you use something like, uh, Llama or Gemma, there are some strings attached. I think it's like a upper limit in terms of how many users you have, and then if you exceed, I don't know, so and so many million users, you have to report your finance, [chuckles] um, situation to, let's say, Meta or something like that. And I think, well, it is a free model, but there are strings attached, and people do like s- things where strings are not attached. So I think that's also one of the reasons, uh, besides performance, why the open-weight models from China are so popular, because you, you can just use them. There's no, there's no catch in that sense. [chuckles] Yeah.

Nathan Lambert 36:46 Mm-hmm. The ecosystem has gotten better on that front, but mostly downstream of these new providers providing such open licenses. That was funny when you pulled up Perplexity. It said, "Kimi K2 Thinking hosted in the US," which is just [chuckles] like an exact-- I've never seen this, but it's an exact example of what we're talking about, where people are sensitive to this.

Sebastian Raschka 37:01 [chuckles]

Nathan Lambert 37:01 Like Kimi K2 Thinking and Kimi K2 is a model that is very popular. People say that it has very good, like, creative writing and also in doing some software things. So it's just these little quirks that people pick up on with different models that they like.

Lex Fridman 37:14 ... uh, what are some interesting ideas that some of these models have explored that you can speak to, like, that are particularly interesting to you?

Sebastian Raschka 37:21 Maybe we can go chronologically. I mean, there was, of course, DeepSeek, um, DeepSeek R1 that came out in January, if we just focus on 2025. However, this was based on DeepSeek version 3, which came out the year, um, before in December 2024. There are multiple things on the architecture side. What is fascinating is you, you can still-- I mean, that's what I do in my from scratch coding projects, you can still start with GPT-2, and you get, can add things to that model to make it into this other model. So it, it's all still kind of like the same lineage, the same... It is a very close, uh, relationship between those. But off the top of my head, DeepSeek, what was, uh, unique there is the Mixture of Ex-- Uh, not, I mean, they were not inventing Mixture of Experts. We can maybe talk a bit more what Mixture of Experts means. Um, but just to list these things first before we dive into detail. Mixture of Experts, but then they also had a multi-head latent attention, which is a tweak to the attention mechanism, where this was, I would say, 2025, the main distinguishing factor be- between these, uh, open weight models. Different tweaks to make inference or KV cache size. We can also define KV cache in a few moments, but to kind of make it more economical, to have long context, to shrink the KV cache size. So what are tweaks, um, that we can do? And most of them focused on the attention mechanism. There is multi-head latent attention in, in DeepSeek. There is a group query attention, which is still very popular. It's not invented by any of those models. It goes back a few years, but that, that would be the other option. Sliding window attention, I think almost reuses it, um, if I remember correctly. So there are, like, like, these different tweaks that make the models different. Otherwise, um, I put them all together in an article [chuckles] once, where, um, I just compared them. They are very surprisingly similar. It's just different numbers in terms of how many repetitions of the transformer block you have in the center, uh, and, like, just little, little knobs that people tune. But, but what's so nice about it is it's, it, it works no matter what. You can tweak things, you can move the normalization layers around, you get some performance gains. And i- OLMO is always very good in ablation studies, showing what actually, uh, what it does to the model if you move something around. Ablation studies, does it make it better or worse? But there are so many, let's say, ways you can implement a transformer and make it still work. Big ideas, um, that are still prevalent is Mixture of Experts, multi-head latent attention, um, sliding window attention, group query attention. And then at the end of the year, we saw a focus on making the attention mechanism scale linearly with inference token prediction. So there were Qwen-3 Next, for example, which added a gated delta net. It's, it's like, um, kind of like inspired by, um, state-space models, where you have a fixed state that you keep updating, but it makes essentially this attention cheaper, or it replaces attention with a cheaper operation.

Lex Fridman 40:08 And then maybe is it useful to step back and talk about transformer architecture in general?

Sebastian Raschka 40:13 Yeah. So maybe we should start with the GPT-2 architecture, the transformer that was derived from the Attention Is All You Need paper.

Lex Fridman 40:21 Mm-hmm.

Sebastian Raschka 40:22 So the Attention, uh, Is All You Need paper had a transformer architecture that had two parts, an encoder and a decoder, and GPT went just focusing in on the decoder part. It is essentially, uh, still a neural network, um, and it has this attention mechanism inside, and you predict one token at a time. You pass it through an embedding layer. There's the transformer block. The transformer block has attention modules and a fully connected layer, and there are some normalization layers in between, but it's essentially neural network layers with this attention mechanism. So coming from GPT-2, uh, when we move on to GPT-OSS, there is, for example, the Mixture of Experts, um, layer. It's not invented by GPT-OSS, it's a few years old, um, but it is essentially a, a tweak to make the model larger without consuming more compute in each forward pass. So there is this, uh, fully connected layer, and if listeners are familiar with, um, multilayer perceptrons, you can think of a mini multilayer perceptron, a fully connected neural network layer inside the transformer, and it's very expensive because it's fully connected. If you have a thousand inputs, a thousand outputs, that's like a one million connections, and it's a very expensive part in this transformer. And the idea is to kind of expand that into multiple feedforward net- uh, networks. So instead of having one, let's say you have two hundred and fifty-six, but it would make it way more expensive because now you have two hundred and fifty-six, but you don't use all of them at the same time. So you now have a router that says, "Okay, based on this input token, it would be useful to use this, um, fully connected network." And in that context, it's called an expert. So a Mixture of Experts means you have multiple experts, and depending on what your input is, uh, let's say it's more math-heavy, it would use different experts compared to, let's say, translating input text from English to Spanish. It would maybe consult different experts. It's not quite clear, I mean, as clear-cut to say, "Okay, this is only an expert for math, and for Spanish," it's a bit more fuzzy. But the idea is essentially that you pack more knowledge into the network, but not all the knowledge is used all the time.

Lex Fridman 42:27 Mm-hmm.

Sebastian Raschka 42:27 That would be very wasteful. So you are kind of like, during the token generation, you are more selective. There's a router that selects which tokens should go to which expert. Adds more complexity, it's harder to train. There's a lot of, you know, that can go wrong, like collapse and everything. So I think that's why OLMO-3 still uses, uh, dense. I mean, you have, I think, OLMO models with Mixture of Experts, but dense models, where dense means... Uh, so also, it's jargon. There's a distinction between dense and sparse. So Mixture of Experts is considered sparse because we have a lot of experts, but only few of them are active, so that's called sparse. And then dense would be the opposite, where you only have, like, one fully connected module, and it's always, you know, utilized.

Lex Fridman 43:08 So may- maybe this is a good place to also talk about KV cache.

Sebastian Raschka 43:11 Mm-hmm.

Lex Fridman 43:11 But actually before that, even zooming out, like, fundamentally, how many new ideas-... have been implemented from, from GPT-2 to today?

Sebastian Raschka 43:22 Mm-hmm.

Lex Fridman 43:22 Like, how different really are these architectures?

Sebastian Raschka 43:25 Picture like the mixture of experts. Um, the attention mechanism in GPT-OSS, that would be the group query attention mechanism. So it's a slight tweak from multi-head attention to group query attention, so that we have two. I think they replaced, uh, LayerNorm by RMSNorm, but it's just like a different normalization layer. Not a big change, it's just like a, a tweak. Um, the nonlinear activation function, people familiar in- with deep ne- neural networks, I mean, it's the same as changing sigmoid with ReLU. It's, it's not changing [chuckles] the network fundamentally, it's just like a tweak, like a little, little tweak. Um, and, and that's about it, I would say. It's not really fundamentally that different. It's still the same, same architecture. So you can convert one from one, uh... You can go from one into the other by just adding these, these changes, basically.

Lex Fridman 44:09 It fundamentally is still the same architecture.

Sebastian Raschka 44:12 Mm-hmm. Yep. So for example, you mentioned my book earlier. That's a GPT-2 model in the book because it's simple and it's very small. Um, so 124, uh, 120 million parameters approximately. But in the bonus materials, I do have OMO 3 from scratch, Gemma 3 from scratch, and other types of from-scratch models. And I always started with my GPT-2 model and just, you know, tweaked a few or added different components, and you get from one to the other. It's like, it's kind of like a lineage in a sense. Yeah.

Lex Fridman 44:37 Can you build up an intuition for people? Because s- sort of when you zoom out, you look at it, there's so much rapid advancement in the AI world, and at the same time, fundamentally, the architectures have not changed.

Sebastian Raschka 44:50 Mm-hmm.

Lex Fridman 44:51 So where is all the turbulence, the turmoil of the advancement happening? Where, where's the gains to be had?

Sebastian Raschka 45:01 So there are the different stages where you develop the network, um, or train the network. You have the pre-training. Now, um, back in the day, it was just pre-training with GPT-2. Now you have pre-training, mid-training, and post-training. Um, so and I think right now we are in the post-training focus stage. I mean, pre-training still gives you, um, advantages if you scale it up to better, higher quality data. But then we have capability unlocks that were not there with GPT-2. For example, uh, ChatGPT, it is basically a GPT-3 model, and, and GPT-3 is the same as GPT-2 in terms of architecture. What was new was adding the, um, supervised fine-tuning and the reinforcement learning with human feedback. So it's more on the algorithmic side rather than the architecture.

Nathan Lambert 45:44 I would say that the systems also change a lot. I think if you listen to NVIDIA's announcements, they talk about these things like, "You now do FP8, you can now do FP4," and what is happening is these labs are figuring out how to utilize more compute to put it into one model, which lets them train faster, and that lets them put more data in, and then you can find better configurations faster by doing this. So you can look at, like, the-- essentially, the tokens per second per GPU is a metric that you look at when you're doing large-scale training, and you could get-- you can go from, like, ten K to thirteen K by turning on FP8 training, which means you're using less memory per parameter in the model, and by saving less information, you do less communication, you can train faster. So all of these, like, system things underpin way faster experimentation on data and algorithms that is kind of like-- it's this, it's this kind of loop that keeps going, [chuckles] wh- where it's kind of hard to describe when you look at the architecture, and they're exactly the same, but the code base used to train these models is gonna be vastly different.

Sebastian Raschka 46:44 Mm-hmm.

Nathan Lambert 46:44 And you could probably like, I don't... The GPUs are different, but you probably train GPT-OSS 20B way faster in wall clock time than GPT-2-

Sebastian Raschka 46:52 Mm-hmm

Nathan Lambert 46:53 ... was trained at the time.

Sebastian Raschka 46:54 Yeah, like you said, they had, for example, in the mixture of experts, this NV, uh, FP4 optimization, for example, where you get more throughput. But I, I do think this is for the speed, this is true, but, uh, it, it doesn't give the model new capabilities in a sense. It's just how much can we make, make the computation coarser without suffering in terms of model performance degradation? Um, but I do think, I mean, there are alternatives popping up to the transformer. There's text diffusion models, uh, completely different paradigm. Um, and there's also, I mean, though text diffusion models might use transformer architectures, but it's not an auto- autoregressive, um, transformer, and also Mamba models. Uh, it's a state-space model, but they do have trade-offs, and, uh, what's right is there's nothing that has replaced the autoregressive transformer as state-of-the-art model. So, like, for state of the art, you would still do that, go with that thing, but there are now alternatives for the cheaper and, like, alternatives that are kind of, um, making compromises, but i- it's not just one architecture anymore. There are little ones coming up, but if we talk about the state of the art, it's pretty much still the, the transformer architecture, autoregressive, derived from GPT-2,

Sebastian Raschka 48:05 essentially.

Lex Fridman 48:06 I guess the big question here is, we talked quite a bit here on the architecture behind the, the pre-training. Are the scaling laws holding strong across pre-training, post-training, inference, context size, data, synthetic data?

Nathan Lambert 48:20 I'd like to start with the technical definition of scaling law-

Lex Fridman 48:23 Yes

Nathan Lambert 48:23 ... which kind of informs all of this. The scaling law is a power law relationship between, you can think of the X-axis, so kind of what you are scaling is a combination of compute and data, which are kind of similar, and then the Y-axis is like the held-out prediction accuracy over our next tokens. We talk about models being autoregressive. It's like if you keep a set of m- text that the model has not seen, how accurate will it get when you will train? And the idea of scaling laws came when people figured out that that was a very predictable relationship, and I think that that technical term is continuing, and then the question is like, what do users get out of it? And then there are more types of scaling where, um, OpenAI's o1 was famous for introducing inference time scaling, and I think less famously for also showing that you can scale reinforcement learning training and get kind of this log X-axis and then a linear increase in performance on Y-axis. So there's kind of these three axes now where the traditional scaling laws are talk- talked about for pre-training, which is how big your model is and how big your dataset is, and then-... scaling reinforcement learning, which is like, how long can you do this trial-and-error learning that we will talk about? We'll define more of this, and then this inference time compute, which is just letting the model generate more tokens on a specific problem. So I'm kind of bullish where they, they're all really still working, but the low-hanging fruit has mostly been taken, especially in the last year, on, um, reinforcement learning with verifiable rewards, which is this RLVR, and then inference time scaling, which is just why these models feel so different to use, where previously you would get that first token immediately, and now they'll go off for seconds, minutes, or even hours, th- generating these hidden thoughts before giving you the first word of your than-- answer. And that's all about this inference time scaling, which is such a wonderful kind of step function in terms of how the models change abilities. They've kind of enabled this tool use stuff and enabled this much better software engineering that we were talking about. And this is, when we say enabled, almost entirely downstream of the fact that this reinforced learning with verifiable rewards training just kind of let the models pick up these skills very easily. So it let the models learn. So if you look at the reasoning process, when the models are generating a lot of tokens, what it'll be often doing is it tries a tool, it looks at what it gets back, it tries another API, it sees what it gets back, and if it solves the problem. So the models, when you're training them, very quickly learn to do this. And then at the end of the day, that gives this kind of general foundation where the model can use CLI commands very nicely in your repo and handle Git for you and move things around and organize things, or search to find more information, which if we're sitting in these chairs a year ago, is something that we didn't really think of the models being doing. So this is just kind of something that has happened this year and has totally transformed how we think of using AI, which I think is very magical. It's such an interesting evolution and just so-- unlocks so much value. But it's, it's like, it's not clear what the next avenue will be in terms of unlocking stuff like this. I think that there's, there's-- we'll get to continual learning later, but there's a lot of buzz around certain areas of AI, but no one knows when the next step function will, will really come.

Lex Fridman 51:31 So you've, you've actually said quite a lot of things there and said profound things quickly. It would be nice to unpack them a little bit. You say you're bullish basically on every version of scaling. So can we just even start at the beginning? Pre-training, are we kind of implying that the low-hanging fruit on pre-training scaling has been picked? Is the, is, has pre-training hit a plateau, or is even pre-training still you're bullish on?

Nathan Lambert 52:01 Pre-training has gotten extremely expensive. I think to scale up pre-training, it's also implying that you're gonna serve a very large model to the users. So I think that it's been loosely established, the likes of GPT-4 and similar models were around one trillion, like this order of trillion parameters at the biggest size. There's a lot of rumors that they've actually gotten smaller as training has gotten more efficient. Pe- you want to make the model smaller because then your costs of serving go down proportionally. These models, the cost of training them is really low relative to the cost of serving them to hundreds of millions of users. I think DeepSeek had this famous number of about five million dollars for pre-training at cloud market rates. I think Olmo Three, um, section two point four in the paper, we just detailed how long we had the GPU clusters sitting around for training, which includes engineering issues, multiple seeds, and it was, like, about two million dollars to rent the cluster to, like, deal with all the problems and headaches of training a model. So these models are pretty... like, a lot of people could get one to ten million dollars to train a model, but the recurring costs [chuckles] of serving millions of users is really billions of dollars of compute. I think that you can look at cl- like, a thousand GPU rental, you can pay a hundred grand a day for, and these companies could have millions of GPUs. [chuckles] Like, you can look at how much these things cost to sit around. So that's kind of a big thing, and then it's like, if scaling is actually giving you a better model, like, is it gonna be financially worth it? And I think it'll kind of slowly we'll push it out as AI solves more compelling tasks. So like the likes of Claude Opus four point five, making Claude code just work for things. I think I, I launched this project called, like, the Atom Project, which is like American Truly Open Models in July, and that was, like, a true vibe coded website. [chuckles] And, like, I have a job, um, make plots and stuff, and then I came back to refresh it in the last few weeks, and it's like Claude Opus four point five versus whatever model at the time was, like, just crushed all the issues that it had from building in June and July, and, like, it might be a bigger model. There's a lot of things that go into this, but that's like-- there's still progress coming.

Lex Fridman 54:04 So, so what you're speaking to is the nuance of the Y-axis of the scaling laws, that the, the way it's experienced versus on a benchmark, the actual intelligence is, might, might be different. But still, your intuition about pre-training, if you scale the, the size of compute, will the models get better? Not whether it's financially viable, but just from the law aspect of it, do you think the models will get smarter?

Nathan Lambert 54:28 Yeah, and I think that there's-- and this sometimes comes off as like, almost like disillusioned from people, leadership at AI companies saying this, but they're like: "It's held for thirteen orders of magnitude of computer something. Like, why would it ever end?" So I think fundamentally it is pretty unlikely to stop. It's just like eventually we're not even gonna be able to test the bigger scales because of all the problems that come with more compute. I think that there's a lot of talk on how twenty twenty-six is a year when very large Blackwell compute clusters, it's like gigawatt-scale facilities that hyperscalers are coming online, and these were all contracts for power and data centers that were signed and sought out in, like, twenty twenty-two and twenty twenty-three, so before or right after ChatGPT. So it took this two- to three-year lead time to build these bigger clusters to train the models, while there's obviously immense interest in building even more data centers than that. So that is, like, kind of the crux that people are saying is, like, these new clusters are coming, the labs are gonna have more compute for training, they're going to utilize this, but it's not a given. And it's like I...... I've seen so much progress that I expect it, and I expect a little bit bigger models, and I expect, um, I would say it's more like we will see a two thousand dollar subscription this year. We've seen two hundred dollar subscriptions. It's like that could ten X again, and these are the kind of things that could come, and they're all downstream of this, like, bit big- bit bigger model that offers just a little bit more cutting edge.

Lex Fridman 55:53 So, you know, it's reported that xAI is gonna hit that, uh, one gigawatt scale early '26 and full two gigawatt by year-end. W- how do you think they'll utilize that in the context of scaling laws? Is, is a lot of that inference, is a lot of that training?

Nathan Lambert 56:12 It ends up being all of the above. So I think that all of your decisions when you're training a model come back to pre-training. So if you're gonna scale RL in a model, you still need to decide on your architecture that enables this. We were talking about, like, other architectures and, uh, using different types of attention. We're also talking about mixture of experts models. The sparse nature of MoE models makes it much more efficient to do, um, generation, which becomes a big part of, um, post-training. And it's like you need to have your architecture ready so that you can actually scale up this compute. I still think most of the compute is going in at pre-training because you can still make a model better, you still want to go and revisit this. You still want the best base model that you can. And in a few years, that'll saturate, and the, the RL compute will just go longer.

Lex Fridman 57:00 Is there people who disagree with you that say, basically, pre-training is dead, it's all about scaling inference, scaling post-training, scaling context, continual learning, uh, scaling data, synthetic data?

Nathan Lambert 57:15 People vibe that way and describe it in that way, but I think it's not the practice that is happening.

Lex Fridman 57:19 It's just the general vibe of people saying this thing is dead.

Nathan Lambert 57:21 The excitement is elsewhere.

Lex Fridman 57:23 Yes.

Nathan Lambert 57:23 So the low-hanging fruit in RL is elsewhere. Like, for example, we released our model in November, for every company has deadlines. Our deadline was, like, November twentieth, and our, for that, our RL run was five days, which compared to twenty twenty-four, is a very long time to just be doing post-training at a model of, like, thirty billion parameters. It's not a big model. And then in December, we had another release, which was just-- we let the RL run for go, go for another three and a half weeks, and the model got notably better, so we release it. And, like, that's a big amount of time to just allocate to, like, something that is gonna be your, um, peak-

Lex Fridman 57:55 Mm-hmm

Nathan Lambert 57:55 ... for the year. [chuckles] So it's like-

Lex Fridman 57:57 So reasoning is-

Nathan Lambert 57:58 There's these types of decisions that happen when they're training a model where they just, like, can't, they can't leave it forever. You have to keep, you have to keep pulling in the improvements you have from your researchers. So that's like you redo pre-training. You'll do this post-training for a month, but then you need to give it to your users. You need to do safety testing. So there's kind of just, like... I think there's a lot in place that reinforces this cycle of just keep updating the models. There's things to improve. You get a new compute cluster that lets you do something maybe more stably or faster. It's like you hear a lot about Blackwell having rollout issues, where at Ai2, most of the models were pre-training around, like, one to two thousand GPUs. But when you're pre-training on ten thousand or a hundred thousand GPUs, you hit very different failures. So GPUs are known to break in weird ways, and doing a hundred thousand GPU run is like, you're pretty much guaranteed to always have at least one GPU that is down. And you need to have your training code handle that redundancy, which is just a very different problem.

Lex Fridman 58:51 Mm-hmm.

Nathan Lambert 58:51 Whereas, like, what we're doing, like, I'm playing with post-training on a DJX Spark, or you have your book, it's like... Or people learning ML, it's like what they're battling to train these biggest models is just, like, mass distributed scale, and it's very different. But that's somewhat different than, like, are these- like, that's a systems problem-

Lex Fridman 59:11 Mm-hmm

Nathan Lambert 59:11 ... in order to enable the scaling laws, especially at pre-training, you need all of these GPUs at once. When we shift to reinforcement learning, it actually lends itself to heterogeneous compute because you have many copies of the model. And, uh, to do a primer for a language model, reinforcement learning, what you're doing is you have two sets of GPUs. One is, you can call it the actor, and one you call the learner. The learner is where your actual reinforcement learning updates are gonna do. These are traditionally policy gradient algorithms, um, proximal policy optimization, PPO, and group relative po- policy optimization, GRPO, are the two popular classes. And on the other side, you're gonna have actors, which are generating completions, and these completions are the things that you're gonna grade. So reinforcement learning is all about optimizing reward. And in practice, what you can do is that you can have a lot of different actors in different parts of the world doing different types of problems, and then you send it back to this highly networked compute cluster to do this actual learning, where you, where you take the po- where you take the gradients, and you need to have a tightly meshed network where you can do different types of parallelism and spread out your model for efficient training. So there's just, like, a lot of-- every different type of training and serving has these considerations you need to scale. Like we talked about pre-training, we talked about RL, and then inference time scaling is like: how do you serve a model that's thinking for an hour to a hundred million users? I'm like, I don't really know about that, but I know that's a hard problem, and in order to give people this intelligence, there's all these systems problems, and we need more compute, and you need more stable compute to do it.

Lex Fridman 1:00:46 But you're bullish on all of these kinds of scaling, is what I'm hearing, on the inference, on the reasoning, even on the pre-training.

Sebastian Raschka 1:00:54 Yeah. So that's a, a big can of worms here, but so there are basically two... The knobs are the training and the, uh, inference scaling, where you can get gains. And so in an, in a world where we had, let's say, infinite compute resources, you want to do all of them. But, like, so you have training, you have inference scaling, and training is like a hierarchy. It's pre-training, mid-training, post-training. Changing the model size, more training data, making, training a bigger model gives you more knowledge in the model. Then the model, um, let's say, has a better-- it's like a better base model back in the day, or we still we call it foundation model, and it unlocks... So you un- but you don't, let's say, have the model be able to solve your most complex task-... tasks during pre-training or after pre-training, you still have these other unlock phases where you have mid-training or long context, for example, post-training with, uh, LR/VR that unlocks capabilities that the model has in terms of just knowledge in the pre-training. And I think, sure, if you, uh, so do more pre-training, you get a better base model that you can unlock later. But like Nathan said, it, it just becomes too expensive, so we don't have infinite compute. So you have to decide: Do I want to spend that compute more on making the model larger? But, you know, it's like a trade-off. It's, it's like an ideal world, you want to do all of them. And I think in that sense, scaling is still pretty much alive. You would still get a better model, but like we saw with GPT 4.5, it's just not worth it. I mean, it's like ... 'cause you can, let's say you can unlock more performance with other techniques at that current moment. Especially, um, if you look at inference scaling, that's one of the biggest gains this year with o1, um, where it took a smaller model further than pre-training a larger model like GPT 4.5. So it, it's like, I wouldn't say pre-training scaling is dead, it's just like there are other more attractive ways to scale right now at the moment. But at some point, you know, we'll- y- you will still wanna make some progress on the pre-training. The thing is also to consider, um, where you, where do you want to spend your money? If you spend it more on the pre-training, it's like a fixed cost. You train the model, and then it has this capability forever. You, you can always use it, uh, and so forth. With inference scaling, you don't spend money during training, you spend money later per query. And then it's also like the math, how long is my model gonna be on the market? If I replace it in half a year, maybe it's not worth spending $5 million, $10 million, $100 million on the training it longer. Maybe it's just, I will just do more inference scaling and get the performance from there. It maybe cost me $2 million in terms of user queries. It becomes a question of how many users you have and then doing the math. Um, and I think that's also where it's interesting, where ChatGPT is in a position, I think they have a lot of users, where they need to go a bit cheaper, where they have that, uh, GPT-5 model that is a bit smaller. Other companies that have ... Let's say if your customers have other, uh, other, um, trade-offs, for example, there was also the Math Olympiad or some of these, these math, uh, problems where ChatGPT or, or OpenAI, they had a proprietary model, and I'm pretty sure it's just like a model that has been maybe fine-tuned a little bit more, but most of it was during inference scaling to achieve this peak performance in certain tasks, where you don't need that all the time. And but yeah, long [chuckles] story short, I do think all of these, uh, pre-training, mid-training, post-training, inference scaling, they are all still things you want to do. It's just finding, at the moment, in this year, it's finding the right ratio that gives you the best bang for the buck, basically.

Lex Fridman 1:04:13 I think this might be a good place to define pre-training, mid-training, and post-training.

Sebastian Raschka 1:04:18 So pre-training is the classic training, uh, one next token prediction at a time. You have a big corpus of data, and, uh, Nathan also has very interesting insights there because of MoE-3. It's a big portion of the paper focuses on the right data mix. So pre-training is essentially just, you know, train across entropy loss, training on next token prediction on a, a vast corpus of internet data, books, papers, and so forth. It has changed a little bit over the years in the sense people used to throw in everything they can. Now, it's not just raw data, it's also synthetic data where people re- um, let's say, rephrase certain things. Uh, so synthetic data doesn't necessarily mean purely AI made-up, uh, data. It's also taking something from an article, Wikipedia article, and then rephrasing it as a Q&A question, or, um, summarizing it, rewording it, and, and making, uh, better data that way. 'Cause I think of it also like with humans, if someone, let's say, reads a book compared to a messy, I don't know, no offense, but like Reddit post or something like that. I do think you learn- [laughing] You ... No offense, uh, but I think-

Lex Fridman 1:05:25 There's gonna be a post about this-

Sebastian Raschka 1:05:26 Yeah [chuckles]

Lex Fridman 1:05:28 ... Sebastian.

Sebastian Raschka 1:05:28 Some Reddit data is very coveted and excellent for training.

1:05:31 Yep, yep.

1:05:31 You just have to filter it. [chuckles]

Lex Fridman 1:05:32 Yeah.

Sebastian Raschka 1:05:33 I, I think that's the idea. Uh, I, I think it, it's like if someone took that and rephrases that in a s- let's say, con- more concise and structured way-

Lex Fridman 1:05:42 Mm-hmm

Sebastian Raschka 1:05:43 ... I think it's higher quality data that gets the LLM maybe the same-- you get the same LLM out of it at the end, but it gets there faster, it trains faster because the ... Let's say, if the grammar and the punctuation is correct, it already learns the correct way, versus getting information from a messy way and then learning later how to correct that, and stuff like that. So I think that is how pre-training evolved and how, um, how still, while-- why scaling still works is that it's not about just the amount of data, it's also the tricks to make that data better for you, in a sense. And, and mid-training is, I mean, it used to be called, uh, pre-training. It's, I think it's called mid-training because it was awkward to have pre-training and post-training, but nothing in the middle, right? It sounds a bit weird to have pre-training and post-training, but what's the actual training? So the mid-training is usually similar [chuckles] to, uh, pre-training, but, you know, it's a bit more, I would say, specialized than pre-training. It's the same algorithm, but what you do is you focus, for example, on long context. Like, as one example, you have long context documents. The reason you don't do that during just pure pre-training is because you don't have that many long context documents, so you have a specific phase. And one problem of LLMs is also still, it's a neural network. It has the problem of catastrophic forgetting. So you teach it something, it forgets other things. And you wanna ... It's not 100% forgetting, but, you know, it's like no free lunch, you can't-- it's also the same with humans. If you ask me some math I learned 10 years ago, I don't know, [chuckles] I would have to look at it again.

Lex Fridman 1:07:09 Oh, Nathan was actually saying that he's consuming so much content that there's a catastrophic forgetting issue.

Sebastian Raschka 1:07:14 Yeah.

1:07:14 Yep.

1:07:14 I'm, like, trying to learn so much about AI, and it's like I was learning about pre-training parallelism, I'm like, I lost something and I don't know what it was. [laughing]

Lex Fridman 1:07:22 [laughing]

Sebastian Raschka 1:07:22 I don't want to anthropomorphize, uh, LLMs, but it's, I think, the same kind of in that sense, how humans learn. I mean-... the quantity is not always better because, yeah, you, it, it's like being selective. And I, and the mid-training is being selective in terms of quality content at the end. So the last thing the LLM has seen is the quality stuff. And then post-training is all the, uh, fine-tuning, supervised fine-tuning, uh, DPO, um, reinforcement learning with verifiable rewards, hu- uh, with human feedback, and so forth. So the refinement stages. And it's also interesting, it's like the cost thing, right? I mean, it's like pre-training, you spend a lot of money on that right now. RL, a bit less. RL, you don't really, I would say, teach it knowledge. It's more like unlocking the knowledge. Uh, it's, it's more like a skill learning, like how to solve problems with the knowledge that it has from pre-training. There are paper-- there are actually three papers [chuckles] this year, or last year, two thousand and twenty-five, on, uh, RL for pre-training, but I, I mean, I don't think anyone does that in production.

Nathan Lambert 1:08:17 Toy, toy examples for now. [chuckles]

Sebastian Raschka 1:08:18 Toy examples, right. But to generalize, RL, uh, post-training is more like the skill unlock, where pre-training is like soaking up the knowledge, essentially, yeah.

Nathan Lambert 1:08:26 A few things that could be helpful for people. A lot of people get, like, have-- think of synthetic data as being bad for training the models. You mentioned, like, the DeepSeek, got Olmo, uh, OCR, which is optical character recognition paper. A lot of labs did. Ai2 had one, uh, like, had multiple. And the reason that each of these labs has these is because there's vast amounts of PDFs and other digital documents on the web that are in formats that aren't encoded with text easily. So you use these Olmo-CR, these... or DeepSeek OCR, and we called ours Olmo-OCR, to extract what can be trillions of tokens of, um, candidate data for pre-training. And pre-training data set size is on the order of trillions, is measured in trillions of tokens. Smaller models from researchers can be something like five to ten trillion. Um, Qwen is documented going up to, like, fifty trillion, and there's rumors that these closed labs can go to, like, a hundred trillion tokens. And just getting this potential data to put in, I think they, they have a very big funnel, and then the data you actually train the model on is a small percentage of this. Like the syn-- this character recognition data would be described as synthetic data for pre-training in a lab. And then there's also the things like ChatGPT now gives wonderful answers, and you can train on those best answers, and that's synthetic data. It's very different than, like, early ChatGPT, lots of hallucinations, data when people became grounded in synthetic data.

Sebastian Raschka 1:09:48 One interesting question is, if I recall correctly, Olmo-III was trained with less data than, uh, specifically some o- other OpenWeight models, maybe even Olmo-II, but you still got better performance, and that might be one of the examples how the data helped.

Nathan Lambert 1:10:01 It's mostly down to data quality.

Sebastian Raschka 1:10:02 Mm-hmm.

Nathan Lambert 1:10:02 I think if we had more compute, we would train for longer. I think we'd ultimately see that as a, like, just, like, something we would want to do. And especially with big models, you need to have more compute because we talked about having more parameters, and we talked about knowledge. And essentially, there's a ratio where big models can absorb m- more from data, and then you're gonna-- you get more benefit out of this. It's, it's, it's like one of these, any logarithmic graph in your mind is like, a small model will level off sooner if you're measuring trillions of tokens, and bigger, bigger models need more. But mostly is we aren't training that big of models right now at Ai2, and getting the highest quality data we can is the natural starting point.

Lex Fridman 1:10:38 Is there something to be said, uh, about the topic of data quality? Is there some low-hanging fruit there still, where the quality could be improved?

Nathan Lambert 1:10:46 It's like turning the crank. So I think historically, in the open, there's been, like, a canonical best pre-training data set that has moved around between who has the most recent one or the best recent effort. Like Ai2's Dolmo was very early with the first Olmo, and Hugging Face had FinWeb, and there's a DCLM project, which has been kind of like a-- which is, it stands for Data Comp Language Model. There's been data comp for other machine learning projects, and they have a, had a very strong data set. And a lot of it is the Internet is becoming fairly closed off. So we have Common Crawl, which I think is hundreds of trillions of tokens, and you filter it, and it looks like being a lot of scientific work, where you're training classifiers and making decisions based on how do you prune down this dese-- this data set into the highest quality stuff and the stuff that suits your tasks. So previously, language models were tested a lot more on, like, knowledge and just kind of conversational things, but now they're expected to do math and code. So to train a reasoning model, you need to remix your whole data set. And there's a lot of actually wonderful scientific methods here where you can, you can, like, take your gigantic data set, you sample a lot of really tiny things from different sources. So you say you have GitHub, Stack Exchange, Reddit, Wikipedia. You can sample small things from them, and you train small models on each of these mixes and measure their performance on your evaluations. And you can just do, like, basic linear regression, and it's like, here's your optimal data set. But if your evaluations change, your data set changes a lot. So a lot of Olmo-III was new sources for reasoning to be better at math and code to... And then you do this mixing procedure, and it gives you the answer. [chuckles] And I think that's a lot of that's happened at labs this year. So there's new hot things, whether it's like coding environments or web navigation, and you just need to bring in new data. You need to change your whole pre-training so your post-training can work better and stuff like this. So that's, like, the constant re- re-evolution and the redetermining of what they care about as their, for their models.

Lex Fridman 1:12:35 Are there fun anecdotes of what sources of data are particularly high quality that we wouldn't expect? You mentioned Reddit sometimes can be a source.

Nathan Lambert 1:12:45 Reddit was very useful. I think that, um, like, PDFs is definitely one.

Sebastian Raschka 1:12:51 Especially archive.

Nathan Lambert 1:12:52 Yeah, so like Ai2 has run Semantic Scholar for a long time, which is a, um, like a, what you can say, is a competitor to Google Scholar with a lot more features. And to do this, Ai2 has found and scraped a lot of PDFs for openly accessible papers that might not be, um, like, behind the closed paid garden of a certain publisher. So, like, truly open scientific, scientific PDFs, and if you, like, you sit on all of these and you process it, and you can get value out of it. And I think that, like, a lot of that style of work has been done by the frontier labs much earlier, and it's just like you need to have a pretty skilled researcher that understands how things change models, and they bring it in, and they clean it, and it's, that's a lot of labor that... Like, I think of a lot of frontier labs, when they scale researchers, a lot more goes into data-... you have people, [chuckles] like, if you wanna be, if you join a frontier lab and you wanna have impact, the best way to do it is just make, find new data that's better. [chuckles] And then, like, the fancy, glamorous, algorithmic things, like figuring out how to make o1 is, like, the sexiest thought of a scientist. Of like, "Oh, I figured out to scale RL," and there's a group that did that, but I think in most of the contributions, it's like-

Lex Fridman 1:13:58 On the data set.

Nathan Lambert 1:13:58 - I'm gonna make the data better, or I'm gonna make the infrastructure better so that everybody in my team can run experiments 5% faster.

Sebastian Raschka 1:14:04 At the same time, I think it's also one of the closest guarded secrets, what your training data is for legal reasons. [chuckles] And so there's also, I think, a lot of work that goes into hiding what your training data was, essentially. Like, trying the model to not give away the sources because, yeah, of legal reasons.

Nathan Lambert 1:14:19 The other thing, to be complete, is that some people are trying to train on only licensed data-

Sebastian Raschka 1:14:24 Mm

Nathan Lambert 1:14:24 ... where Common Crawl is a scrape of, like, the whole internet. So, um, if I, I host multiple websites, I'm happy to have them train language models, but I'm not explicitly licensing-

Sebastian Raschka 1:14:34 Mm

Nathan Lambert 1:14:35 ... what governs it, and therefore, this li- like, Common Crawl is largely unlicensed, which means that your consent really hasn't been provided for how to use the data. There's another idea where you can train language models only on data that has been licensed explicitly, so that the kind of governing contract is provided. And I'm not sure if Apertess is the copyright thing or the license thing.

Sebastian Raschka 1:14:53 Yeah.

Nathan Lambert 1:14:53 I know that the reason that they did it was for an EU compliance thing, where they wanted to make sure that their model, um, fit one of those checks.

Sebastian Raschka 1:15:02 Mm. And also on that note, also, for example, there's also the distinction between, um, the, the licensing. So some people, like you said, they just purchase the license. Let's say they buy a book online, let's say an Amazon Kindle book or let's say a Manning book or something, and then use that in the training data. And that is, like, the gray zone because you paid for the content, and you might wanna train it. But then there are also restrictions where even that shouldn't be allowed, and so that, that is, like, where, where it gets a bit fuzzy. And yeah, I think that is, right now, still a hot topic. And also, big companies like OpenAI, they approached private companies for their proprietary data. And private companies, they become more and more, let's say, uh, protective of their data because they know, "Okay, this is gonna be my moat in, in a few years." And I do think, um, that's, like, the interesting question where if LLMs become more commoditized, and I think a lot of people learn about LLMs, there will be a lot more people able to train LLMs. Of course, there are infrastructure challenges, but if you think of big industries like, uh, pharmaceutical industries, law, finance industries, I do think they, at some point, will hire people from other, uh, frontier labs to build their in-house models on their proprietary data, which will be then, again, another unlock with pre-training that is currently not there. Because even if you wanted to, you can't get that data. You can't get access to clinical trials most of the time and these types of things. So I, I do think scaling in that sense might be still pretty much alive if you also look in domain-specific applications. Because we are still, right now, in this year, just looking at general-purpose LLMs on, on ChatGPT, Anthropic, and so forth. They are just general purpose. They're not even, I think, scratching the surface of what an LLM can do if it is really specifically trained and designed for a specific task.

Nathan Lambert 1:16:47 I think on the data thing, some-- this is one of the things where, like, this happened in 2025, and we totally forget it, is Anthropic lost in court and was owed [chuckles] $1.5 billion to authors. Anthropic, I think, bought thousands of books and scanned them and was cleared legally for that because they bought the books, and that is kind of going through the system. And then the other side, they also torrented some books, and I think this torrenting was the path where the court said that they were then culpable to pay this billions of dollars to authors, which is just, like, such a mind-boggling [chuckles] lawsuit that kind of just came and went. Like, that is so much money [chuckles] from-

Sebastian Raschka 1:17:20 Yeah

Nathan Lambert 1:17:20 ... the VC ecosystem. [chuckles]

Lex Fridman 1:17:22 These are court cases that will define the future of human civilization 'cause it, it's clearly that data drives a lot of this.

Nathan Lambert 1:17:27 Mm.

Lex Fridman 1:17:27 And there's this very complicated human tension of... I mean, you can empathize. You're both authors.

Sebastian Raschka 1:17:33 Yeah, yeah.

Lex Fridman 1:17:34 Yeah, there's some degree to which, I mean, you put your heart and soul and your, your sweat and tears into the, the writing that you do. Uh, it feels a little bit like theft-

Nathan Lambert 1:17:45 Mm

Lex Fridman 1:17:46 ... for somebody to train your data without giving you credit.

Sebastian Raschka 1:17:49 And there are, like Nathan said, also two layers to it. Someone might buy the book and then train on it, which is, uh, could be argued fair or not fair, but then there are literally straight-up, um, companies who use pirated books, where it's not even compensating the author. It's that-- that is, I think, where people got a bit, uh, angry about it specifically.

Lex Fridman 1:18:06 Yeah, but there has to be some kind of compensation scheme. This is like-

Sebastian Raschka 1:18:09 Mm, yep

Lex Fridman 1:18:09 ... moving towards, towards something like Spotify streaming did-

Sebastian Raschka 1:18:13 Mm-hmm

Lex Fridman 1:18:13 ... originally for music.

Sebastian Raschka 1:18:14 Mm-hmm.

Lex Fridman 1:18:14 You know, what does that compensation look like? You have to define those kinds of models. You have to think through all of that. Uh, one other thing I think people are generally curious about, I'd love to get your thoughts. As LLMs are used more and more, if you look at even Archive, but GitHub, more and more of the data is generated by LLMs. What do you do in that kind of world? Is- how big of a problem is that?

Nathan Lambert 1:18:38 Largest problem is infrastructure and systems, but from an AI point of view, it's kind of inevitable.

Lex Fridman 1:18:45 So it's basically LLM-generated data that's curated by humans, essentially, right?

Nathan Lambert 1:18:49 Yes, and I think that a lot of open source contributors are legitimately burning out. If you have a popular open source repo, somebody's like, "Oh, I wanna do open source AI, it's good for my career," and they just vibe code something, [chuckles] and they throw it into the p-

Sebastian Raschka 1:19:03 I, I have-

Nathan Lambert 1:19:03 You might get more of this than I do. [chuckles]

Sebastian Raschka 1:19:05 Yeah. So I have a actually a case study here. [chuckles]

Lex Fridman 1:19:08 Yeah.

Sebastian Raschka 1:19:08 Um, I have a, a repository called ML Extend, uh, that I developed as a student, around 15 years, 10 years ago, and it is a reasonably popular library still for certain algorithms, I think, especially like frequent data mining stuff. And there was recently, I think, two or three people who submitted a lot of PRs in a very [chuckles] short amount of time. I do think LLMs have been involved in submitting these PRs. Me, as the maintainer, there are two things. First, I'm a bit [chuckles] overwhelmed. Like, I don't have time to read through it because especially it's an older library, that is not a priority for me. At the same time, I kind of also appreciate it because I think something people forget is it's not just using the LLM. There's still a human-- you have a human layer that verifies something, and, uh, and that is, in a sense, also how data is labeled, right? So that's, like, um, one of the most expensive things is get- getting labeled data for RL, like, in human feedback, uh, phases. And, and this is kind of like that, where-... it goes through phases, and then you get actually higher quality data out of it, you know? So I, I don't mind it in a sense. It, it can feel overwhelming, but I do think there is also value in it.

Lex Fridman 1:20:11 It feels like there's a fundamental difference between raw LLM-generated data and LLM-generated data with human in the loop-

Sebastian Raschka 1:20:19 Mm-hmm

Lex Fridman 1:20:19 ... that does some kind of verification, even if that verification is a small percent-

Sebastian Raschka 1:20:23 Mm-hmm

Lex Fridman 1:20:23 ... of the lines of code.

Sebastian Raschka 1:20:25 Mm-hmm. I think this goes with anything, um, like where the people think also sometimes, oh, I can just use an LLM to learn about XYZ, which is true, you can, but there might be a person who is an expert who might have used an LLM to write specific code. There is kind of like this human work that went into it to make it nice and throwing out the not-so-nice part, to make it, to kind of like pre-digest it for you, and that saves you time. And I think that's, uh, it's, that's the value add, where you have someone filtering things or even using the LLMs correctly. I think this is still labor that, that you get for free. If you, for example, read an article, let's say a Substack article, I could maybe ask an LLM to give me opinions on that, but I wouldn't even maybe know what to ask. I think there is still value in reading that article compared to me going to the LLM because you are the expert, you select what knowledge is actually spot on, should be included, and you give me this very... this, this, um, executive summary. And, and this is kind of a huge value add because now I don't have to waste three, five hours to go through this myself, maybe get some incorrect information, and so on. And so I think that's also where the future still is for writers, even though there are, uh, LLMs, that expert can kind of, like, save you time.

Lex Fridman 1:21:43 It's kind of fascinating to actually watch, and I'm sure you guys do this, but, uh, for me to look at the difference between a summary and the original content, even if it's a page-long summary of a page-long content. It's interesting to see how the summary, LLM-based summary takes the edge off. Like, what, what is the signal it removes from the thing?

Nathan Lambert 1:22:07 The voice is what I talk about a lot.

Lex Fridman 1:22:09 Voice. Well, voice, I, I would love to hear what you mean by voice. That's really powerful. But sometimes there's, like, literally insights.

Nathan Lambert 1:22:15 Yeah. [chuckles]

Lex Fridman 1:22:16 Like, in removing an insight, you're actually fundamentally changing the meaning of the thing. So I'm continuously disappointed how bad LLMs are at really getting to the core insights, which is what a great summary does. Yet, even if you go, and I have these extensive, extremely elaborate prompts where I'm, like, really trying to dig for the insights, [chuckles] and it's still not quite there, which, um, I mean, that's a whole deep philosophical question about what is human knowledge and wisdom, and what does it mean to be insightful, and so on. But when you talk about the voice, what do you mean?

Nathan Lambert 1:22:52 So when I write, I think a lot of what I'm trying to do is take what you think as a researcher, which is very raw, which a researcher is trying to encapsulate an idea at the frontier of their understanding, and they're trying to put what is a feeling into words. And I think that my writing, I try to do this as the writing, which makes it come across as raw, but also high information in, in a way that it's like, some people will get it and some won't, and that's kind of the nature of research.

Lex Fridman 1:23:18 Mm-hmm.

Nathan Lambert 1:23:19 And I think this is something that language models don't do well. Particularly, they're all trained with this reinforcement learning from human feedback, which is designed to take feedback from a lot of people and, in a way, average how the model behaves from this.

Lex Fridman 1:23:30 Mm-hmm.

Nathan Lambert 1:23:30 And I think that there's-- it's going to be hard for a model to be very incisive when there's that sort of filter in it. And I think this is kind of a wonderful fundamental problem for researchers in RLHF, is like, this provides so much utility in making the models better, but also the problem formulation is kind of like there, there's this knot in it that you can't get past. So that's what I think of as like, these language models don't have this prior in their deep expression that they're trying to get at. I don't think it, it's impossible to do. I think there's stories of models that really shock people. Like, I think of, like, I would love to have tried Bing Sydney, and does like, does that have more voice? 'Cause it would so often go off the rails on people, and in fact, in what is historically obviously a scary way. Like, telling a reporter to leave his wife is a crazy model to potentially put in general, [chuckles] general adoption. But that's kind of like a trade-off, like, is this RLHF process like, in some ways adding limitations?

Lex Fridman 1:24:28 That's a terrifying place to be as one of these frontier labs and, and companies, because millions of people are using them.

Nathan Lambert 1:24:35 There was a lot of backlash last year with the GPT-4.0 getting removed, and I personally never used the model, but I've talked to people at OpenAI where they're to the point where they, like, get emails from users that might be detecting subtle differences in the deployments in the middle of the night, and they email them, and they're like: My friend is different. And they, like, find these people, employees' emails and send them things because they are so attached to this set, what is a set of model weights and a configuration that is deployed to the users. We see this with TikTok. You open it-- I, I don't use TikTok. Supposedly, in, like, five minutes, the mo- the algorithm gets you. It's like, it's locked in, and I don't-- like, those are language models doing recommendations. Like, I think there are ways that you could do this with a language model. Within, like, five minutes of chatting with it, the model just gets you, and that is something that people aren't really ready for. Like, I think that, like, kid- like, don't give that to kids. [chuckles]

Lex Fridman 1:25:27 Yeah.

Nathan Lambert 1:25:27 Like, don't give that to kids, at least until we know what's happening.

Lex Fridman 1:25:30 But there, there's also going to be this mechanism. What's going to happen with these LLMs is they're used more and more. Uh, unfortunately, the nature of the human condition is such that people commit suicide. And so what journalists would do is they will report extensively on the people who commit suicide, and they would very likely link it to the LLMs because they have that data about the conversations. If you're really struggling in your life, if you're depressed, if you're thinking about suicide, you're going to probably talk to LLMs about it. And so what journalists will do is they will say, "Well, the suicide was committed because of the LLM," and that's going to lead to the companies-... because of legal issues and so on, more and more and more taking the edge off of the LLM. So it's going to be as generic as possible. It's so difficult to operate in this space because of course, you don't want an LLM to cause harm to humans at that level. But also, this is also the nature of the human experience, is to have a rich conversation, a fulfilling conversation, and one that challenges you, from which you grow. You need that edge, and that, that, that's something extremely difficult for AI researchers on the RLHF front to actually have to solve, 'cause you're actually s- dealing with the human condition.

Nathan Lambert 1:26:47 Like, a lot of researchers at these companies are so well-motivated, and there's definitely the likes of Anthropic and OpenAI are culturally so want to do good through this for the world. And there is-- it's such a, like, I'm like, "Ooh, I don't wanna work on this," because on the one hand, a lot of people see AI as a health ally, as somebody they can talk to about their health confidentially, but then it bleeds all the way into this, like, talking about mental health and things, where that's, it's heartbreaking that this will push-- like, be the thing where somebody goes over the edge, but other people might be saved. And I'm like, I don't... Like, there's things that, as a researcher training models, it's like I don't wanna train image generation models and release them openly, 'cause I don't wanna enable somebody to have a tool on their laptop that can harm other people. Like, I don't have the infrastructure at my company to do that safely. But it's like, like there's a lot of areas like this, where it's just, it needs people that will approach it with the complexity and just kind of conviction of like... It's just such a hard problem.

Lex Fridman 1:27:47 But also, we as a society, as users of these technologies, need to make sure that we're having the complicated conversation about it versus just fear-mongering. Big Tech is, is causing harm to humans or stealing your data, all that kind of stuff. There-- it's, it's more complicated than that. And you're right, there's a very large number of people inside these companies, many of which you know, many of which I know, that deeply care about helping people. They are considering the full human experience of people from across the world, not just Silicon Valley, people across the United States, people across the world, what that means, what their needs are. It's really difficult to design this one system that is able to help all these different kinds of people across the different age groups, cultures-

Nathan Lambert 1:28:27 Mm

Lex Fridman 1:28:27 ... mental states, mental conditions, all that kind of stuff.

Nathan Lambert 1:28:31 I wish that the timing of AI was different with the relationship of Big Tech to the average person. So, like, Big Tech's reputation was so low, and with how AI is so expensive, it's like inevitably gonna be a Big Tech thing, where it takes so many resources, and people say the US is, quote, unquote, "betting the economy on AI with this build-out." And it's like to have these be intertwined at the same time is just makes for such a hard communication environment. It would be good for me to go talk to more people in the world that hate Big Tech and see AI as a continuation of this.

Lex Fridman 1:29:02 And one of the things you actually recommend, one of the antidotes that you, you talk about is, uh, to find agency in this whole system, as opposed to sort of sitting back in a powerless way and consuming the AI slop as it quickly, rapidly takes over the internet. More find agency by using it to build stuff, build apps, build... So you-- one, that actually helps you build the intuition, but two, it's empowering because you, you're gonna understand how it works, what the weaknesses are, and allows-- it gives your voice power to say, like: "This is fucked up. This is bad. This is bad use of the technology, and this is good use of the technology." And you're more plugged into the system then, so you can understand it better, and you can steer it better as a, as a consumer.

Sebastian Raschka 1:29:48 I think that's a good point you brought up, agency. Instead of ignoring it and saying, "Okay, I'm not gonna use it," I think it's probably long-term healthier to say, "Okay, it's out there. I can't put it back." You know, like internet, computers back then, when they came out. How do I make best use of it, and how does it help me to uplevel myself? The one thing I worry here, though, is like if you just fully use it for something you love to do, the, the thing you love to do is not, not, no longer there, and that could potentially, I feel, like, lead to burnout. For example, if I use an LLM to do all my coding for me, now there's no coding. I'm just managing something that is coding for me. Two years, let's say, later, if I just do that eight hours a day, have something code for me, do I feel fulfilled still? Like, is this like... Yeah, I mean, is this, like, hurting me in terms of being excited about my job, excited about what I'm doing? Am I still proud to build something?

Lex Fridman 1:30:43 So there's, uh, on that topic of enjoyment, it's, it's quite interesting. We should just throw this in there, that there's this recent survey of about seven hundred and ninety-one professional developers, professional meaning ten-plus years of experience.

Nathan Lambert 1:30:55 That's a long time. [chuckles]

Lex Fridman 1:30:57 Yeah, but [laughing]

Nathan Lambert 1:30:58 As a junior developer. [laughing]

Lex Fridman 1:31:01 Uh, yeah, in this day and age. Uh, so the, the results here on many fronts are, uh, surprising. So they break it down by junior and senior developers, but, I mean, it just shows that both junior and senior developers u- use AI-generated code in code they ship. So this is not just for fun, sort of intermediate kind of learning things. This is code they ship, and so it's twenty-five percent men-- like, most of them use around fifty percent or more, and what's interesting is, for the category of over fifty percent of your code that you ship as AI-generated, senior developers are much more likely to do so. But you don't want AI to take away the thing you love.

Nathan Lambert 1:31:45 Yep.

Lex Fridman 1:31:46 I think this speaks to my experience, these particular results I'm about to say. So together, about eighty percent of people find it either somewhat more enjoyable or significantly more enjoyable to use AI as part of the work.

Sebastian Raschka 1:31:59 I think it depends on the task. Um, for my personal usage, for example, I have a website where I sometimes tweak things on the website. I personally don't enjoy this. [chuckles] So in that s- sense, if the AI can help me to implement something on my website, I'm all here for it. It's, it's great.... but then at the same time, when I solve a complex problem, well, if there's a bug and I hunt this bug, and I find the bug, it's the best feeling in the world. It's like you get so much joy, like, oh, it's like you feel like, great. But now, if you don't even think about thinking about the bug, you just go directly to the LLM, well, you never have this kind of feeling, right? But then there could be the middle ground, where, well, you try yourself, you can't find it, you use the LLM, and then you don't get frustrated because it helps you, and you move on to something that you enjoy. And so I think looking at these statistics, I think also the difference is or w- what is not factored in, it's averaging over all the different scenarios where we don't-- So we don't know if it's for the core task or if it's for something mundane that people would not have enjoyed, uh, otherwise. So in a sense, AI is really great for doing mundane things that, um, take a lot of work. Um, so for example, my wife the other day, uh, she has, like, a podcast for, like, book, uh, like, book discussions, a book club, [chuckles] and she was, like, transferring the show notes from, um, Spotify to YouTube, and then the links somehow broke. Uh, and she had in some episodes, because it discussed so many books, like hundred links or something, and it would have been really painful to go in there and fix each link manually. And so I suggested, "Hey, let's try ChatGPT." We copied the text into ChatGPT, and it fixed them. And in- instead of two hours going from link to link, fixing that, you know, it made that type of work much more seamless. There was no frustration. Fixed. I think everyone has a use case where AI is useful for something like that, that would be really boring, really mundane.

Lex Fridman 1:33:51 I, for me personally, since we're talking about coding, uh, and you mentioned debugging, uh, what-- A lot of the source of the enjoyment for me on, more on the cursor side than the Claude code side, is the... I have a friend, I have a co, what's that called? A pair programmer. Like, uh, it's less l- lonely. You, you made debugging sound like this great joy. [laughing]

Sebastian Raschka 1:34:15 Yeah.

Lex Fridman 1:34:15 No, I would say, I would say debugging is like a drink of water after you've been going through a desert-

Sebastian Raschka 1:34:21 Yep

Lex Fridman 1:34:21 ... for, for, for, for days. [chuckles] So, like, you, you skip the whole desert part where you're suffering. So, like, there-- sometimes it's nice to have a friend who's, who can't really find the bug but can give you some intuition about the code, and you're together-

Sebastian Raschka 1:34:36 Mm-hmm

Lex Fridman 1:34:37 ... uh, with that friend, going through the desert, and then together find that drink of water. So at least for me, uh, maybe it speaks to the loneliness of the programming experience. It's-- That is a source of joy.

Sebastian Raschka 1:34:48 It's maybe also related to delayed gratification. I'm a person who, you know, [chuckles] even as a kid, I liked the idea of Christmas presents, having them, getting them, better than actually getting the presents. I would look forward to the day I get the presents, but then it's over, and I'm disappointed. And maybe it's something like also with, let's say, food. I think food tastes better when you're really hungry. Uh, and with-- Yeah, you're right, with debugging, it is not always, you know, uh, great. It's, like, often frustrating, but, um, then if you can solve it, then it's great. But there's also, like, a sweet Goldilocks zone, where if it's too hard, then it's, you know, wasting your time. But I think that is another challenge, though. How will people learn? I mean, the chart we looked at, um, we saw that more senior developers are shipping more AI-generated code than the junior ones. And I think it's very interesting because intuitively, you would think it's the junior developers because they don't know, let's say, how to do the thing yet because they are more junior, and so they use AI to do that thing. It could either mean the AI is not good enough yet to solve that task, but it could also mean experts are more effective at using it. They know where and better how to use it and review the code, and they trust the code then more. And so I think one issue in the society in the future will be, though, how do you become an expert if you never try to do the thing yourself? And I think one way, it's always like, for me, how, how I learn is by trying things myself. Like math textbooks, if you look at the solutions, yeah, you learn something, but I think you learn actually better if you try first, and then you appreciate the solution differently because you know how to put it into your mental framework. And, um, if LLMs are here all the time, would you actually go through the length at struggling? Would you be willing to struggle? Because struggle is not nice, right? I mean, it's struggling, and if you use the LLM to do everything, at some point, you will never really take the next step, and then you will maybe not get that unlock that you would get as an expert using an LLM. So it's like, you know, it's like... I think there's, like, a Goldilocks sweet spot, where maybe, maybe the trick here is you make dedicated offline time, where you study two hours a day, and the rest of the day, use LLMs. But I, I think it's important also for people to still invest in themselves, in my opinion, to not just, you know, LLM everything.

Lex Fridman 1:37:10 Yeah, there is, and, uh, we together as a civilization, that we each individually have to find that Goldilocks zone-

Sebastian Raschka 1:37:16 Yeah

Lex Fridman 1:37:16 ... uh, and in the programming context

Lex Fridman 1:37:18 as developers. Now, we've had this fascinating conversation that started with pre-training and mid-training. Let's get to post-training. A lot of fun stuff in post-training. So what are some of the interesting ideas in post-training?

Nathan Lambert 1:37:31 The biggest one from 2025 is learning this reinforcement learning with verifiable rewards. You can scale up the training there, which means doing a lot of this kind of iterative generate grade loop, and that lets the models learn both interesting behaviors on the tool use and software side. This could be searching, running commands on their own, and seeing the outputs, and then also that training enables this inference time scaling very nicely. And it just turned out that this paradigm was very nicely linked in this, where it's this kind of RL training enables inference time scaling, but inference time scaling could have been found in different ways. So it's kind of this perfect storm of the models change a lot, and the way that they're trained is a major factor in doing so. And this has changed how people approach post-training.... dramatically.

Lex Fridman 1:38:20 Can you describe RLVR, popularized by DeepSeek R1? Can you describe how it works?

Nathan Lambert 1:38:25 Yeah, fun fact, um, I was on the team that came up with the term RLVR, which is from our two to three work before DeepSeek, which is-- we don't take a lot of credit for the, being the people to popularize the scaling RL. But as fun as what academics get as an aside, is the ability to name and influence-

Lex Fridman 1:38:44 Mm

Nathan Lambert 1:38:44 - the discourse, because the closed labs can only say so much. That one of the things you can do as an academic is like, you might not have the compute to train the, the model, but you can frame things in a way that ends up being... I describe it as like a community can come together around this RLVR term, which is very fun. And then DeepSeek is the people that did the training breakthrough, which is they scaled the reinforcement learning, which was, you have the model generate answers and then grade the completion if it was right, and then that accuracy is your reward for reinforcement learning. So reinforcement learning is classically an agent that acts in an environment, and the environment gives it a state and, and a reward back, and you try to maximize this reward. In the case of language models, the reward is normally accuracy on a set of verifiable tasks, whether it's math problems, coding tasks, and it starts to get blurry with things like factual domains. Like, that is also in some ways verifiable or constraints on your instruction, like respond only with sen-- words that start with A. Like, all of these things are verifiable in some way, and the core idea of this is you find a lot more of these problems that are verifiable, and you let the model try it many times while taking these RL steps, these RL gradient updates. The infrastructure evolved from this reinforced learning from human feedback, where in that era, the score they were trying to optimize was a learned reward model of aggregate human preferences. So you kind of change the problem domains, and that let the optimization go on to much bigger scales, which kind of kickstarted a major change in what the models can do and how people use them.

Lex Fridman 1:40:24 What kind of domains is, uh, RLVR amenable to?

Nathan Lambert 1:40:28 Math and code are the famous ones, and then there's a lot of work kind of on s- what is called the rubrics, which is related to a word people might have heard as LM, as a judge, which is like, for each problem, I'll have a set of problems in my training data set. I'll then have another language model and ask it: "What would a good answer to this problem look like?" And then you could try the problem a bunch of times over and over again and assign a score based on this rubric. So that's not necessarily verifiable, like a math and code domain, but this rubric's idea and other scientific problems that it might be a little bit more vague, is where a lot of the attention is, where they're trying to push this set of methods into these kind of, uh, more open-ended domains, so the models can learn a lot more.

Sebastian Raschka 1:41:11 I think that's called reinforcement learning with AI feedback, right?

Nathan Lambert 1:41:14 That's the older term from it that was coined in Anthropic's constitutional AI paper. [chuckles] So it's like a lot of these things come in cycles.

Sebastian Raschka 1:41:21 Also, just one step back for the RLVR. So I think the interesting, beautiful thing here is that you j-- you ask the LLM, let's say, a math question, and then you know the correct answer, and you let the LLM, like you said, figure it out. But how it does it, I mean, you don't really constrain it much. There are some constraints you can add, like use the same language. Don't switch between Spanish and English. But let's say you're pretty much hands-off. You only give the question and the answer, and then the LLM has to, you know, just the task to arrive at the, uh, right answer. But the beautiful thing here is, what happens in practice is that the LLM will do a step-by-step description. Like, you know, like as a student or like as a, yeah, m- mathematician, how you would derive the solution. It will give you, or it will use those steps, and that helps actually the model to improve its own accuracy. And then, like you said, the inference scaling. So inference scaling loosely means basically spending more compute during, using the LLM, during inference. And here, the inference scaling is that the model would use more tokens. And, and also, I think in the R1 paper, they showed the longer they train the model, the longer the responses are. They, they grow over time, they use more tokens, so it becomes more expensive, becomes more expensive for simple tasks. But these explanations, they help the model with the accuracy. There are also interesting [chuckles] lot, uh, lot of papers showing what the model explains does not necessarily have to be correct, or maybe it's even unrelated to the answer, but for some reason, it still helps the model. Like, this is the fact that it is, um, explaining. And I think it's also, again, I don't want to anthropomorphize these LLMs, but it's kind of like how we humans operate, right? If there's a complex math problem, let's say in a math, uh, class, you, you usually have a note paper, and you do it step by step, you cross out things, and the model also self-corrects. And that, that was, I think, the aha moment in the R1 paper. They called it aha moment because the model itself recognized it made a mistake and then said, "Ah, I did something wrong, and so let me try." And I think that's just so cool that this falls out of just giving it the correct answer and having it figure out how to do it, that it kind of does, in a sense, what a human would do. Although LLMs don't think like humans, [chuckles] it's kind of like an interesting coincidence. And, and the, the other side-- nice side effect is it's great for us humans often to see these steps. It builds trust, but also we learn, we can double-check things.

Nathan Lambert 1:43:40 There's a lot in here.

Sebastian Raschka 1:43:40 Mm.

Nathan Lambert 1:43:40 I think some of the debate-- There's been a lot of debate this year on if the language models like these aha... I think the aha moments are kind of fake because in pre-training, you essentially have seen the whole Internet.

Sebastian Raschka 1:43:51 Yep.

Nathan Lambert 1:43:51 So you have definitely seen people explaining their work, even, even verbally, like a transcript of a math lecture. You try this, "Oh, I messed this up." And what reinforcement learning is, this RLVR is very good at doing, is amplifying these behaviors-

Sebastian Raschka 1:44:03 Mm-hmm

Nathan Lambert 1:44:03 ... because they're very useful in enabling the model to think longer and to check its work. And-... I agree that it is very beautiful that this training kind of, the model learns to amplify this in a way that is just so useful at the final answers being better.

Sebastian Raschka 1:44:16 I can give you also a hands-on example. I was training the GPT-3 base model with RLVR on MATH five hundred. The base model had an accuracy of about fifteen percent. Just fifty steps, like in a few minutes, with RLVR, the model went from fifteen percent to fifty percent accuracy. And the model, you can't tell me it's learning anything about, fundamentally about math in-

Nathan Lambert 1:44:38 The Qwen example is weird because there's been two papers this year, one of which I was on, that talks about data contamination in Qwen-

Sebastian Raschka 1:44:44 Mm-hmm, mm-hmm

Nathan Lambert 1:44:44 ... and specifically that they trained on a lot of this special mid-training phase that we did-

Sebastian Raschka 1:44:48 Exactly

Nathan Lambert 1:44:48 -like a minute on, but it's weird since they train on problems that are almost identical to that.

Sebastian Raschka 1:44:53 Exactly. And so you can see that basically the RL, it's not teaching the model any new knowledge about math. You can't do that in fifty steps. So the knowledge is already there in the pre-training, you're just unlocking it.

Nathan Lambert 1:45:03 I still disagree with the kind of premise, because there's a lot of weird complexities that you can't prove. Because one of the things that points to weirdness is that if you take the Qwen-three so-called base model, and you can, you could Google on the screen, you could Google, like, math dataset hugging face, and you could take a problem. And what you do if you put it into Qwen-three base, the-- all these math problems have words. So it'd be like, Alice has five apples and takes one and gives three to whoever, and they're these word problems. With these Qwen-based models, why people are suspicious of them is if you change the numbers but keep the words-

Sebastian Raschka 1:45:36 Mm.

Nathan Lambert 1:45:36 -Qwen will produce like a very high de... wi-without tools, will produce a very high accuracy, like decimal representation-

Sebastian Raschka 1:45:43 Mm-hmm, mm-hmm

Nathan Lambert 1:45:43 ... of the answer. Which means there's some like, at some time, it was shown problems that were almost identical to the test set, and it was using tools to get a very high precision answer. But a language model without tools will never actually have this. So it's kind of been this big debate in the research community is like how much of these reinforcement learning papers that are training on Qwen and measuring specifically on this like math benchmark, where there's been multiple papers talking about contamination, it's like, how much can you believe them? And I think this is what caused the reputation of RLVR being about formatting, because you can get these gains so quickly, and therefore, it must already be in the model. But there's a lot of complexity here that we-- it's not really like controlled experimentation.

Sebastian Raschka 1:46:25 Yep.

Nathan Lambert 1:46:25 So we don't really know. [chuckles]

Sebastian Raschka 1:46:26 But, uh, if it weren't, weren't true, um, I would say distillation wouldn't work, right? I mean, distillation can work to some extent, but the, the thing is, that is, I think, the biggest problem in LLM research, this contamination, because we don't know what's in the data. It's-- unless you have wr- a new data set, it, it's really impossible. And the same, you mentioned, um, math, uh, the math data set, which is give a question and an answer and an explanation is given. But then also even something simpler like, uh, MMLU, which is a multiple-choice, um, benchmark. If you just change the format slightly, um, like, I don't know, you use a pre-- a dot instead of a parenthesis or something like that, the model accuracy will vastly differ.

Nathan Lambert 1:47:04 I think that that could be like a model issue rather than a general issue.

Sebastian Raschka 1:47:09 It's not even malicious by the developers of the LLM, like, "Hey, we want to cheat at that benchmark." It's just-- it has seen something at some point, and I think the only fair way to evaluate an LLM is to have a new benchmark that is after the cut-off date when the LLM was deployed.

Lex Fridman 1:47:22 Can we lay out what would be the sort of the recipe of all the things that would be going to post-training? And you mentioned our RLVR was a really exciting, effective thing.

Sebastian Raschka 1:47:32 Mm-hmm.

Lex Fridman 1:47:32 Maybe we should elaborate. RLHF still has a really important component to play. What kind of other ideas are there on post-training?

Nathan Lambert 1:47:40 I think you can kind of take this in order. I think you could view it as what made o1, which was this first reasoning model possible, or what will the latest model be, and they actually have... You're gonna have similar interventions at these, where you start with mid-training, and the thing that is rumored to enable o1 and similar models is really careful data curation, where you're providing a broad set of like, what is called reasoning traces, which is just the model generating words in a forward process that is reflecting, like breaking down a problem into intermediate steps and trying to solve them. So at mid-training, you need to have data that is similar to this to make it so that when you move into post-training, primarily with this verifiable rewards, it can learn. And then what is happening today is you're figuring out which problems to give the model and how long you can train it for, and like, how much inference you can enable the model to use when solving these verifiable problems. So as models get better, certain problems are no longer-- like, the model will solve them a hundred percent of the time, and therefore, there's very little signal in this. If we pull-- if we look at the GRPO equation, this one is famous for this because essentially, the reward given to the agent is based on how good a given action, a action is a completion, is relative to the other answers to that same problem. So if all the problems get the same answer, there's no signal in these types of algorithms. So what they're doing is they're finding harder problems, which is why you hear about things like scientific domains, which is like, that's so hard, like getting anything right there. If you have a lab or something, it just generates so many tokens or much harder software problems. So the frontier models are all pushing into these harder domains, and they can train on more problems, and the model will learn more skills at once. The RLHF link to this is kind of like RLHF has been and still is kind of like the finishing touch on the models, where it makes the models more useful by improving the organization or style or tone. There's different things that resonates to different audiences. Like, some people like a really quirky model, and RLHF could be good at enabling that personality. And some people hate this like markdown bulleted list thing that the models do, but it's actually really good for quickly parsing information. In RLHF, this human feedback stage is really great for just give-- putting this into the model at the end of the day. So it's what it made ChatGPT so magical for people, and that use has actually remained fairly stable. This formatting can also help-... the models get better at math problems, for example. So it's like the border between style and formatting and like the method that you use to answer a problem is actually, uh, they're all very closely linked in terms of when you're training these models. Which is why RLHF can still say, make a model better at math, but these verifiable domains are a much more direct process to doing this, because it's kind of makes more sense with the problem formulation. Which is why it kind of ends up all forming together. But to summarize, it's like mid-training is give the model the skills it needs to then learn. RL and verifiable rewards is let the model try a lot of times. So put a lot of compute into trial-and-error learning across hard problems, and then RLHF would be like, finish the model, make it easy to use, and kind of just round the model out.

Lex Fridman 1:51:02 Can you comment on the amount of compute required for RLVR?

Nathan Lambert 1:51:06 It's only gotten up and up, so I think Grok four was famous for saying they used a similar amount of compute for pre-training and post-training. Back to the scaling discussion, they involve very different hardware for scaling. Pre-training is very compute-bound, which is like this FLOPS discussion, which is just how many matrix multiplications can you get through in one time? And because RL, you're generating these answers, you're trying the model in the real-world environments, it ends up being much more memory-bound because you're generating long sequences, and the attention mechanisms have this behavior where you get a quadratic increase in memory as you're getting to longer sequences. So the compute becomes very different. So you-- when in pre-training, we would talk about a model. I think if we go back to, like, the Biden administration executive order, it's like ten to the twenty-fifth FLOPS to train a model. If you're using FLOPS in post-training, it's a lot weirder because the reality is just like, how many hours are you allocating, how many GPUs for? And I think in terms of time, the RL compute is getting much closer because you just can't put it all into one system. Like pre-training is so computationally dense, where all the GPUs are talking to each other, and it's extremely efficient, where RL has all these moving parts, and it can just take a long time to generate a sequence of a hundred thousand tokens. It could, like, if you think about GPT 5.2 Pro taking an hour, it's like, what if your training run [chuckles] has a sample for an hour, and you have to make it so that's handled efficiently? So I think in GPU hours or just like wall clock hours, the RL runs are probably approaching the number of days as pre-training, but they probably aren't using as many GPUs at the same time. There's rules of thumb where in labs, it's like you don't want your pre-training runs to last more than like a month because they fail catastrophically. And if you are planning a huge cluster to be held for two months and then it fails on day fifty, the opportunity costs are just so big. So you kind of don't want to just... People don't want to put all their eggs in one basket, which is like GPT-4 was like the ultimate YOLO run, and nobody ever wanted to do it before, where it took, like, three months to train, and everybody was shocked that it worked, where I think people are a little bit more cautious and incremental now.

Sebastian Raschka 1:53:07 So RLVR is more, let's say, un- unlimited, how much you can train or get still benefit, where RLHF, because it's a preference tuning, it-- you reach a certain point where it doesn't really make sense to spend more RL budget on that. So just a step back with, um, preference tuning. So there are multiple people that can give multiple, let's say, explanations for the same thing, and they can both be correct, but at some point, you learn a certain style, and it doesn't make sense to, you know, iterate on it. My [chuckles] favorite example is like if relatives ask me what laptop they should buy, I give them an explanation or ask them like: "Yeah, what is your, um, use case?" Like they, for example, prioritize battery life and storage. Other people like us, for example, we would prioritize RAM and, uh, compute. And so, but both, both answers are correct, but different people require different answers. And with preference tuning, well, you're trying to average somehow, like you are asking the data labelers to give you the right-- or not the right, the preferred answer, and then you train on that. But at some point, yeah, you learn that average preferred answer, and, and there's no, I think, reason to keep training longer on it because, you know, it's just a style. Where with RLVR, you literally give the model... Well, you let the lo- model solve more and more complex, difficult problems. And so I think that it makes more sense to allocate more budget long term to LRVR. And also that right now we are in, uh, LRVR one point O land, where it's still like that simple thing, where we have a question and answer, but we don't do anything with the one stuff in between. So there are also, I mean, multiple research papers, also by Google, for example, on process reward models that also give scores for the explanation, how correct is the ex- explanation? And I think that will be the next thing, let's say, RLVR two point O for this year. Focusing in between question and answer, like how to leverage that information, the explanation, to improve the explanation and help it to get better accuracy. But then, uh, so that, that's one angle. And there was a Deep Seek math version two paper where they also had interesting, uh, inference scaling there, where... Well, first they had, um, developed models that grade themselves, a separate model. And I think that that will be one aspect, and the other, like Nathan mentioned, it will be for R- LRVR branching into other domains.

Nathan Lambert 1:55:23 The place where people are excited are value functions-

Sebastian Raschka 1:55:26 Mm-hmm.

Nathan Lambert 1:55:26 -which is very, pretty similar. So process reward models are kind of like... Process reward models assign how good something is to each kind of intermediate step in a reasoning process, where value functions apply value to every token the language model generates. Both of these have been largely unproven in the language modeling and this reasoning model era. People are more optimistic about value functions forever, for whatever reason now. I think process reward models were tried a lot more in this pre-'01, pre-reasoning model era, and a lot of people had a lot of headaches with them. So I think a lot of it is the human nature of, like, value models have a very deep history in reinforcement learning. They're one of the first things that were core to, like, deep reinforcement learning existing, is like training value models in this. So right now, the literature people are excited about trying value models, but there's very little proof in it, and there are negative examples in trying to scale up process reward models.

Sebastian Raschka 1:56:21 Mm-hmm.

Nathan Lambert 1:56:22 These things don't always hold in the future. I think-... we came to this discussion by talking about scaling, and a simple way to summarize what you're saying with, like, you don't want to do too much RLHF, which is essentially the signal scales, is people have worked on RLHF for language models for years, especially in intense interest after ChatGPT. And this, the first release of a reasoning model trained with RLVR, OpenAI's o1, had a scaling plot where if you increase the training compute logarithmically, you get a linear increase in evaluations, and this has been reproduced multiple times. I think DeepSeek had a plot like this. But there's no scaling law for RLHF, where if you log increase the compute, you get some performance. In fact, the seminal scaling paper for RLHF is scaling laws for reward model over optimization. [chuckles] So it's like that's a big line to draw with RLVR and the methods we have now and in the future. Like, they will follow the scaling paradigm, which is like the best runs you can let to run for an extra ten X and you get a few X performance, but you can't do this with RLHF, and that is just going to be field defining in how people approach them, where I'm a shill for people academically to do RLHF, and that's a good way to describe it, is like, to do the best RLHF, you might not need the extra ten or a hundred X of compute, but to do the best RLVR, you do. So I think there's a, what I say is a seminal paper from what was a meta internship. It's called, it's like the art of scaling reinforcement learning with language models. They're-- what they describe as a framework is scale RL, and their incremental experiment was like ten thousand V200 hours, which is like thousands or tens of thousands of dollars per experiment, and they [chuckles] do a lot of them, which is just like, this cost is not accessible to the average academic, which is a hard equilibrium where it's trying to figure out how to learn from each community.

Lex Fridman 1:58:11 I was wondering if we could take at this point a bit of a tangent and talk about education and learning. If you're somebody listening to this who's a smart person, interested in programming, interested in AI, so I presume building something from scratch is a good beginning. So can you just take me through, like, what you would recommend people do?

Sebastian Raschka 1:58:32 So I would personally start, like you said, uh, implementing a simple model from scratch that you can run on your computer. The goal is not if you build a model from scratch, to have like something you use every day for your personal projects. Like, it's not going to be your personal assistant replacing an existing open weight model or ChatGPT. It's to see what exactly goes into the LLM, what exactly comes out of the LLM, how the pre-training works in that sense, uh, on your own computer, preferably. Um, and then you learn about the pre-training, the supervised fine-tuning, the attention mechanism. You get a solid understanding of how things work, but at some point you will reach a limit because small models can only do so much. And the problem with learning about LLMs at scale is, I would say it's exponentially more complex to make a larger model because it's not that the model just becomes larger, you have to now think about sharding your parameters across multiple GPUs. Even for the KV cache, there are multiple ways you can implement it. One is just to understand how it works, just to grow the cache. That's-- It's like a cache you grow step by step by, let's say, concatenating lists, um, growing it, but then that wouldn't be optimal. In GPUs, you wouldn't do that. You would pre-allocate a tensor and then fill it in, but that adds, again, another twenty, thirty line, lines of code, and for each thing, you add so much code. And I think the trick with the book is basically to understand how the LLM works. It's not gonna be your production level LLM, but once you have that, you can understand the production level LLM.

Lex Fridman 1:59:56 So you're trying to always build an LLM that's gonna fit on one GPU?

Sebastian Raschka 2:00:00 Yes. The-- Most of them I have, they-- I have some bonus materials on some, uh, MoE models. I think one or, or two of them, they may require multiple GPUs, but the goal is to have it on one GPU. And the beautiful thing is also you can self-verify. [chuckles] It's almost like RLVR when you code these from scratch. You can take, uh, an existing model from the Hugging Face transformer library. Um, so the Hugging Face transformer library is great, but if you want to learn about LLMs, I think that's not the best place to start because the code is so complex, because it has to full-- it has to fit so many use cases. Also, some people use it in production. It has to be really sophisticated, and it's really intertwined and really hard. It's not linear to read.

Nathan Lambert 2:00:39 It was started as a fine-tuning library, and then it grew to be like the standard representation of every model architecture and the way it is loaded. So Hugging Face is like the default place to get a model, and Transformers is the software that enables it-

Sebastian Raschka 2:00:52 Mm-hmm.

Nathan Lambert 2:00:52 -so people can easily load a model-

Sebastian Raschka 2:00:55 Mm

Nathan Lambert 2:00:55 ... and do something basic with it.

Sebastian Raschka 2:00:56 And all frontier labs that have open weight models have a Hugging Face Transformers version of it, like from DeepSeek to GPT-OSS. That's like the canonical weight, uh, that you can load there. But again, also even Transformers, the library is not used in production. People use then SGLang or vLLM, and it adds another layer of complexity.

Lex Fridman 2:01:15 We should say that the Transformers Library has, like, four hundred models.

Sebastian Raschka 2:01:19 So it's a one library that tries to implement a lot of LLMs, and so you have a huge code base. Basically, it's, like, huge. It's like, uh, it's, I don't know, maybe millions-

Nathan Lambert 2:01:30 It's crazy [chuckles]

Sebastian Raschka 2:01:30 ... hundreds of thousands of lines of code, and find-- It's like understanding the part that you want to understand is finding the needle in the haystack. But what's beautiful about it is y- you have a working implementation, and so you can work backwards from it. What I would recommend doing, or what I also do, is if I want to understand, for example, how Almost 3 is implemented, I would, uh, look at the weights in the model hub, the config file, and then you can see, oh, they use so many layers. They use, let's say, group query attention or multi-head attention in that case. Then you see all the components in, like, a human-readable, I don't know, hundred lines of config file. And then you start, let's say, with your GPT 2 model and add these things, you know? And the cool thing here is you can then load the pre-trained weights and see if they work in your model. And you want to match the same output that you get with a transformer model, and then you can use that as a, basically as a verifiable reward to make your architecture correct. And then it's kind of-- Sometimes it takes me a day to [chuckles] with Almost 3, the, the challenge was, uh, rope for the position embeddings. They had a yarn, uh, extension, and there was some custom, uh, scaling there, and I couldn't quite match the, these things. And in this struggle, you kind of understand things, but the cool thing is, at the end, you know you have it correct because you can unit test it, you can check against the reference implementation, and I think that's maybe one of the best ways to learn, really, like to basically reverse engineer something. Yeah.

Nathan Lambert 2:02:51 ... I think that that is something that everybody that's interested in getting to AI today should do. And I think that's why I liked your book, is like I came to language models from this RL and robotics field. Like I never had taken the time to just, like, learn all the fundamentals. And this transformer architecture I describe as being like so fundamental as like deep learning was a thing that I had to learn in the past, and people need to do this. I think that where a lot of people kind of get overwhelmed is: how do I apply this to have impact or find like a career path? Because like AI and language models make this fundamental stuff so accessible, and people with motivation will learn it. And then it's like, how do I get the cycles on goal to contribute to research? And I think that I'm actually fairly optimistic in this because the field moves so fast that a lot of times the best people, like, don't fully solve a problem because there's a bigger, lower hanging, [chuckles] like a bigger problem to solve that's very low-hanging fruit, so they move on. And I think that a lot of what I was trying to do in this RLHF book is like, take post-training techniques and just describe how people think about them influencing the model and what people are doing. And then it's remarkable how many things I just think are just like, people stop studying them or don't. So I think people trying to get narrow after doing the fundamentals is good, and then reading the relevant papers and being engaged in the ecosystem, it's like you actually, the proximity that random people have online from [chuckles] the leading researchers, like no one knows who all the... The anonymous account on X and ML is very popular for whatever reason, and no one knows who all these people are. Like, it could just be random people that study the stuff deeply, especially with the AI tools, and just be like, "Keep-- I don't understand this, keep digging into it," I think is a very useful thing. But there's a lot of research areas that just like are maybe three papers that you need to read, [chuckles] and then one of the authors will probably email you back. But you have to put in a lot of effort into these emails to understand the field. Like, I think it would be, for a newcomer, easily weeks of work to feel like they can truly grasp, like, what is a very narrow ara-- area. But I think going narrow after you have the fundamentals will be very useful to people because it's like I became very interested in, like, character training, which is like how you make the model funny or sarcastic or serious, and like, what do you do to the data to do this? And it's like a student at Oxford reached out to me and was like: "Hey, I'm interested in this," and I advised him. And I was like: "That paper now exists." And it's like, I don't know, there's like two or three people in the world that were very interested in this. He's a PhD student, which gives you an advantage, but like, for me, that was a topic I was waiting for someone to be like: "Hey, I have time to spend cycles on this," and I'm sure there's a lot more very narrow things where you're just like: "Oh, it doesn't make sense that there was no answer to this." And I think that it's just like there's so much information coming that people are like, "I can't grab on to any of these," but if you just actually stick in an area, I think there's a lot of interesting things to learn.

Sebastian Raschka 2:05:48 Yeah, I think you can't try to do it all because it would be very overwhelming, and you would burn out if you tried to keep up with everything. For me, for example, I haven't kept up with computer vision in a long time, just focused on LLMs. But coming back to your book, for example, I think this is also a really great book and a really good bang for the buck because you want to learn about RLHF. I wouldn't go out there and read RLHF papers because I would be-- you would be spending two years-

Nathan Lambert 2:06:10 Some of them contradict.

Sebastian Raschka 2:06:11 Yeah.

Nathan Lambert 2:06:11 There's-- I've just edited the book, and I was like, there's a chapter where I had to be like, "X papers say one thing, and X papers say another thing, and we'll see what comes out to be true." [chuckles]

Lex Fridman 2:06:21 What, what are some of the-- just to go through some of the table of contents, some of the ideas we might have missed in the bigger picture of the post-training. So first of all, you do the problem setup, training overview, what are preferences, preferences data and the optimization tools, reward modeling, regularization, instruction tuning, rejection sampling, reinforcement learning, i.e., policy gradients, direct alignment algorithms. Then constitutional AI and AI feedback, reasoning and inference time scaling, tool use and function calling, synthetic data and distillation, evaluation, and then open question section over optimization, style and information, and then product UX, character, and post-training. So what are some ideas worth mentioning that connect both the educational component and the research component? You mentioned the character training, which is pretty interesting.

Nathan Lambert 2:07:08 Character training is interesting because there's so little out of it, but we talked about how people engage with these models and like, like we feel good using them because they're positive, but that can go too far. It could be too positive, and it's like essentially it's how do you change your data and, or decision making to make it exactly what you want? And I-- OpenAI has this thing called a model spec, which is essentially their internal guideline for what they want the model to do, and they publish this to developers. So essentially, you can know what is a failure of OpenAI's training, which is like they have the intentions, and they haven't met it yet, versus what is something that they like actually wanted to do and that you don't like. And that transparency is very nice. But all the methods for curating these documents and how easy it is to follow them is not very well known. I think the way the book is designed is that the reinforcement learning chapter is obviously what people want because everybody hears about it with RLVR, and it's the same algorithms [chuckles] and the same math, but it's just like you can use it in, in very different documents. So I think the core of prefe-- of RLHF is like how messy preferences are, is essentially a rehash of a paper that I wrote years ago. But this is essentially the chapter that will tell you why RLHF is never, ever fully solvable. Because, like, the way that even RL is set up is that, um, it assumes that preferences can be quantified and that multiple preferences can be reduced to single values. And I think it relates in the economics literature to the Von Neumann-Morgenstern utility theorem. And like, that is the chapter where all of that philosophical, economic, and like psychological context, it tells you what gets compressed into doing RLHF. So it's like you have all of this, and then at the-- later in the book, it's like you use this RL math to make the number go up. And I think that that's why I think it would be very rewarding for people to do research on, is because it's like quantifying preferences is something that is just like-... humans have designed the problem in order to make preferences studiable. But there's kind of fundamental debates on, like, an example is in a language model response, you have different things you care about, whether it's accuracy or in style, and when you're collecting the data, they all get compressed into like a, "I like this more than another." And it's like, like that is happening. And there's a lot of philoso- [chuckles] there's a lot of research in other areas of the world that go into like, how should you actually do this? I think social choice theory is the subfield of economics around how you should aggregate preferences. [chuckles] And there's like, uh, I was-- I went to a workshop that published a white paper on like, how can you think about using social choice theory for RLHF? So I mostly would want people that get excited about the math to come and have things where they can stumble into and learn this kind of broader context. I think there's a fun thing. I just keep a list of all the tech reports that I like of reasoning models. So in the, in chapter fourteen, which is kind of like a short summary of RLVR, there's just like a gigantic table where I just, like, list every single reasoning model that I like. So there's just like, I think in education, a lot of it needs to be like... At this point, it's like what I like. Mm-hmm. Because the language models are so good at the math, where it's like famous paper, direct preference optimization, which is like a much simpler way of pro- solving the problem than RL. Um, the derivations in the appendix skip steps of math, and it's like I tried for this book. Like, I redid the derivations, and I'm like: What the heck is this log trick that they use to change the math? But doing it with language models, they're like: "This is the log trick." And I'm like, I don't know if I like this, that the math is so commoditized. I think, like, some of the struggle in reading this appendix- Mm-hmm ... and following the math, I think is good for learning. And I, [chuckles] - Yeah, so we're actually returning to this often just on the topic of education. You bo- both have brought up the word struggle- Mm-hmm ... quite a bit. So there is value. If you're not struggling as part of this process, you're not fully following the, the proper process for learning, I suppose. Some of the providers are starting to work on models for education, which are designed to not give-- actually, I haven't used them, but I would guess they're designed to not give all the information at once- Right ... and make people work to do this. So I think you could train models to do this, and it would be a wonderful contribution where, like all of this stuff in the book, you have to reevaluate every decision for it- Yeah ... which is such a great example. I, I think there's, there's a chance you work on an Ai2, which I, which I was like: "Oh, I think this is, be so fun." Makes sense. I do something like that, uh, did that the other day for video games, for example. [chuckles] I sometimes, for my pastime, play video games. Like, I like, uh, video games with, uh, puzzles, you know, like Zelda and Metroid, and there's this new game where I got stuck, and I really got stuck, and I was okay. I, you know, I don't want to struggle f- uh, like, two, uh, two days. And so I used an LLM. But then you, uh, say: "Hey, please don't, uh, add any spoilers, just, you know, I'm here and there. What do I have to do next?" And the same thing you can do, I guess, for math, where you say: "Okay, I'm here at this point. I'm getting stuck. Don't give me the full solution, but w- what is something I could try?" You know, like where you kind of carefully probe it. But the problem here is, I think it requires discipline, and a lot of people do math for like... I mean, there are a lot of people who enjoy math, but there are also a lot of people who need to do it for their homework. And then it's like the shortcut. And yeah, we can develop an educational LLM, but the other LLM is still there, and there's still a temptation to use the other LLMs. I think a lot of people, especially in college, they, they understand the stuff they're passionate about. Mm-hmm. They're self-aware about it, and they understand it shouldn't be easy. Mm-hmm. Mm-hmm. Like, I think you just have to develop a good taste. Mm-hmm. We talk about research taste, like school taste, about stuff that you should be struggling on and, and stuff you shouldn't be struggling on. Which is tricky to know, 'cause sometimes you don't have good, um, long-term vision about what would be actually useful to you in your career. But you have to, you have to d- develop that taste. Yeah. I was talking to maybe my fiancée or friends about this, and it's like there's this brief ten-year window where all of the homework and all the exams could be digital. But before that, everybody had to do all the exams in blue book because there was no other way. And now, after AI, everybody's gonna need to be in blue books and oral exams because everybody could cheat so easily. [chuckles] It's like this brief generation that had a different education system that, like, everything could be digital and, but you still couldn't cheat, and now it's just gonna go back. [chuckles] And it's just very funny. You mentioned character training. Just zooming out on, on a more general topic. For that topic, how much compute was required? And in general, to contribute as a researcher, are there places where not too much compute is required, where you can actually contribute as an individual researcher? For-- on the character training thing, I think this research is built on fine-tuning about seven billion parameter models with LoRA, which is like a, essentially, you only fine-tune a small subset of the weights of the model. I don't know exactly how many GPU hours that would take- But it's doable. Not doable for every academic. So the situation for some academics is, like, so dire that the only work you can do is doing inference, where you have closed models or open models, and you get completions from them, and you can look at them and understand the models. And that's very well suited to evaluation, which you become ex-- you want to be the best at creating representative problems that the models fail on or show certain abilities, which I think that you can break through with this. So I've-- like, I think that the top-end goal for a researcher working on evaluation, if you want to have career momentum, is the frontier labs pick up your evaluation. So it's like, you don't need to have every project do this, but if you go from a small university with no compute, and you figure out something that Claude struggles with, and then the next Claude model has it in the blog post, like there, there's your career rocket ship. I think that that's hard, but it's like if you want to scope the maximum possible impact with minimum compute, it's something like that, which is just get very narrow, and it takes learning of where the models are going. So you need to, like, build a tool that tests where not Claude Four point five will fail. If you're gonna do a re-- if I'm gonna start a research project, I need to think where the models in eight months are gonna be struggling. But what about developing totally novel ideas?... this is a trade-off. I think that if you're doing a PhD, you could also be like, "It's too risky to work in language models. I'm going way longer term," which is like, what is-

Lex Fridman 2:15:18 Mm.

Nathan Lambert 2:15:19 -what is the thing that's gonna define language model development in ten years? Which I think that I end up being a person that's pretty practical. I mean, I went to my PhD, where it's like, "Oh, I got into Berkeley. Worst case, I get a master's, and then I go work in tech." It's like I'm very practical about it, so I'm like, the life afforded to people to work at these AI companies, the amount of-- like, OpenAI's average compensation is over a million dollars in stock a year for an employee. Any normal person in the US, to get into this AI lab is transformative for your life. So I'm pretty practical of like-

Lex Fridman 2:15:50 Mm-hmm

Nathan Lambert 2:15:50 ... there's still a lot of upward mobility working in language models if you're focused, and the outcomes is like, look at these jobs. But from a research perspective, the transformative impact and these academic awards, it's like, be the next Yann LeCun, is from not working on, not caring about language model development very much.

Lex Fridman 2:16:07 It's a big financial sacrifice in that case.

Nathan Lambert 2:16:09 So I get to work with some awesome students, and they're like, "Should I go work in an AI lab?" And I'm like, "Uh, like, you're getting a PhD at a top school, or you're gonna leave to go to a lab?" I'm like: I don't know. Like, if you go work at a top lab, I don't blame you. Don't go work at some random start-up that might go to zero, but if you're going to OpenAI, I'm like, it could be worth leaving a PhD for. [chuckles]

Lex Fridman 2:16:30 Let's more rigorously think through this. Where would you give a recommendation for people to do a research contribution? So the options are academia, so get, get a PhD, spend five years publishing. Compute resources are constrained. There's, uh, there's research labs that are more focused on open weight models, and so working there, or closed frontier labs, research labs-

Nathan Lambert 2:17:01 The-

Lex Fridman 2:17:01 ... OpenAI, Anthropic, xAI, so on.

Nathan Lambert 2:17:04 Um, the two gradients are the more closed, the more money you tend to get, and but also the le- you get less credit. So in terms of building a, like a portfolio of things that you've done, like, it's very clear of what you have done as an academic, and you have done this. And versus if you are gonna go be, like, trade this fairly reasonable progression for being a cog in the machine, which could also be very fun. So I think it's very different career paths, but the, like, the opportunity cost for being a researcher is very high 'cause PhD students are paid essentially nothing. So I think it ends up rewarding people that have a fairly stable safety net, and they've realized that they can operate in the long term, which is they want to do very interesting work and get a very interesting job. So it is a fairly-- like, it is a privileged position to be like, "I'm gonna see out my PhD and figure it out after because I, I want to do this," and I think a lot of academ-- like, at the same time, the academic ecosystem is getting bombarded by funding getting cut and stuff. So there's just, like, so many different trade-offs, where I understand plenty of people that are like, "Oh, I'm doing-- I can't deal with this funding search. My grant got cut for no reason by the government," or, "I don't know what's gonna happen." So I think there's a lot of uncertainty and trade-offs that, in my opinion, favor just like, take the, take the well-paying job with meaningful impact. 'Cause like, not also like you're getting paid to sit around at OpenAI. You're building, like, the cutting edge of things that are-

Lex Fridman 2:18:30 Mm-hmm

Nathan Lambert 2:18:30 ... changing mi- millions of people's relationship to tech.

Lex Fridman 2:18:34 But publication-wise, they're being more secretive, increasingly so. So you're publishing less and less and less and less, and so you're, you are having a, uh, positive impact at scale, but it's, you're a cog in the machine.

Sebastian Raschka 2:18:47 I think it's b-- honestly, it hasn't changed that much. Uh, so I, I've been in academia. I'm not in academia anymore. At the same time, I wouldn't want to miss my time in academia, but what I wanted to say before I get to that part, I think it hasn't changed that much. I, um, was working in, um-- like, I was using AI or machine learning methods, um, for applications in computational biology with collaborators, and a lot of people went from academia directly to Google, and I, I think it's the same thing. Back then, the professors were like, you know, sad that their students went into, uh, industry w- because they couldn't carry on their legacy in that sense. And I, I think it's the same thing. It's like it, it hasn't changed, I think, that much. The only thing that has changed is the scale. But, you know, cool stuff was always developed in industry that was closed. You couldn't, couldn't talk about it, and I think the difference now is, um, well, your preference. Do you like to talk about your work, publish, uh, or you know, you, you are more in a closed lab? Uh, the o-- that's one difference, the compensation, of course, but it's always been like that, I think. So it, it really depends on, you know, where you feel comfortable, and it's also nothing is forever. The only thing right now is there's a third option, which is, um, starting a start-up. That's a lot of, uh, people doing start-ups. Very risky move, [chuckles] uh, but can be high-- is a high risk, high reward type of situ- situation where, uh, joining an industry lab, I think, is pretty safe. You know, also upward mo- mobility. I, honestly, I think if once you have been at a industry lab, it will be easier to find future jobs. But then again, you know, you know, it's like, yeah, how much do you en-enjoy the team and working on propriety things? Versus how do you like the publishing work? I mean, publishing is stressful. It is, um, you know, like, uh, acceptance rate at conferences can be arbitrary, can be very frustrating, but also high reward. If you have a paper pub-published, you feel good because your name is on there, you have a high accomplishment, and, you know-

Nathan Lambert 2:20:48 I feel like my friends who are professors seem, on average, happier than my friends who work at a frontier lab [chuckles] to be totally honest. 'Cause that's just grounding, and the frontier labs definitely do this nine nine six-

Sebastian Raschka 2:21:00 Exactly

Nathan Lambert 2:21:00 ... which essentially is shorthand for work all the time.

Sebastian Raschka 2:21:03 Yep.

Lex Fridman 2:21:03 Can you describe nine nine six as culture that's, I believe you could say, invented in China and, uh, adopted in Silicon Valley? What's, what's nine nine six? It's nine AM to nine PM and-

Sebastian Raschka 2:21:14 Six days a week. [chuckles]

Lex Fridman 2:21:15 Six days a week. What is that? Seventy-two hours. Okay, so what-- is this basically the standard in AI companies in Silicon Valley?... more and more of this kind of grind mindset?

Sebastian Raschka 2:21:26 Yeah, I mean, not, maybe not exactly like that, but I think there is a trend towards it. And it's interesting, I think it almost flipped because when I was in academia, I felt like that because, uh, as a professor, you had to write grants, you had to do t-- you had to teach, and you had to do your research. It's like three jobs in one, and it is more than a full-time job if you want to be successful. And, um, I feel like now, like Nathan just said, [chuckles] the professors, in comparison to a lab, I think they have less, like, even maybe pressure or workload than at a frontier lab because-

Nathan Lambert 2:21:57 I think they work a lot. They're just so fulfilled by-

Sebastian Raschka 2:21:59 Yeah. Yeah.

Nathan Lambert 2:21:59 Like working with students and having a constant runway of mentorship and, like, a mission that is very people-oriented, I think in an era when things are moving very fast and are very chaotic, is very rewarding to people.

Sebastian Raschka 2:22:11 Yeah, and I think at a start-up, I think it's this pressure. It's like you have to make it, and it's like it is really important that people put in the time, but, well, it is really hard because you have to deliver constantly. And I've been at a start-up. I had a good time, but [chuckles] I don't know if I could do it forever. It's like a interesting pace, uh, and it's exactly like we talked about in the beginning. Uh, these models are leapfrogging each other, and they are just constantly, like, trying to take the next step compared to the competitors. It's just ruthless, I think, right now.

Nathan Lambert 2:22:42 I think this leapfrogging nature and having multiple players is actually an underrated driver of language modeling process-

Sebastian Raschka 2:22:48 Mm-hmm, mm

Nathan Lambert 2:22:48 ... where competition is so deeply ingrained to people, and these companies have intentionally created very strong culture. Like, Anthropic is known to be so culturally, like, deeply committed and organized. I mean, like, we hear so little from them, and everybody at Anthropic seems very aligned, and it's like being in a culture that is super tight and having this competitive dynamic is like, talk about a thing that's gonna make you work hard and create things that are better.

Sebastian Raschka 2:23:20 Mm-hmm.

Nathan Lambert 2:23:20 So I think that this-- but that comes at the cost of human capital, which is like, you can only do this for so long, and people are definitely burning out. I think I've-- I wrote a post on burnout as like I've tread in and out of this myself, especially trying to, like, be a manager full mode training. It's a crazy job doing this. The book Apple in China by Patrick McGee, he talked about the-- how hard the Apple engineers worked to set up the supply chains in China, and he was like... They had saving marriage programs, and he told in a, a podcast, he was like, "People died from this level of working hard." So I think this is just like, it's a perfect environment for creating progress based on [chuckles] human expense, and I-- it's-- there's gonna be a lot of-- there's a lot of... The human expense is the nine-nine-six that we started this with, which is like-

Sebastian Raschka 2:24:06 Yeah

Nathan Lambert 2:24:06 ... people do really grind.

Sebastian Raschka 2:24:08 I also read this book. I think they had a code word for if someone had to go home to spend time with their family to save the marriage, and, uh, it's crazy. Then, then colleagues honestly say, "Okay, this is like red alert for this situation. We have to let that person go home this weekend," and, um, but at the same time, I don't think they were forced to work. It's really they were so passionate about the product, I guess, that it is, it is-- you, you get into that mindset, and I, I had that sometimes as an academic but also as an independent person. I have that sometimes. I overwork, and it's unhealthy. I had, you know, I had back issues, I had neck issues because I did not take the breaks that I maybe should have taken, but it's not because no one forced me to. It's because I wanted to work, because it's exciting stuff.

Nathan Lambert 2:24:46 That's what OpenAI and Anthropic-

Sebastian Raschka 2:24:47 Yeah

Nathan Lambert 2:24:47 ... are like. They want to do this work.

Sebastian Raschka 2:24:49 Yeah.

Lex Fridman 2:24:49 Yeah, but there's also, there's also a, a feeling, a fervor that's building, especially in Silicon Valley, aligned with the scaling laws idea, where there's this hype where the world will be transformed on a scale of weeks, and you want to be at the center of it. And then, you know, um, I'm-- I have this great fortune of having conversations with wide variety of human beings, and from there I get to see all these bubbles and echo chambers across the world, and it's fascinating to see how we humans form them. And I think it's fair to say that Silicon Valley is a kind of echo chamber, uh, a kind of, um, silo and bubble. I think bubbles are actually really useful and effective. It's not necessarily a negative thing because it could be ultra productive. It could be the, the Steve Jobs reality distortion field, 'cause you just convince each other the breakthroughs are imminent, and by convincing each other of that, you make the breakthroughs imminent.

Nathan Lambert 2:25:48 Mm-hmm. Bernd Hobart wrote a book classifying bubbles, but essentially one of them is financial bubbles, which is like speculation, which is bad, and the other one is like, I don't know the term, but effectively for build-outs because it pushes people to build these things. And I do think AI is in this, but I worry about it transitioning to a financial bubble, which is like, it's-

Lex Fridman 2:26:05 Yeah, but also in the space of ideas, that bubble, you are doing a reality distortion field, and that means you are deviating from reality. And if you go too far from reality while also working, you know, nine-nine-six, and you, you might miss some fundamental aspects of the human experience, including in Silicon Valley, and this is a common problem in Silicon Valley. It's like, it's a very specific geographic area. You might not understand the Midwest perspective, the full experience of all the other different humans in the United States and across the world, and you, and you speak a certain way to each other, you convince each other of a certain thing, and that, that can get you into real trouble. Whether AI is a big success and becomes a powerful technology or it's not, in either trajectory, you can get yourself into trouble. So you have to consider all of that. Here you are, a young person trying to decide what you want to do with your life.

Nathan Lambert 2:27:02 The thing that is... I don't even really understand this, but the SFAI memes have gotten to the point where permanent underclass was one of them, which was the idea that