Dwarkesh Podcast - AI researchers debate how close we are to recursive self-improvement

Episode Date: September 11, 2026

New episode with John Schulman, Beren Millidge and Charlie O’Neill. I got together with some of the most insightful AI researchers I know who are at the openish companies, because I wanted to hear the details of what's actually happening at the frontier and what comes next.Watch on YouTube; read the transcript.Sponsors* Antithesis helps you trust your code. As agents generate more and more of your software, the bottleneck shifts from your engineers actually writing code to verifying it. Antithesis does that testing for you. Ron Minsky, who co-leads Jane Street’s tech group, told me that Antithesis was able to help his team shake out bugs in software that had already undergone heavy review. If you want to see how it fits into your development process, go to antithesis.com/dwarkesh* Grok Bot has been a great way to hand off tasks. My team uses it as a producer: whenever my editor posts a rough cut of an interview in Slack, Grok Bot opens the transcript on its own computer, matches my notes to the exact moments they refer to, and uses a file of my preferences to suggest edits. Then it sends me its top clip candidates so I can review everything from my phone, which saves my editors from sorting through hours of footage. Try Grok Bot for yourself at x.ai/bot* Jane Street just launched its most ambitious competition yet: design a protocol-emulator ASIC. Basically, if you have a chip you want to test outside of a live system, you should be able to connect it to your design and have it simulate realistic traffic. Jane Street wants general-purpose, reprogrammable designs that can work across multiple protocols and remain useful as new ones emerge. The most novel submissions will actually get taped out, and the winners will receive a physical copy! The competition is open until January 18, 2027, and teams are encouraged. To get started download the template code at janestreet.com/dwarkeshTimestamps(00:00:00) – Steelmanning the case against RSI(00:18:39) – What’s driving the Chinese labs’ progress(00:28:06) – How will automated AI researchers be trained(00:33:51) – Will long-horizon RL elicit AGI?(00:45:24) – The sim-to-real gap(01:00:33) – How much progress is explained by data?(01:18:03) – Why is RL working so well?(01:24:54) – Move 37 and entropy collapse(01:28:32) – Rapid-fire timelines This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.dwarkesh.com

Today, I'm chatting with three of my AI researcher friends, from whom I learn a lot every time we talk, and who also happen to be at somewhat open-ish labs and company, so you guys can actually say things on the record. I'm joined by Baron Millich, who is the CTO of Zipra, which is developing open-source models. John Shulman, who is the chief scientist at Thinking Machines, previously, the co-founder of Open AI, led the RLHF work that led to Chethept, and Charlie O'Neill, who is head of model training at Base Ten. The first question I have, if we're in 2036, it's been 10 years, and we don't have like billions of crazy super intelligences that are running around that have like radically transformed the world, what is the most likely reason that that doesn't end up being the case?

Other than sort of exogenous political shocks or like there's a war or they ban AI or something, but what is the most likely technical reason that we don't, like 2036 isn't like a crazy alien super intelligence world? I mean, like my reason would just be like, it's got to be the, it's sort of, like, there's been a classic thing, almost like Marvex paradox, right? Where, like, we see, like, you know, we think of the AI, be like, if it can do this, is going to be amazing, right? Like, if it can solve these hot maths problems, if it can win a chess, blah, blah,

and then it solves these things, and then it's, like, not that impactful. Obviously, it's somewhat impactful, but, like, not everything. It's like, if somehow that continues, and, like, there's never, like, the true, like, spark of generalization that occurs, I think that could lead to, like, the AI is just being, like, extremely good at, kind of everything that people, like, put into a benchmark, put into an environment, but, like, there's still some, persistent like SimTreal, which is somehow blocking everything. I think this is kind of unlikely.

I think we do actually see this kind of generalization even from our own practice already. But like if it is just like ridiculously hard to like generalize meta learning, plus like we don't solve continual learning, it's just like super hard and impossible. Yeah. Like this would be my like default scenario in that case. Yeah, I agree with that. Humans have a lot of advantages over models now. And each time a new model comes out, it'll sort of,

it'll catch up in some of these areas, but you end up getting bottlenecked by the places where the model is weaker and where it has worse judgment or the models can't check themselves well enough. Yeah, so there's this cycle that keeps repeating where people think, where a new model comes out and people are blown away and they're like, this is it, this is AGI, but then they use it a bit and then it starts to feel dumb after a month or so. So that cycle just might keep going. it's hard to predict how many times it's going to repeat.

And like right now, you don't get explosive growth in capabilities because you still get bottlenecked enough when you're trying to do research and engineering that even if the model can write way more code than a person, it doesn't make you like a hundred times more productive. But, yeah, so maybe they're just more of these cycles than we would expect. For me, it's like a question of how far off like this global optimum of a level. learner you could have on a chip is like the transformer plus like RL basically like the current recipe. So like I think people imagine that even once like once you have a an agent which is better than all humans at AI research, even if it's like 0.1% better than all humans,

then the fact that you can run like hundreds of thousands, if not millions of these in parallel, you can run them much faster like chip's going to speed up. That's going to outweigh every other like bottleneck and like you're eventually just going to like hit this like very fast takeoff with percursive self-improvement. I could imagine that if we continue along the trajectory that we're currently on with that paradigm where, you know, it's basically just like self-attention, RL, scaling up RL environments. The, I guess like, if you think about what happened with Moore's law, right, like, we had this very, like, nice straight line and that held for a really, really long time. But there were so many, like, discrete, like, discontinuities and innovations that

had to happen to keep that scaling law going. And the same thing has kind of happened with LMs. Like, we had this pre-training, like, scaling law, and then that was kind of, like, you know, hitting the diminishing returns. And then we came up with, like, RL and solved that. And then we got this new, like, you know, diminishing returns curve to hit that made it keep looking like a straight line going up. And so, like, if it requires another one of those discontinuities to solve, like, I'm not sure that, like, the current method of, like, training LNs with these oral environments, even, like, RSI-targeted oral environments would be able to discover that discontinuity. and if not, like, we're probably going to hit this like asymptotic, like, curve where, like... Sorry, but do you think discontinuity will be harder than anything that's come since 2012? If we had the answer to that, that we kind of have the ability to implement it, but, like,

maybe there's the distinguish, we should distinguish between a discontinuity, which adds to the current paradigm. Again, it's, like, cumulative. Like, there's some thing beyond the R.L. That we have to discover, and maybe they're capable of, like, you know, connecting the dots in that straight line. Or, like, like, again, how far. far off the global optimal optimal are we? Do we have to go back and throw out like, you know, gradient descent and like neural nets in general? And I don't think like if you continue to scale up the current paradigm and LLM, no matter how many LMs you're running, are capable of

necessarily discovering that if it's too far away. Yeah, the only hope really is if deep learning just can't get us to an AI, which is at least can dominate human research and human development, including the human ability to come up with new paradigms and so forth. Or like, I don't know, maybe humans would also never have discovered the next learning architecture, but to the extent humans could have discovered it eventually. But it just seems like, I don't know, if you just look at the progress this happens in 2012 till now, and you just continue that on, I mean, I know if it's been powered by a huge amounts of compute scaling and so forth, but it would be weird if, like, it just didn't get to the point where it could, like, dominate humans, at least.

in R&D, especially over the next few years, there's going to be Ryan Greenblab was on the podcast recently. And he made this point that you could imagine as the AIs get more and more capable and are capable of making progress on simulations, which incentivize getting better at not only AIAR&D, but generally at science. So this is a thing that all the labs are targeting, many startups are targeting. Or another intuition pump is if you look at the ELO score of chess bots since the 80s, There's just like a very linear increase in ELO over time, but there's a huge discontinuity as they cross the human range of human experts always win against AIs to like human experts never win against AI's as this linear increase in EEL happen. And you could think, I agree with your point that so far, AI capabilities have not been that big of a deal in terms of their end economic impact in the world. But that's because like they're slowly rising in ELO relative to humans. Yeah. Yeah, I mean, I agree it would be very, I mean, the only way for this to not happen is if like, as you said, somehow asymptote, like, just before, basically, because we're already pretty close, in my opinion,

to, like, where we'll start crossing, like, the human ELO score. And so we'll need to asymptote before that. And, like, that's the only way in, you know, in this scenario you post where, like, somehow we're sitting here in 2035 and, like, everything is normal for this to happen, I think. I mean, the only other way is, like, some dramatic, like, regulation on AI. It's, like, this is kind of what I see is, like,

the most likely way for this scenario to happen, actually, rather than the technical thing. Yeah. I think there's different kinds of research. There's, like, research where it's, like, the auto-research style, where the objective is already specified very cleanly, and you're optimizing that objective.

And I think everyone is picturing, like, if we continue along this path, making pre-training loss go down, making our environment, getting a water on those go up, that's going to lead to, like, improvement. But, like, yeah, maybe what Ryan is talking about is like this much more open-ended type of science,

which is required for, like, paradigm shifts, where we can't specify the objective, and the AIs are definitely not able to specify that objective either. Like, we have to be really, really careful about how we specify objectives for any of these things. And maybe your point is that, like, the nature of the breakthroughs that have happened since 2012, is that we have found

in 2012 people weren't saying or I'm assuming I don't know you guys were there or at least John you were there but I was in primary school I actually John I'm curious where you're like sort of wisdom of the ages of

or wisdom of like being in the trenches way back when but presumably a big breakthrough was realizing that next token prediction is the like you wouldn't have thought that nano gpte speed run is the thing to be optimizing for in 2014

But now that we have come to this new paradigm, you wouldn't think to do a speed run on that and have AIs get really good at that. But maybe there's like a next inner loop to optimize that the AIs wouldn't anticipate. And there's an outer loop of like revenue or something that eventually should be strong, but it's a very slow outer loop.

Yeah. In fact, I remember in the early Open AI days having the intuition that actually just minimizing log loss wasn't going to get you intelligence because like the important bits are accounting for such a small fraction of the loss that

it was going to be overwhelmed by noise. So just training a language model on next token prediction just wasn't going to learn the interesting things you wanted to learn. And we needed to craft better objectives that would put more emphasis on the important things.

And like you can make all sorts of arguments for this and you could say, oh humans, probably don't learn how to, like, we don't learn how to model everything in our environment. We can't, like, people can't create a photorealistic reproduction of some kind of scene they've looked at. So there must be, we must need a better objective, but then it turned out that it just worked anyway. As you were pointing out, the inner loop, even in current AI research of, like, post-training benchmarks or whatever, doesn't necessarily translate it to what users like.

Oh, yeah, I mean, the whole field relies a lot on generalization, and it's very hard to predict when you're going to get generalization or when you're going to get some kind of out-of-distribution generalization. So we know that if you train on the task you care about, you're going to do better. Yeah. But, like, the most important advances are often the types of generalization that we have no right to expect.

So, for example, from just preaching, training on this very naive next token prediction objective to like various tasks of interest where some very, uh, that require understanding of the input in some deep way or learning some skill from pre-training that's like very rare and like not very like heavily represented. And then like also generalization from these verifiable tasks to less verifiable ones. This is also a type of generalization that there's no reason a priori to expect it. Yeah, yeah. So this is an interesting question because one intuition pump that you could have

for why you would see some sort of singularity very rapidly without even scaling up the inputs to AI progress that are not just AI labor, is that before every single experiment you run, that's, you know, like a seven-figure experiment, you spend an equivalent amount of compute on AI labor. And so you just have automated versions of you guys. spending a century, thinking about, like, what is the optimal experiment to run, doing small-scale ablations, developing literally, like, a century's worth of theory. So going back even before, like, deep learning, before you decide what experiment to run,

doing extremely optimal, like, setting up of the experiment, then you do a century of thinking after the experiment is over, where you're, like, analyzing what happened and what the next experiment to run is. Well, I think if you think hard enough, you probably could have expected some of these things beforehand. Like, there is probably some very clever way to do a small-scale experiment that will let you build the theory that then will generalize to the large-scale experiment. So I would expect that, like, we're nowhere near the ceiling of how well you can do research. And, like, I would imagine a future where AI is, like, is doing a lot of, like, analysis and theory

building, like spending a comparable amount of compute to the amount that you're spending on the experiments themselves doing various kinds of analysis and building a theory around what we've seen so far. I think there's like really concrete examples of this when the objective is well specified. So again, like all thinking can do is like update your posterior based on like, you know, the bits that you've gotten since you formed your prior. Like you can't gain any new bits from like just thinking. But when the objective is well specified and there is like this data sitting around, like, I imagine there will be this big speed up in the current paradigm we're in. And like good example, this is like, you know, if you got an AI to think about like, like,

like the Kaplan scaling laws. Like, an AI at this point would have noticed that like, you know, oh, like they've just taken these intermediate checkpoints and didn't account for like the annealing. And so like this is wrong. And like that would have caught that like years earlier. We would have made like, you know, progress like would have cut off a year or two of progress just from that like observation from an AI.

And like again, once the objective is well specified, which is like lower pre-training loss or whatever, like there's many, many good examples where if you just thought about it a bit more, you would have been able to cut down a significant on things that you've done. So like Mupi and like how learning rate scales with like model size and like realizing that model width is important in that as well.

Like I feel like you can you can really back out a lot of these things and cut off like a lot of like hanging fruits. I would imagine like a 10 times speed up if our thing is just like maximize the objective we're currently on. But I don't see that how that generalizes it all to come up with the right objective in the first place.

Like just thinking doesn't necessarily buy you the right objective in the first place. I mean, yeah, I think this is really the key question to like any kind of like very rapid RSIs like from current AIs. It's like how well can AI is generalize to like learning their own objectives? Because to have any kind of like self-propelling automated loop, you need the AI to like propose objectives, optimize them, figure without out, proposing your objective. And like have this like not go off the rails at like any point for like a long, long time.

Coming back to Movix products, there might be like a case of Moavx products where like we think this kind of like autonomy and sort of like being like self-encapulated so we can, you know, think of what we should do ourselves and then go do it and like have this. loop is like super easy because we always do this. And like obviously evolution needs to create creatures that can like survive on them by themselves like long period of time. And like this just might be something that for some reason it's like really hard for the AI in the same way that like locomotion stuff is really hard. It's like math is super easy despite being super hard for us. I don't know. Doesn't the time horizon increasing suggest that that's you would yeah exactly. I mean this is another possibility which like but I agree like there's no

obviously evidence for this like in fact the fact that you know our agent's not like super persistent and it's quite easy to do this is kind of evidence against this. But like this would be potentially like one of the reasons why like we just don't get this like immediate takeoff is like if this is hard. If you look back from 2012 till now, or maybe from when you started doing your research till now, what part of all the innovations that have happened since that time, including purely engineering ones, including purely conceptual ones, what seems like the thing that is the thing that would be the last things humans would have to do before AI is totally, automate AI R&D. Probably just like iteratively asking the right questions.

Like if you can get the AI to like do any experiment, but like you need to decide what experiments are doing. Like right now I think AI is not very good at this compared to coding the experiments at all. Yeah. Like whenever we talk about research, they propose like a bunch of like miscellaneous things which are like very, very tiny steps.

Or even going from like, you know, deep minds approach of like we're going to solve intelligence by learning to play games at a superhuman level. That's going to be the approach to like one random researcher like Radford being like, I'm going to try and just predict the next token of a very wide swathlet of data. And then even once Radford had discovered that, right,

it took a while before people decided to scale it up because we had to come with the idea of scaling levels and the fact that you could very reliably predict these things. Yeah. I would say that the last job for humans, or the role for humans that will last the longest is defining the objective and deciding what we actually want. So, like, in that vein, something like deciding what, how the AI, the assistance should behave or what it means to be helpful or what's like the objective when we're doing are all from human feedback is one such thing. And then later, like, defining, like, constitutions and model specs is another one.

And I think even if the AIs can do all the technical work, we'll have to still do a lot of that and decide what we're. we actually want. Alignment is the final job. Yeah, alignment is sort of the answer, but it's also, alignment itself can be kind of decomposed into, into like specification of the objective or figuring out what the right objective should be. And then, like, actually, like, achieving or optimizing the objective you've defined.

And I think the first one is not going to go away anytime soon. And, like, if I think about, like, a post-training team and, why you need a lot of people to be on the team. It's just because there are a lot of different areas where you have to figure out how the model should behave. And like there's no way of, like, there's no way of, it would be very hard to automate the whole thing just because someone has to think about how should the model behave in this area.

Jane Street started using antithesis to test their software in early 2025. And they were so impressed by the product, that they decided to invest in the company. I recently caught up with Ron Minsky, who co-leads Jane Street's tech group to ask about how Antithesis actually plugs in. The thing that I think is most impressive about Antithesis is we started using it in a team

that was building high assurance software and being really careful, and nonetheless, it was able to shake out bugs that were otherwise going to be really hard to find. And that's important both because it helps make those systems more reliable, but also because it helps the teams that build it to just move faster.

This matters more and more as code production is increasingly automated. I think in general, as we've been using agents more and more, the key problem that you run into is the verification bottleneck. Just the time it takes from people to look at code and figure out, is that actually something you want to accept in your production software? And tools that make testing better are just incredibly helpful there. They just ease the verification bottleneck and make it possible for you to get more stuff done and move faster because you can have more confidence that the code generated by the agent is actually not important. introducing new problems. To see how antithesis fits into your development process, go to antithesis.com slash Thorekesh. What is the story for why there isn't huge consolidation in model providers?

There's just so many things that point to centralization here. Is there, yeah, if he step back over the course of years, is there something that is going to prevent that? Yeah, I think distillation is the main thing that fights against the centralizing force, because basically anything that can be learned through RL can be distilled very easily because it's a small number of bits. It's something that you can learn from a small amount of data. So if you can get trajectories from the model that show a behavior, you can easily distill it. So I think distillation is one of the things that fight centralization. There's also, I mean, there is a possibility.

that there'll be company-specific models, that it'll be possible to learn from deployment and have a company continually improving its own model. And such a system could be provided by the current oligopoly of model providers or some other currently smaller company. But I think that'll change the game a bit. Yeah, and I also want to point out that, like, continual learning and honestly it doesn't stop distillation, right?

Like, even if your model is improving every day, like people could be distilling it every day. So it's like the loops could just operate at the same pace. Right, that makes sense. Okay, so copying model behavior. I guess you need to know yourself what the right distribution to prompt is in order to get like the relevant model behavior. Well, yeah, for just distilling with supervised learning, the prompt distribution is extremely

important. So it's very non-trivial to distill a model even if you have full access to it and have the cot. the chain of thought and everything. Yeah, it's non-trivial to distill all of the useful capabilities from it because you need to prompt the model with something. And you need to prompt it with realistic prompts. You need to have a really wide distribution of realistic prompts. So, yeah, one thing that's been coming out recently is some of the Chinese companies

are probably using these router services, which are just designed to allow people in China to use the U.S. frontier models, which would otherwise be blocked in China, but there are all these router or proxy services that allow people in China to use these models mostly for coding, and these router services are collecting and selling some of the data. So I think this is like a very useful data set for distillation because it gives you the perfect prompt distribution. I think this is one of those things where AI's help a lot here. Like, if you actually look at, like, you know, the frontier pipelines, or say, like, the Chinese models that they actually put in their papers,

it's a lot of, like, humans or, like, they get seed prompts from somewhere, which is some combination of humans, this kind of data. And then they, like, synthesize a vast coverage from those seed prompts using their existing models or, like, the other frontier models. And so it's like you can automate, like, an awful lot of this, like, prompt distribution gathering and, like, environment creation. It's just, like, humans need to provide, like, increasingly few amounts of bits. It's like the models get better.

Right. Or, but it still seems to you're bottlenecked by, like, having, a service which has users or users are going through. Not necessarily. I mean, like, yeah, that's obviously very helpful, but theoretically you can just think about, like, what users want or like, no task.

The whole point is that we don't, the user says, make me an application like this. Oh, that didn't work. I actually want you to make this new feature. But actually, let's step back and do this other thing in capturing that whole trace is the, or to the extent you could have done that anyways, then you just have like RSI.

Yeah, I mean, like, ultimately, like, if you have this like fully automated loop, that is basically our side, right? Like, the AI is deciding the data, it's deciding the training, that that is the loop. But, yeah, I mean, like, it depends how much human information you need. Like, at some point, if you're just, like, I want traces that look like this.

You prompt that to the model. The model will be able to, like, come up with, like, a pretty good approximation. But what if you want to do, like, make me a really good politician? And then it does, like, anticipate de novo. Like, how would have a discussion in, like, the Senate halls go or something? Yeah, I mean, like, there's going to be a lot of things, which I want to, I think easier for the distillers than the frontier labs, right? Because the distillers just like, I want a good politician.

They go to the like frontier model. The frontier model already knows how to be a good politician. So it just like generates those traces. Whereas like if you actually want to build the first model that does this, you have to like actually somehow like get data on like what politicians do every day and like build that. So it's actually much easier to like say like I want something like this and then like get like the AI to produce like a billion variations than to like actually create the thing like this to begin with. I think you can actually make a really concrete prediction based off like this observation

that the Chinese labs have this router dollar. So, like, I think the thing that just stated this originally was I was saying, isn't it weird how Sonnet 5 and Opus 5 are, like, almost objectively worse models than, like, GLM 5.3, Kimi K3, even though they've had access to, like, not only distillation, but logic distillation from, like, mythos. And so the counter here was that, like,

okay, the prompt distribution really, really matters. Like, you need to see what users are doing so that you can distill, like, kind of these behaviors and things in. I think the prediction from this is that the frontier labs don't necessarily have much of advantage, if at all, in oral environments now. Because, yes, like, user distribution matters for, like, general, like, behavior and so on. But, like, the best measure of a capability is the very, very hard oral environments you've made at the frontier. And so if you have access to those oral environments as anthropic and you have access to logic deflation and you've still made a worse model, then maybe like... Then real world deployment matters more than the environment.

Yeah. That's really interesting. So, but they had to incentivize those capabilities in the first place in Fable or the Frontier model. And so it's weird that they can't incentivize them again with like a smaller model or something. Maybe like maybe we're just in this weird like uncanny valley where, you know, like actually trying to copy that frontier model too much. Like the student teacher gap or whatever it is, it's just like too large. And like I think people have made this point with Opus is it's like the difference between Opus 4.6 and Opus 5 is that Opus 5 really. really feels like it's got this like AI as a judge checking every possible thing it's done.

It's like, that's why it uses so many tokens. It like tries to think about all these things. But it doesn't necessarily have the big model smell of Fable to know when to like stop doing that or like when's a good path to go down or whatever. The research sees the grass. Yeah. Yeah, I would offer a slightly different hypothesis. So I would say there are a couple of different axes for the environments you can create. And like one of them is difficulty and the other is realism.

it's sort of easy to create, or it's comparatively easy to create a lot of difficult environments like, that are just, like, involve, like, doing a much more complicated task or doing something that requires a lot more cleverness. And you could say this is like the benchmaxing distribution

because a lot of the most prominent benchmarks just involved doing some very hard puzzle-like task that's easy to verify. And then there's sort of like the realism axis where you want the model to be good in the realistic coding agent setting where there's like multiple back and forth with the human and there's like multiple objectives. And like I'd say like the people, like the labs who are crafting the model behavior

for the first time need to push in both directions. And to get good model behavior, you need to really push on the realism axis and have like rubrics or some kind of human feedback that's informing the reward function you use there. But I think when, if you try to do distillation naively, you end up just sort of matching the teacher on the benchmaxing distribution. And, but if you don't have enough of the environments that really exercise the capabilities in these like trickier realistic settings, then you're not going to, you're not going to get those into your student model. And I think maybe one thing that's happening is the big models generalize better from the like the tricky narrow tasks to the, sort of more realistic tasks. So if you have a really good, like, realistic prompt distribution

for distillation, you can match the big model really well. But if you only have this, like, this distribution of easily verifiable tasks, then you can match the big model on all the benchmarks, but you do worse on this broader distribution. So you might, that might even explain something about the smaller anthropic models like Sonnet 5, though it's hard to predict exactly what they're doing to post-training those models. It could also be that they're always changing their post-training stack, and they just got a few things wrong in some of these models. So they, like, I don't know, they turned up like something too high and created some quirks that people really don't like. So it's like really easy to screw up post-training in some way

that doesn't show up in benchmarks. I mean, just one of the sort of very basic point is just like the AI company, like the frontier AI labs buy all their data from their data companies. And like the Chinese can also just buy the same data from data companies. And they are. And like they are.

Exactly. There's a lot of people like, you know, being annoyed about this. But like if they have exactly the same data and like they can buy that, they can also distill. It's like, it's, it means it's quite easy to like keep up really. Yeah, yeah, yeah. Okay. The other question I had is how the first models that are capable of automating

AI R&D will actually be trained. Because there's a toy version, which is this thing that Ryan was talking about, which is you just have GPT-8 try to build GPT3-size models that are really good at like interloop-type challenges of beating video games that require continual learning or just getting to a certain wask with like the least amount of compute, et cetera. But John, I think you had an interesting point that maybe that's not the way it actually will happen in practice. So I'd be curious about, yeah, by the point of which you have AI that are at, you know,

actually capable of automating AIR&D, how are they probably trained? Yeah, I think we'll probably do some combination of learning from human feedback to absorb, like the researchers' taste, and just like creating a lot of practice environments, which involve doing multi-step research projects. So I think, yeah, people will in practice

do some combination of those two things, and just each iteration, like patch, whatever seems to be most broken in the last iteration. So like researchers will be using the AIs a lot and and we'll notice that they have some consistent weaknesses and then those things will either be patched by like collecting human feedback or like creating environments. Yeah, makes sense. Maybe maybe useful when I think about this is like how much of the lineage we roll back and then let self play from there. Like I think in the limit like you're picturing like you know

just giving them like a GPU and maybe neural nets or something and saying, like, okay, figure out how to train a model to, like, do this particular task. Like, the way it currently works is, like, we go up to the very, like, edge of the lineage and say, okay, like, here are the bugs, like, you know, Anthropica's found in their training stuck in the last few months. We'll turn those into environments. Like, you need to train and get better on the frontier. And so you obviously lock in all the previous history of the lineage, but you could imagine a world in which you roll back to, like, you know, before GRPO or something. And then you have environments which, like, trying to get it to discover, like, the best will form

to, like, RL models on. and maybe you roll further and further back, but I think we will be still so compute bottlenecked that people will just keep staying at the frontier and, like, diffing essentially the bugs and whatever improvements they found since the last model version and turning those into training environments. Which is also really good for having non-steal, like, new data

between model generations. It's just, again, this is basically a continual learning within the AI lab of distilling the last three months of AI research progress through environments and like RLHF's type stuff back into the model itself. And it is distilled. feeling right. And that's maybe why some of us feel like it's asymptotic. It's like you're always like just trying to get the last three months of progress. And that progress is being

contributed to by airs, of course. But it also still has humans in the loop. And it feels like, you know, you're just constantly inching closer and closer to what the human researchers are like finding capable of doing. Yeah. I mean, the one thing I'll say, though, is like obviously if you're just distilling on like trajectories, you can never go above it. But environments can go quite a far away above what the humans can do. Like it's very easy to design an environment that like no human can solve. But the AI can obviously still try and solve it. And so That would be the path to go ahead of just what the human AI research is. Do you have an example of, like, in terms of RSI or like, like, you know, working on a speed run,

but doing it even faster than a human speed runner? Yeah. I mean, I feel like in AI research especially is very easy to define, like, goals, which, like, you know, you could say, like, the loss needs to be, like, 1.3 or something. And, like, no human can get there, you know, now. But, like, that's a very extremely measurable, verifiable task. And if the AI gets there, then great. Right. And building, like, a hundred million parameter model that beats Minecraft.

That's maybe too easy. like, be a much more complicated game or something. Isn't it crazy that 100 million permanent models to bait Minecraft way calling that too easy? Like I can imagine she said that like five years ago. I would say a lot of research is not exactly like that, though, where it's like hill climbing on a well-defined goal. It's sort of more like, here's an intuition we have about some way models

should be better. And then we also have some idea for an algorithm that seems to go a little bit in this direction. in. So let's come up with a task that is sort of designed to show signs of life on this approach and see if we get those signs of life. And then if we do, we can make successively more realistic versions of the task. Right. It's like a lot more guided by intuition. And then the inner loop is to elicit the or make tests for that intuition rather than like the test itself leading to the insight. Right. Like you're not directly optimizing for the eventual objective you care about or the

practical, like, production objective. It's sort of you're relaxing your objective a little bit. You're saying, yeah, let's relax on the realism axis a little bit and find some methods that actually work and then like then try to get back to realism later after the method matures a little bit. Yeah, yeah, yeah. And then there's also like more, there's research that's more oriented towards explaining things and like, developing a theory or a sort of, yeah, often we don't have like mathematical theories in machine learning that are that predictive, but we have like a lot of like more informal theories for what's going on. Yeah, I mean, like, presumably the models will be trained on like some combination of all of these tasks and like some will be very easily verifiable. Some will be like LMS judge or like, just. ask the human, like, does this look reasonable?

And then the hope would be that, like, these would all generalize to like these much sort of hard, sort of more vague, fuzzy kind of tasks. And, like, it probably will to some extent whether it generalized enough that, like, the loop can become, like, self-sealing without humans being in the loop at all. It's, like, unclear. Yeah, yeah, yeah. Maybe taking a step back, here's what I, here's what it seems to me that the plan for AI research going forward is.

And you tell me if you think it's going to work or if you agree with this characterization. So the bet is that we will scale up our OVR training across millions of diverse environments across hundreds of different kinds of domains. And what will emerge at the other end is an agent which has learned these basic skills or less than basic skills around being persistent, being able to triage information and context, eventually having like end-to-end optimization of working with other agents and things like that. and such an agent will be very sample efficient within the context.

You know, you've done research on how you actually scale up in context learning to make it like arbitrarily long, but you keep scaling it up. And so what comes out of the other end will be something that it basically functions like a drop in remote worker over the course of a week or a month.

First of all, do you agree that that is the bet the labs are making? And second, is that enough? Like basically learning how to learn within the simulacra within a data center and then getting deployed into the real world but not actually learning from real world deployment, only learning these metaskills from the simulated environments in the data center. Yeah, I think it's now hard to separate out

how much of the lab's effort is going towards direct RSI versus making generally intelligent models that they can continue to deploy to collect revenue to fund the next big training run. I think for the latter, yes, that's probably just the bet they're making. And it's very clear, like, the pattern of where these environments are going over the last few years. I mean, like, Anthropics lineage of environments is like a very clear example of this. Like, you know, first, like, we just focus on coding and, like, we're going to get really, really good at that. And then the task horizon that we've got from coding, which is probably

the lowest hanging fruit in terms of, like, data available on the internet to create environments, like their own internal stuff that they can turn into environments. Then we're going to generalize, we're going to go after finance next. And like, literally, like, just so much Excel data and all that sort of stuff in the in the oral training. And then, you know, it's PowerPoints. It's like this long tail of the working economy, and, like, that seemed to work really well in, like, a little of the other labs and things, even the open source labs have now realized that that was the correct bit to me. But what is the implication from that? When I had Dario on the podcast, the thing I asked him was, if you truly expect models, which will be human-like in their ability

to learn on the job, why would you try to beacon all these skills of, like, working with PowerPoint or something? Wouldn't you just expect the model to be able to pick that up on while it's deployed? And so, yeah, there's multiple different explanations. One is just that this is, we expect models to get there soon, but they're not there yet. So why not amortize these skills into the model training? Another is that we're not concentrated on making it really good at widely deployed work. We just want it really good at RSI. And this is just like a way for us to like get revenue so that we can pour it back into a model

that is actually like really good at doing RSI development. And then like once the singularity happens, the thing that comes out the other end will be really good at. all the things which seem like bottlenecks to the current generation of models. Yeah, John, I don't know if you have a taste on like what, how much it construe why there is so much task-specific knowledge in these models, if the path is like this kind of generalization. Yeah, I mean, if the models were good enough at learning in context, then in theory you wouldn't be,

you wouldn't need to train them on finance. They would just be able to figure out, read all the books on the fly and figure out how to do everything in the appropriate jurisdiction. Yeah, and you could argue that you need to do a lot of this domain-specific training just to make them more efficient. So even if they were smart enough to figure this out on the fly, you still might want to do a bunch of RL and bake all these intuitions into the weights, so the model would be more efficient at runtime. Yeah. Yeah, I'd say in practice, it does seem like model providers are going domain by,

domain and trying to strengthen the models in the highest value domain. And I'd say that that's one of the answers to why the models have gotten so much better. It's just because the model providers have covered a lot of the high value domains and the most common types of skills. I mean, I think another thing is just like it's not that expensive to do both at the same time, right? Because like the models are massive they can easily afford in terms of their parameters to like learn everything.

And like there is likely some transfer in sort of even just, even if like finance is not specifically like the information is important for like, RSI, just the general meta learning of like how to figure what's important, how to have taste, how to like do Long Horizon work is potentially generalizable. And like, there's not that much RSI like data in the world as well. Like it's kind of hard to generate and like that requires a lot of effort to like if you can sort of amortizing this other data, get some transfer from it, you already have masses of compute and massive parameters based to like why not do that as well as like obviously the direct like commercial intent of like selling a model.

Yeah, that makes sense. Oh yeah, I'll add that I mean there's one question about whether this current paradigm of doing like sim to real will be the dominant one forever. So basically you look at what the real world tasks are like and then you try to create a bunch of environments that can be simulated like in the data center and you can do RL on them. And I think obviously this has been very successful, but it has a lot of weaknesses because a lot of things are just kind of hard to simulate, especially if they involve like interacting, interacting with a bunch of humans in real time. Yeah, so there's some question about

like whether Simterreal will be the dominant framework forever. I think Simterial has to be the dominant framework where, like, sample efficiency is kind of low. Because, like, right now you need, like, thousands and thousands of interactions with the humans and no human is going to sit there and, like, deal with this, basically be in the loop of RL training.

And so, like, we kind of have to simulate that now to, like, get the samples you need. But, like, obviously, if sample efficiency improves a lot, you'd expect learning from deployment to, like, become, like, a much bigger part of it. Though there are also other things you can do like you can learn off policy so you can take all the traces and even without resimulating everything you can potentially learn something from them jane street just launched a new competition

and it's their most ambitious one yet design a protocol emulator asic basically if you have a chip that you want to test you can connect it to this asic and then this asic will simulate realistic traffic that way you can see how the chip responds without having to plug it into a live system jane street is looking for flexible, general-purpose designs, not single protocol emulators. When I was chatting with them, they suggested that I start off by trying to implement what are apparently three very common protocols, U-Art, SBI, and I-squared-C. Jane Street also mentioned that they hoped that more ambitious designs will also tackle low-speed USB and Ethernet and any other protocols that flex your ship's specific architecture. Importantly, your design should be reprogrammable rather than smashing a bunch of

specific protocols onto a ship. protocol comes out after your ASIC is taped out. Your chip still needs to be able to handle it. How exactly does it is up to you. But there is one hard constraint. Your design must target an open source 130 nanometer process node. That's because Jane Street will pay to tape out the most novel submissions and send the physical copies to the winners. The competition is open till January 18th, 2027, and working in teams is highly encouraged. Go to jane street.com slash thorkesh to download the template code and get started. I want to ask more about this, because it's sort of weird that you have 50% of compute

that's spent on inference that is not directly helping the model become better. Like, one of the key advantages you'd expect eventually digital minds to have is, unlike a human who gets to have 50 years of, like, real world of experience, a model will get to, through all its instances, will get to experience, I don't know, millions of years of deployment across all kinds of economically relevant work in the economy. And right now, that data is just not in a meaningful sense helping the model get better. It just seems so obvious that eventually model should be able to learn from this data. And once they do, you would have something that almost feels like a widely deployed

intelligence explosion because the model is assimilating so much information across all these deployed instances. But when do you expect this kind of hive mind kind of crazy shit to be started happening? I think broadly, like at a very basic level, this is already happening. just in the next generation of models. So right now you can always take your deployment data and put this in the pre-trained or the mid-train of future models, especially if you do like some kind of filtering

or some kind of judgment or annotation or recent synthesization of that. How much do you think that explains the generation over-generation improvement? I think it explains quite a bit. I mean, especially like, I mean, this is, you know, I don't know whether the labs do this because, you know, theoretically they claim not to train on people's data. But like the Chinese 100% do. And like they definitely get this advantage,

both like obviously deploying. This is basically what distillation is. Like, they take other models, they get some of their, like, deployment data. They get some fraction of that by, like, pinging the model. And then they train their next generation of models on it. And they can suddenly do it on their own models as well. Like, there's no reason not to whatsoever. I completely agree.

I think if you zoom out far enough, this is, like, definitely happening. Like, you're all picturing this, like, and we're all picturing this. This is, like, what continual learning, like, the Holy Grail is, like, this very, very organic, like, live loop of, like, an individual model, like, getting an experience and, like, live updating on the spot and learning from that. And, like, a lot of things break when you, like, zoom into that level of granularity. Like, yeah, the big labs are doing this, like the closed models are doing this. There's also early signs of life of, like, people using open source models doing this

at a much faster cadence. So, like, a good example is probably, like, composer. Like, you have some sort of model and you are able to, or like, you know, Harvey's doing the same thing with, like, legal agents. Like, it is getting very specific environments from the data that you have for that particular tasks and things that, like, you know, users are complaining about. And, like, all the feedback that you're somehow extracting from, like, your specific deployments and a lot of these companies have the advantage over the big labs

in that they can use this data really, really well. And then they will create environments. They will like, you know, do a big post train of Kimmy K3. They will go deploy it. They might do some online learning as well. Like Composer did online basically like reinforce for a long time. So yeah, like there's still a human in Lube. There's still a human saying, okay, these are the signals we care about. Here's how we're going to create environments from the data that we have. And it's like still a longer cadence than maybe the one that you're thinking of. But like it really is happening. And like eventually that loop will become like faster and faster. I mean, the composer thing is interesting because this is where the model, like, in cursor, people like press tab or they don't press tab

on the next completion that the model suggests. And based on that, every single day, composer gets better at, like, predicting the next. So that was the old tab model. Like, they actually did the same thing for the actual, not just like the tab model, but the actual, like, generative model. Oh, that's interesting. And they, it's hard because when you do online reinforcement learning, you don't have groups, right? You just have one user, say one thing, and then you get one rollout. And so, like, you have a big variance reduction problem. And, like, cursors' kind of fuzzy answer to this was, like, oh, you know, we have very good heuristics, which are able to estimate, like, what the, how much better than average, like, this response was or how

much worse than average this response was. And then they would do, like, this big, you know, like, reinforce update. And then their solution to, like, whether it got worse or not was, like, if it improved on cursor bench, they would deploy the new model, like, every five hours. And if it didn't, they would, like, throw that version out. Interesting. Yeah, I think your biggest problem is actually just not knowing what the reward function should be from natural data. And if you use some kind of superficial signal, like did they accept the code, the edit, you might, that might get reward hacked in some way. But this seems like a bigger issue with the SIM to Real thing, where the longer and longer

horizon tasks get, the harder they are to simulate within a data center, right? It seems to me already, potentially, even in coding, we're getting to it with a point where there's like not some year-long coding task that doesn't eventually require you to talk to a client or interact with the company or interact with users. And if you think about the gamut of things we would want AI to be capable at,

you want eventually a super intelligence to be able to run a business or start a new business and make it profitable or like have a profitable day trading in the markets or win a court case. And these are all things which are very hard to simulate in a data center. Like an inherent part of the learning there is interacting with the real world. And so maybe

maybe they may have to learn how to get better at these things from the transfer between SimTorio. But alternatively, maybe you do need weight updates from these kinds of interactions in order to get better at them. And then if that is the case, if transfer isn't strong enough and you do need weight updates, then the fact that the models are quite sample inefficient is like maybe a deeper problem. And the reason I'm curious about this, I feel like by default, I don't see how you don't get some kind of crazy recursal self-improvement within the next 10 years.

But the one reason why that might not happen is in terms of like weight updates, the sample efficiency of weight updates, they just seem way far behind humans, right? Like plausibly million-fold behind humans in terms of how much data human sees from birth to adulthood versus how much a model sees from, you know, like cold start to like finishing training.

And so, yeah, this is all to say, first of all, is there going to be a good transfer between simulations and extremely long horizon, really complicated real shit that we want the EIs to do in the real world? And if not, does that really mean, that like the aqua sample efficiency in these models comes to bite us. I think maybe the way I'd break down like the two types of tasks in which models get good and models will like still continue to struggle is whether the task is like cumulative

or like you kind of have to, you have this like non-stationary distribution you have to keep learning and like relitigating a bunch of stuff. So like maybe an example of a cumulative task might be RSI. Like it's theoretically possible to maybe like have less than a million token like you know Python file, which, like, from scratch, trains a model that is capable of recursive self-improvement. And, like, every discovery that you make is kind of a line and sand that you hold. Like, if it's true that, you know, for RSI, we don't need to discover a new attention variant or whatever.

Like, once you've discovered attention and then once you've discovered, you know, mixture of experts, once you discover JRP, you just add that to the training stack and, like, that's there. And, like, a good example of this is, like, you know, 5.6 all training, 5.6 terrol, whichever one opening I told it to train. Like, it didn't have to go back and discover attention. And it basically probably would have called a bunch of scripts, which is like pre-training.sh and post-training.sh and just did that.

So like that's an example of like a cumulative task. I think the real world and the reason like people are thinking so much about like continual learning is it's not really a cumulative task. Like imagine like in a law firm you have an agent acting as a legal associate. Like that's a very non-stationing distribution. You have to be able to fit in your context like all the relationships between all the important people of that company, which are also changing all the time. You have all these implicit ways about how things are done,

where to find information, et cetera. And that's not as clean of an example of a cumulative task like RSI is. So I think that there will be this breakdown between tiles, but if the labs realize that, and they do believe that RSI is cumulative in the sense that we don't need to go back and discover some brand new architecture or whatever,

then maybe more and more effort and compute gets focused on that versus the... It's so unfortunate that RSI happened to be easier than they're inherently all. Yeah, I don't know if you guys have thoughts on this. Yeah, I would say there's, like, models are, today's models are weaker than humans in a lot of different ways, and some of them might have to do with sample efficiency in a certain regime

where, I mean, in some regimes, models are very sample efficient, like learning in context, but then there might be some, like, medium-length regime where they're less sample efficient because humans can do, some kind of weight update more efficiently than models. So I think like being less sample efficient in certain regimes might be one of the sources of weakness, but then I think there are other sources of weaknesses that are completely different than that. For example, having lower diversity of thought than humans or being bad at certain kinds of long horizon judgments. I mean, I think a lot of what people call taste is something about,

behavior that works in the long run and that people have realized works in the long run. Not everything, but like some aspect of taste, especially for something like software engineering, like I think a lot of taste is like what are the systems that are going to be maintainable and work well in the long run of this project. So, yeah, I think the weaknesses of humans which limit RSI, along with other things, there's a variety of them, and some of them are related to sample efficiency,

and some of them aren't. Maybe an interesting thought experiment is, like, if you were able to give a model, like a context window of, I don't know, a trillion tokens, or whatever you would have needed to fit in, like, your experience prior to, like, let's say, RLHF.

And, like, it's got all that experience in the context window and it has the same sample efficiency and in context learning ability as it does at a million tokens. Like, do you think taste is then solved? Like, would it be? able to make the same judgments that you did?

Or is there like something fundamentally missing, apart from just a longer context window with the same sample of efficiency? Yeah, I mean, it would have to be trained to learn from that context. So I'm not sure. Yeah, either it would have to be trained to learn the right update to make from that context or we don't have to generalize. You can just like dump it all in, like your whole like life like research experience.

I mean like you still need the data to train it long context. right, like, even if you could theoretically get, like, a trillion context, you would need a trillion lengths of data to train it. Like, right now, you have, like, I'm just asking if you had that. Like, that way, I think, yes. I mean, this really just comes down to the question of, like, how meta-learnable is taste from, like, shorter horizon episodes. And, like, I feel like there's no obvious reason it's super long, because, like, humans somehow develop taste with not having many long episodes. Like, we don't live to be, like, 10,000. We have, like, you know, we develop pretty quickly, right? And so, like, you know, if you think about, like, even, like, in a PhD, the difference between, like, a first

a PhD student and like a final postdoc or something, that's like five years maybe. And they've only done like maybe like 10, 5th, 30 research projects in total, but somehow they develop taste quite quickly from like a relatively short succession of like small things. And so like theoretically it's possible to develop it like that. The AI obviously will have vastly more experience in which to develop taste to like meta-learned. And there's like how well does that generalize to like really long horizon things is I think the question, which I think is really unsolved at this point. Like we don't know.

Going back to this question, eventually, there should be a regime where AIs are learning a ton from each individual instance of deployment that they have. Well, currently, you could say there's a meta, fuzzy process by which models do improve for deployment, but I feel like it's a very weak, very weak feedback loop. Do you see this around the horizon where there's this like hive mind kind of learning that's very rapid? And if so, how exactly does it happen? Actually, I would say that around, will we get a hive mind that learns from all of its deployment, experience. I mean, a big part of that is actually about incentives rather than being a technical

question. So, like, companies aren't going to want to have the model provider learn from all of their deployment because that might just reduce the advantage of their business. I think that maybe the economics of this will pressure not necessarily weight updates to one big common shared model, but, like, kind of, like, modules that get subbed in. So, like, a very obvious example, this is a law, but it might be something else like, you know, there's been a lot of work to try and fit like an arbitrary context length into a fixed size. Like, this is all the linear tension stuff and all that sort of stuff and like cartridges, which are essentially KVKHs trained to be very, very compressed KVKHs to fit in a lot of information. That's another

example of like, you know, something that like companies may be willing to sign up for if that's get subbed into the model and it's not like actually changing the base underlying model itself. So, like, there's many different versions of like learning from, from your data in real time and like the latter ones are not really helping the big labs because they are just these modules. But I think the like the economic pressure will force like the labs to go down that path first before they can embark on this like, you know. Which economic pressure there? Because I feel like even if you have like a bunch of cartridges or laws or whatnot,

you can still just like take all these traces and just like distill this, dump this to the pre-training of like your next generation of longness. So it may be a more indirect form of learning that the big labs are getting and that's obviously still really valuable to them. But I can't imagine a world in which we start off. with like, you know, we're going to just like directly train this one big model on like all the exact data. No, I think it will definitely like go through stages because, I mean, this is assuming there's like one discontinuous event where it's like suddenly we fix like weight updates continuously

and like in practice, I think it's much more likely to be like the cartridges and stuff allow you specialize in deployment. Then you generate traces. You put that in your model. Like three months later, you come out with a model, which is better with this stuff. You specialize it again and you like consolidate it again. And then eventually we'll just like make the sleep faster and faster. So instead of like every three months, we release a model. Now it's like every week and then every day and then every hour and in which point we basically have obviously solved it. Yeah.

And I think this is a good point as well because you asked like kind of how far off the current paradigm we are from being able to do this. We've done a bit of research to this and people have done a lot of research. Like at a really large scale, like when you wash out enough noise and you have large enough batches, like this outer loop process of like putting data into mid-training and creating our own environments, like it does work in like some sort of continual learning regime. But the problem is like when you zoom in close enough at like a micro level, it's like I've

got one model and I'm trying to update it again for like a law firm or something. And I'm trying to do that very continuously with a relatively small amount of data. Like all the methods kind of break down a bit. So like if I sFT the model on just like, you know, successful traces off policy on policy, like eventually in the very iterative regime, like when you're doing like hundreds of like these micro updates, you see both catastrophic forgetting, you see forgetting of like previous information I've learned on top of the base model that was much earlier on. And I see degradation of general like use, general capabilities. You know, on policy distillation seems to like push this horizon out a little bit,

but it still eventually succumbs with the same thing. And RL is not very good at like, it is good at like getting capabilities in, but it's not as good as getting like knowledge in and like just this very explicit knowledge of like, okay, like this person does this at this law firm and like this is a very specific process we find. And you have to pour in a lot of compute to create the environment environments to get the knowledge of Do you think that the fundamental issue here, why you get worse at these other skills or there's forgetting and stuff, do you think it's fundamentally an issue of capacity or it's an issue of techniques? A little bit of both. I think, like, SFT and even like distillate, like, on policy distillation can be, like, way too destructive. Like, the reason RL is so nice is because, like, yeah, it changes a very, very small amount about the model and there's, like, a lot of evidence for why this is the case.

And so like it kind of just like tweaks it in this very, very, very small like loss value to like get it into the right point. But that also then limits what you can do with REL, like how much you can actually change the model. Or so sorry, you're saying like the reason this isn't the winner take all potentially is that it's just like very hard to distill that much information into the base model. Without ruining something in an id or version. Like it's easy to distill it into like a different base model. Like this is where I think it's mostly technique. It's not like it's definitely not like just like there isn't capacity. Like if you had some model, you know, with all this data and you take it.

like nearly the same size model and pre-trained it from scratch with like all of this stuff in mid-training, it will be better. And I think that's a lot of what's happening today. Yeah. And so it's very much like there's, you know, a bottleneck that stops us from just keeping training the same model forever versus just like getting all the data from the old model and like training a new model from scratch. And this is exactly, I'm saying like some combination, like plasticity and like catastrophic are forgetting. And it like, you know, if you just naively train on like non-stationary data because you're adding new data as you go, basically this is messing with the day distribution. So like the old stuff is just forgotten.

I see. We don't really have good methods to like stop that from happening. So maybe at the limit, you're like just bottleneck by retraining the model from scratch with all this new information. Yes. Which, of course, is, like, very expensive. Like, training model from scratch is expensive. But you're going to do that anyways. I mean, not necessarily. I mean, like, maybe eventually, if you have continued learning, you never train any model.

You just, like, just have a model and it keeps learning and, like, expanding, right? But there might be, like, some deep technical reason why that's very difficult because of these, like, plasticity. That's the question. I think we have pushed back, like, how much from scratch we need to do. Like, it is definitely possible now to take, like, the pre-trained base and, like, do very good mid-training on top of that, like, kind of continuously plus some RL from, like, different checkpoints that are later on in the training. And, like, that's looking more like a continual learning, but certainly not the case of, like, you know, take the most recent model,

apply a couple of very small updates and, like, iteratively, like, never lose it. So, but isn't this, like, I'm a bit confused, because isn't this literally what happens during training or something? You just have, like, you have a model that's already gone through so much training, and then you, like, distill some fork that's been further RLed or something. Isn't that literally what happens? And, like, it's still at a large enough scale, I think, that you're washing out a lot of, like, the noise. And, like, you're not just focused on one distribution, which, as Barron said, it's like, you know, that is now a very, if you're just like focusing on one task, right, like, that's not a very, I mean, in the eventual

regime, you'd be doing, I don't know, there's billions of deployed instances. You're, like, doing, you're learning from all of them at once. And so hopefully there's some washing out of noise and stuff from that, right? Maybe that scale, yeah. Yeah, I mean, I think, like, definitely is, as far as saying, like, you can do continual mid-training for, like, a long time, And you can roll back to a checkpoint, give a new mid-training data. But at the same time, like, you can't do this, like, indefinitely. Like, if you just keep continuing mid-training the same base forever, it just, like, get, it does, it's sort of asymptote at some point.

Like, you can't just learn new stuff in that base. And this is why people end up training new bases. Like, otherwise, you just keep mid-training the same base forever. Whenever I finish recording an interview, I immediately brain dump all my thoughts into Slack. Things like, what was the most interesting and what should get cut? This ensures that my editors have all the context they need to start editing the episode. But it's not like these brain dumps have any clear times. stamps and my unedited recordings are many hours long. It can take a ton of editor time to even

find the exact moments that I was referencing. So we decided to try adding a Grockbot producer to our chat. Now, whenever one of my editors posts the rough cut of the episode, Grockbot opens the transcript on its own computer and starts working, usually before I've even seen the message. It takes the notes that I dropped into Slack and it highlights the relevant snippets in the transcript. It also uses a big case file that I've compiled with all my preferences, so it can suggest potential edit. And when it's done, it sends me its top clip candidates so that I can review everything from my phone. This has worked really well. Being able to send informal messages like I'm texting my editor and then having the transcript

immediately reflect my preferences has just been so helpful. Try Grockbot yourself at x.a.i slash bot. Okay, let's talk a bit about data now. So I'm generally interested in this question of how much of AI progress is just explained by data progress. It doesn't mean it will be necessarily hard to automate, but this is that a separate question. So is there some data distribution which if you trained current architectures on would result in a superintelligence that totally dominates human experts across every single field? Are we talking about pre-training plus post-training data like environments as well? Like I think the existence of this is obvious. It's just like whether we can create the right environment.

Yeah, and in the trivial case, we could just train it to output the Python file which trains the actual superintelligence. Like just have them memorize in the way it's like, yes, there's. there's probably like a ladder of oral environments that is possible to construct such that you would get a AI researcher, which is at least as good as a human researcher. But the effort to climb each successive wrong grows like kind of exponentially. And that's going to be the two things that you have to trade off against as to whether like how fast we're going to hit like that final wrong where it's where it's better. I think that's fairly clear. And I think there's like, you know, we're still relatively early in like oral environment.

environment creation. Like there's a lot of asymmetries that we exploit in order to create good environments. So one of those asymmetries, which we've talked about before is like, there's environments where it's easier to go backwards and forwards. And like what I mean by that is like it's very easy to define this like complex data generating process. And this is like this kind of latent variable you keep hidden from the model. You can generate like arbitrarily complex like environments and the model has to do a lot of like irreducible like token spend and irreducible work to figure out what that data generating process was. There's asymmetries in terms of like, you know, you can inject information from the real world. So, like,

Anthropic finds a bug through, like, you know, tens of thousands of human and LMLs combined. And, like, turn that into a very, very neat environment, which is a single LM could theoretically find within, like, you know, a few million tokens. So, like, there's all these asymmetches which we're cherry-picking and, like, we're counting on, like, kind of this task horizon generalization. But I think, yeah, again, there's just going to hit diminishing returns at some point. Like, at some point, it's diminishing returns and how hard it is to create those environments in the first place, like, coming up with them because you can't necessarily just, like, have these really, like, these processes where it's easier to go backwards and forwards. Like, you actually

have to sit down and construct, like, something that looks, like, with humans, like, you know, a long enough time horizon, like, it's going to be a really complex task to create. And then it's also going to be, like, the compute and time bottlenecks for the agent to actually do those tasks. So, like, I think you're just going to start seeing this, like, curve to fly them now. I saw something about how someone fine-tuned the Taki model, which is only trained on data up to 1930 on this modern coding agent data. And it did better the Claude III opus on sweepbench. So this model that has no knowledge of code whatsoever

can be fine-tuned on a moderate amount of data and behave better as a coding agent than this much larger pretrained model is pretty crazy. And it kind of shows you that once you have an example of the right expert behavior, it's actually surprisingly easy to like copy that into a like relatively weak model. Yeah. But a counter example to like that kind of is there.

There was a paper recently where they trained it up to like fifth grade maths. And like also like primary school like English and stuff. So it was like a decent language model. And they tried to RL it to do like, you know, late high school and college maths. And the gap was just too large. Like they couldn't get it to climb at all. Like but if you did like successive wrongs of like, you know, you're seven maths and then you're eight maths and so on. Like you could obviously climb to.

to year 12. So like, again, it's just like, what is the distance between the wrongs on those letters and how hard is it to create? Yeah. And this just comes back to like the RL signal problem. Like RL is not very good at like exploring right now. And so if you if the model come like it in like 128 world out, it's very unlikely to get signal to like progress. And this is why like in RL we need like curricula, whereas like in pre-training we don't because like it's, that's not a problem for pre-training at all. Yeah. And again, pre-training data is different to post-training data. And I imagine as we continue on like yeah, humans will be, involve less and less, but that doesn't change the fact that you're bottlenecked on, like, how much signal you can extract from the real world. So, like, there's a lot of signal on the world, and that's true. Like, you know, there's people doing, like, spreadsheet tasks. There's

people doing, like, legal tasks and all this sort of stuff. But, you know, the capability frontier of where the models are at now, like, how many bits in the world are actually, like, really relevant to, like, improving the model's capability? Like, you know, how many new math's problems are being solved that, like, couldn't, they'd just be on the reach or grasp of the current models? like how many new coding problems are being created or solved that would be on the range of the current models. And like, I think that's why the diminishing returns kicks in because even the world as a whole is not giving you the bits that are useful for tipping you into the next basin of capability. Yeah, I totally agree with this. It's like really a question of like where the signal is coming from.

And so like the signal doesn't, you know, in pre-training the signal is like already in common call, right? Like for the tasks that you care about in pre-training, the problem is there's not just like getting signal at all. It's like filtering out all the noise that exists. Yeah. And that's quite an automatable process. But like as the models get better as we enter mid-training and post-training, the signal just like doesn't exist anywhere in the original data we have. Like no amount of filtering will like get this, you know,

there's no like hidden proof of like the Millennium Prize problem sitting in Common Core. We can just like filter until we see it, right? And so like at that point you have to get bits some other way, either from humans like directly like asking them to like write out their reasoning or like by like creating environments where like humans decide like what environment should be created, what the objectives of these environments are, or like, you know, some kind of like training on like, like the human data that exists in deployment.

Like you have to get the bits from somewhere. Yeah, yeah. There's a question of how much of the progress in pre-training is being driven by data. I did this investigation with Jerry Hahn, who's a student at Princeton, where we basically trained all the recipes from 2019 till now, pairwise with all the data sets from 2019 to now. So you change like GPT2 on the newest dataset, like ultra fine web, and you train Delphi, which is the newest training recipe,

or the open source training recipe on the pile or some old data set. And you do the whole grid. And you see the getting to some level of capabilities, how much less compute does it take across this grid? And you see that the data seems to explain like 9x of a compute efficiency gain. But the architecture improvements explains like a 3x computer efficiency gain at a very small scale. And so to the extent that that is true at large scale, that most of the pre-training computer efficiency gains

are coming from better data, how much can that continue? Can you keep just filtering data more and more and building more and more synthetic data until, yeah, do you have a sense of how much this kind of pre-training progress can continue? I think my prior is that, like, again, the low-hanging fruit is, like, somewhat exhausted

with, like, we got the internet as this big block, and, like, it's not like the internet is necessarily, like, growing at the same, like, rates, all the useful stuff and the internet is growing at the same rate. So, like, we've probably got, like, a bunch of, like, 0.1% loss drops to go, but, like, not definitely not as many as have currently occurred. Yeah.

But, like, that's also really interesting that, like, you know, you find this, like, what, cumulative, like, 27 times improvement across both. I think, like, it was epoch or someone who estimated, like, three times a year since 2019, which would imply something, like, you know, three to the seven, like, over 2,000, like, times improvement. So, like, where's that missing, you know, 100 times or whatever coming from? Like, that probably gives you a good signal of, like, how much. as this is like post-training.

I think the explanation has to be that a lot of the computer efficiency gains are scale-dependent, and we were studying at extremely small scale. And that raises a question of do the data computer efficiency gains or the algorithmic computer efficiency gains have more scale dependence? I don't know if you guys were prior on that. We just didn't have enough computer to investigate that question. I mean, like, just naively, right? Like, theoretically the scale dependence of the architecture is, like, fairly well known.

and like you can fit a straight line to it whereas like I would have no idea how to do that for like combining pre-training plus post-training data and mid-training data. I don't know. I feel like data is actually more like more important with scale. Like I feel like architectures are kind of like a one-time like an architect.

I think like combining like saying just like an X percent efficiency again is kind of misleading because like what an architector does is like let you reach like a qualitatively new regime which you couldn't reach with the other architecture. And then within that regime obviously the data is like the primary thing determining it. But like you know, if we say, didn't have like even like GQA we're doing like full attentional day we wouldn't be

able to do like a million it would be like ridiculous expenses to a million context and then like because of that we couldn't we could never use the data which was like actually at a million context and so we couldn't get these capabilities even though like if you just do a naive like how much does this do at like 2k context where the architecture isn't unlocking anything then like the data you know that will look much more important than in some sense it is right it's unclear to me that these things are like really just like multiplicative gains in this way I see. So then what does this take away

for the scale dependence of data? So I mean, on scale dependence, I think a lot of the mid-training and post-training data we have now is actually gets better with scale. Because a lot of it, like, the very long context, horizon environment stuff really requires, like, big models to be able to make use of them. And, like, this is not, you know,

if you try and train your 100 million parameter model on, like, three bench traces, it's not going to get anywhere. Like, it's not going to show you the same kind of improvement that you would get if you train, like, an actual sensible size model on it. Yeah. And like, it's hard as well now because so many of the architecture changes.

Like, you look at like Kimi, for instance, like, or Deep Seek. They're doing these architectural modifications with not just like dropping the pre-training loss in mind, but like, for instance, how the models are going to be used in the real world. So like, yeah, the inference efficiency, like having some form of compressed attention in the deep-seek models is not necessarily geared around. You know, this is fundamentally like a pre-term improvement. It's just like, okay, we're considering how the models are going to be used. Right, right, right. One question I'm curious about to understand the future is how parameter scaling will go.

as we're getting into more of an RL-heavy regime. Like, I don't know how fast historically. Yeah, you can look at sort of open-source architectures and see how fast parameters have been scaling, and maybe it's like roughly 2X every year for Frontier open-source models. And to the extent that, like, even Frontier-Closed-S models

have, like, 100B or 200-B active parameters. Do you think that keeps 2X in year-over-year or another in an RL regime where you also want to conserve compute on rollouts? And also, maybe there is, like, a threshold, the fact where you have enough capacity and at that point, increasing parameters arbitrarily doesn't matter as much. Do you guys have a sense of, in 2030, how many active parameters will a frontier model have? Yeah, I think for the next few years, we're going to be, like,

because we're so focused on doing longer and longer horizon rollouts for RL, where, like, inference efficiency matters a lot. Yeah. It feels like the moles aren't necessarily saturated on their ability of that, where the bottleneck is still the environments. And so we might see, like, a little bit of plateau. Like, I have a feeling that, you know, like, us and the GPT models are much smaller than like, you know, the 10 trillion parameter range that people are talking about. Even just naively comparing to the open source models, you can probably back out that conclusion.

So, yeah, probably for the next few years, I wouldn't imagine a huge growth in the number of parameters. But again, like, there's so many different things to trade off here. Like, you decide the size of your model based on, like, how much pre-training data do you have and then, like, the difficulty of the oral environments that you've got to train on. And you ideally want to, like, get to the optimal point. where you know you can get like a decent pass at one or something on like the the hardest environments you have and like it wouldn't make sense to like make a bigger model pass there because

then you're just paying like much more inference swaps than you need to um so there's a lot of inputs to this like depends on how quickly you know like mccore and then in-house these these guys can scale up the complexity of the oral environments they're training on i would expect the models to keep getting bigger just because people are scaling up compute and the GPUs are getting bigger but I would say exactly how much they get bigger depends a bit on the scaling laws in non-obvious ways. So one thing is that I think, like, data efficiency is going to be a bigger driver than compute efficiency of, like, the exact architectures people use. Now that we're getting to the regime where we're sort of running low on, like, high-quality

pre-training data. So that might affect how sparsy you want to make the model. And then I also think we don't understand sparsity. out well. And it's like parameters are a different resource than active parameters, but it's, and like, sparsity has definitely increased a bit, but it's not clear that it's going to keep increasing without bound. There might be some kind of sweet spot. There's an argument that sparsity should make data efficiency worse because you might have to learn the same thing on multiple experts, though that's debatable. So I think we don't, I don't think we, I don't think we,

we have a good enough theory of scaling laws that we really understand why sparsity is helping and to how much it'll help and if that'll like plateau at some point at a certain level of sparsity. Can you spell out exactly what the implication of data efficiency would be on? So it sounds like you'd say, well, there should be less sparsity, but what are the other implications on parameter scaling? I guess just that the scaling law, you're not necessarily looking for the most, you're not trying to optimize computer efficiency.

So you have all your choices you can make on the architecture, and each of these gives you a different scaling law. And then, like, traditionally you would look at some kind of envelope based on compute. So you would look at performance versus compute and take the envelope of, like, the best models. But, like, if we're making that decision based on data, so it's like, yeah, we're sort of assuming we can spend a lot of compute. and like we're sort of data is on our x-axis instead of compute, then we just get a different set of Optima

or a different set of models that are on that frontier. Yeah. And I also don't think that we've necessarily like, you know, doubled like the size of the models every year for the last few years. Like the people have been training like one trillion primordid models for at least a few years. Like there was even an open source one called Falcon, but, like, Liam from Periodic Labs, like, I think post yesterday on Twitter about how, like, an early experiment at Open AI was, like, training a one trillion parameter model that was very, very sparse.

Oh, yeah, that was what they did before Open AI. That was, like, at Google the Switch Transformer, yeah. So, like, it was, like, you know, very, very good at, like, knowledge, but terrible at reasoning because it was so sparse. And so, like, yeah, it feels like we've been playing in this, like, 100 billion to, you know, up to 2 trillion parameter range for, like, at least a little bit. And, like, it certainly hasn't been this is, like, nice linear increase, yeah. I mean, I feel like there's two things. So as Charlie was saying, like, inference efficiency is super important for RL relapse.

And so like this will really push down active parameters quite a lot. And then I think the total parameters really depends a lot on the hardware as well. So like you really need to get like very high memory bandwidth and like VRAM size to like actually be able to serve like multi trillion parameter models. And so like, you know, right now, you know, people still are using a lot of like H-100s and stuff. And so as everyone moves to GPs and then Verirubins will get like more actual, like the ability to scale and like actually serve and like do like large

our influence are like different at larger scales. The data question I think is interesting because naively larger models are much more sample efficient in like the actual data points. And so like even if you're like not saturating the model, it's still better to go bigger because like the models will larger models generalize better and like get to a better loss for the same amount of data. And so right now I think we kind of have a lot of data and like that's not the constraint rather than computing.

So we're having like small models which are like very influence efficient. But if computers are along the bottom like it might come back to larger models which like under, sort of undersaturated, but like they have this generalization ability because they're much larger. If you just look at like the basic chinchilla scaling law and you just maximize out parameters, it actually decreases the amount of data you need to get to the same loss very little. Yes. If you go to infinity on parameters, the amount of data you need I think goes on less than 10x, just because of the nature of like the power line.

But we're now on the way too much data side of the chinchilla laws, right? So right now we overtrain more of the chinchilla. And so we could easily go back to a point to which as we're running out of data, we move back to like the chinchilla optimum point, or even like a bit on the overtraining, you know, under training model side. But surely like even with these new chips that come online and stuff, like we're just going to be so compute bottlenecked for the next few years

that that won't necessarily be in a case. I mean, this could well, yeah, this depends on like the ratio you have like training and inference compute really. It's like if you're super bottlenecked on data and not on compute, you should go bigger. If you're super bottleneck on compute, you should always go smaller. And then like, yeah. But you can also use computer generated synthetic data.

So it's like one of these very hard things to do. predict. Yeah, I think part of the reason it took people so long to figure out the scaling laws in the first place was that if you don't get all these things right, then you don't get such a clean relationship. And like the beautiful stray lines on graphs, like hide a lot of complexity on how you have to make sure, like to scale every hyperparameter the right way, or like parameterize your optimizer in a way that scales and where you don't have to change your hyperparameters as you change the model size. And bugs have their own clean scaling laws as well, right? Like, you know, like with Kaplan for getting the cosine and

ailing thing or like even just like not considering embedding parameters, I think. And so that messed up the estimate at smaller models because embedding parameters are a decent size of the model. A bit on RL. So I feel like a year ago, a lot of people were making this argument that RL will not be super successful at scaling for models. I think John you wrote a research paper where you were pointing out that models learn one bit per episode when you RL they basically learn

did I get the answer right or did I get it wrong? And I wrote some blockposts earlier this year I was like it's even worse than that because when the pass rate is low and the model is very unlikely to get the answer right.

It learns almost nothing at all from an RL episode. But I look at the models today and they seem pretty smart and it seems to be the result of scaling up RL. Baron, you had a post I think a few weeks ago where you're trying to explain what's going on.

But why has RL been more successful than one would have I mean, so I think the success of RL comes down to a bunch of different things. So first, I think what is slightly underestimated is actually the mid-training. So an awful lot of what we see as successes of RL actually comes from like very, very good mid-training data, which is basically where we're like essentially doing pre-training, but on like synthetic reasoning data and like the kind of environment that like get the model warm started for RL.

And so this actually takes the model like almost like 80% of the way to like the final RL checkpoint often. And then what RL does on top of that is it does like a lot of, essentially tweaking to the policy. And so this is one of the reasons why it doesn't need like as many bits as you would naively think. It doesn't have to learn all of these behaviors from scratch. It needs just like a few bits from these episodes, which you do get.

And then the other thing that I really point out in my blog is that these bits are actually extremely high signal compared to like regular like pre-training, which is why you need RL at all versus just like sifting on like successful reasoning traces. Because it's exactly the bits about how to get the answer right. Well, there's two things. So yes. One, it's exactly the bits about how to get the answer right. but like this is not exactly how you think of it,

because in SFT you have a trace, right? You have like a bunch of math reasoning and then the answer at the end. The bit is still there. Like you're still SFT on the answer token. So that bit is still there. What's important is that the objective ignores all the other bits. So in SFT, you like have like, you know,

to try and match like the exact reasoning token to the model producers. So you're essentially getting like too many bits about like the exact way this other model you're training on reasons. For RL, you only get the one bit. And that means that like this, that signal is not drowned out in the noise of all the other bits the model has.

And so that's what really like, it's really super dramatic increase into the signal to noise ratio during training, which is why like RL is like so dramatically efficient in terms of steps. Hmm. And if you guys have thoughts on that. Yeah, I, like there's been so much debate

about like what RL does to the model versus like, you know, mid-training or SOT or whatever. And like, you know, everyone talks about how, you know, parts of one will go up, but parts of 256 will go down. Like very rare, correct reasoning traces will be like downweighted and kind of like outweighed by gradient signal from like easier kind of reasoning traces. And I think the simple like way to view RL now is that if you have a large enough,

like a large enough amount of compute to sample a large enough group size, such that your probability of getting a bunch of correct answers is like pass some, like not insignificant probability, then like it will be upweighted. And like to, to Beron's point, like basically mid training and, you know, more pre-training, like the pass at one, the starting point for RL, like scales in the long number of pre-training tokens. You can ask a very basic question. I guess that answer makes sense, and maybe there's empirical research which shows that this is what's happening. But then I just look at the models themselves, and I don't know what's happened. Maybe you can give me a sense of what is the

basis of the AI progress over the last year. But maybe it's just upweighing the policies which we're going to do the correct thinking anyways. But it just seems like qualitatively, the models have gotten so much more capable. And anyways, maybe there's nothing to, there's no inherent contradiction there. But how do we square the relatively small impact this take would imply that RL would have

from the actual qualitative capabilities the model seemed to be gaining? So, like, one thing I want to point out here is that it doesn't necessarily imply that RL has a small effect, right? Even if you have a few bits and you only change the parameters a small amount. Like the actual impact on like function space,

the model lands, like the input to upper mapping can still be like super dramatic. You know, even if it's like, even like one bit can change like your function space a lot and it can like rule out like half the hypothesis space, which is huge. So like I don't think it's necessary with the case it's like small amounts of bits, small amounts of RL.

Once you're starting from a really good point, means that like you don't have dramatic impacts in behavior, at least like not necessarily. I think it comes down to two things. I think the first thing is that everyone was hoping that RL would like generalize this reasoning across like all these different domains. And I don't think we necessarily got this like horizontal generalization.

Like just training on math doesn't necessarily make you the greatest code. Like you do have to do RL on code environments. I think what we did get though is like Horizon generalization. Like the models just learned how to use more tokens for longer and still make progress on some sort of task. And so like you can train on environments where they get longer and longer and longer and then put them into a completely new environment. And yes, they may not have generalized the reasoning patterns

which allow them to do well in that environment. but they've at least generalized the ability to continue on that task for longer, which is correlated with success. I think there's a paper called EdgeBench, which show that the rate at which models can work for longer is doubling every three months. So that's a clear evidence of generalization. And I think the final way to think about it is, like, in pre-training, there's this idea of quanta.

So you have this very smooth pre-training loss curve. And when you actually look at what's happening in the model, the model is learning all these very discrete tasks, and there's like all these like emergent points. But there's like kind of a phase transition. Like it didn't have induction heads. Now it has induction heads. And there's like tens of thousands, millions,

probably like hundreds of millions of these things. And you average them all together and you get this very smooth loss curve. I think like to an extent, like a similar thing is happening for RL. Like there is this very slow out of loop, as Berrin mentioned of, you know, we will train a model and then REL and then like the next kind of model iteration of training, we will dump a bunch of these synthetic reasoning traces into the mid-training data. like we're kind of hitting all these quanta for all these different tasks. And like on an individual task level,

it may look like a phase transition. And like you're suddenly going from like a 0.5% pass rate to a 90% pass rate on like a particular like finance task or Excel task or whatever. But you average all these things together and plus the horizon generalization, you kind of would go, wow, we've got like qualitatively better models. Yeah. I mean, I think a lot of this as well is just like, I think RL does generalize a bit. Like suddenly you get like some transfer between like math and code or like puzzles and math and this kind of stuff.

Also, just like the share amount of environments, I think the people are targeting is just vastly greater. So, like, before, when you try to do some task, which, like, you do in your daily life, like, two years ago, like, the labs wouldn't really care about this. They wouldn't, like, train the model for it. And now, like, it's just so much broader. They have a lot of environments targeting this specific thing. Earlier in the conversation, we're talking about R.R.L. in the context of causing this entropy collapse or just, you know, concentrating probability on solutions the base model were already done

and causing relatively sparse updates in the policy. But when I think about I think about I think there's also another story about RL which is going back to the Atari games and then Alpha go coming up with Move 37, the super creative move

that because it was never initialized on human data it can think in ways that humans are not even thinking and come up with extremely creative solutions. Yeah, do you have a sense on when we should expect or if we should expect RL on LLMs to result in things like Move 37, just extreme

creativity, even beyond human creativity, because there's just de novo, de novo initialization of intelligence? I mean, so a couple of things here. Like, first off, I think that the AlphaGo is using MCTS, which obviously does, like, more exploration and, like, stuff than regular positive gradients. But I kind of also think that, like, RL doesn't necessarily, like, reduce the creativity. And, like, I mean, even if we look, I think, you know, this is obviously qualitative, but if we look at, like, the, you know, the open air-hanging phase incident, like, these models

we're coming up with like multiple zero days at a time to like break out of the sandbox and like this is clearly like some level of like Move 37 creativity, I think already which we just get from just like the general generalization properties of the alarms. Like I don't think it's definitely not the case of like RL's like totally destroying their like entropy.

Yeah. Especially on long horizons. Yeah, I mean one thing that people call creativity is just solving hard search problems. So and so that's like like Move 37's obviously an example of that or like writing some kind of poem that satisfies a ton of different constraints.

So that's something AI is obviously going to be extremely good at if trained for it. Then there's another way in which the models, like the diversity of their outputs is a lot lower after RL, and they sort of develop these ticks. And even though the models seem like they're good at writing, when you do some kind of like distributional analysis, you find that, like, they're reusing certain themes, like, all the time, and they're using the same character names all the time. So there's actually, it's not like you're getting the same kind of diversity that you get when you, like, from human authors, you're sort of getting one really good, like, style.

So I think that, like, that kind of diversity has definitely been, like, cut down by RL a lot. And in fact, no, yeah, since we were talking about, about distillation earlier, that's sort of something, yeah, one thing that's happening is that so many people are distilling mostly from Claude that like all the open weight models write the same way as Claude and use the same, like have the same ticks. So this seems kind of concerning to me that we're having this like this monoculture emerge. Yeah. Again, I don't think this is like fundamental to RL as like a method though. And same with distillation. Like even with distillation, like you're just training on the data. It's like,

just because your data is not super broad, that doesn't mean like the training method itself is somehow wrong. It's like a problem with the data. And I think a lot of, for instance, like the RL, like entropy collapses basically due to like exploitation of fairly simple like verifiers when you don't have like a huge diversity of environments. Because like for instance, like the writing, I think the writing is presumably graded by some judge and like the judge has some specific ticks

and like the model is learning to award hack the judge. And that's why like it collapses. But like this is really a problem with the judge. It's not a problem with like RL in general. Okay, super rapid fire predictions about the future. So I want timelines on the following couple questions. By when do we have models, which you can, here's what it feels like to a user. You basically hire them as a drop-in remote worker for all kinds of white collar work,

not just coding, but I don't know, video editing, law, paralegal, etc. Like it's like literally an actual remote worker. Like full computer use with like literally a month of seamless learning and operation and executing on like complex projects and required interacting with other people, et cetera, et cetera. It's like everything a human worker could do over a month. If you like mandated to use like a browser or whatever rather than like these, again the firm setting up the information to be like programmatically accessible like maybe a couple of years. But if it's not like browser based, like you can send Slack messages. They can do all this stuff. But I'd still probably say it around a year.

Yeah, I mean, I would say maybe like, for the like full generality, maybe like three years, but I think to Charlie's point, we will end up with like a lot of people like making their organizations easier for the AIS to use. And so you get like 80, 90% of the way there before that. So but the thing that's the, the diff between one year and three years, there was just literally like, I think there's going to be like a long tail of like miscellaneous stuff, which like some human can do, which like will take the models like quite a while to do. Yeah, like, are you thinking of sort of computer your stuff or like basic cognitive

capabilities. I mean, I think this is this really comes down to a question of like how quickly can we solve this kind of like online learning and like whether we can like get like 80, 90% of the way there with like compaction and like writing files to yourself and stuff. And like that's my big uncertainty. I really don't know. And another like maybe an example of something that wouldn't be good at is like, you know, if I have to like yell at someone to get something at work or like really push someone to get something done, like the model isn't just going to do that. It's just going to be too nice. Yeah. I'd say there's a wide variation in quality of human remote work. So if you try to hire someone off of Upwork to do a software engineering project, there's going to be a huge variation.

It's often quite hard to get them to do a good job or like pay attention and all the feedback you're getting. And like I would guess that in some cases it'll be worse. Like the pre-AI version of this was worse than what you can get now from existing AI. So I think it might end up being a little complicated. Because maybe to some extent we already have this, like for some, like, not so high quality of work. But then, like, then it's obviously, like, we're not matching human level in certain, like, higher quality, like, forms of work. So, but I basically agree with Charlie and Barron that maybe, yeah, we'll have some version of this in a year. So it's, like, okay.

And it will be able to do, maybe we'll have that form factor. and it'll be able to do some things really well, some things not so well, and things will be improving from there. Like, we ship the goalpost based on the very long tail all the time. Like, I feel like he used this example before of like doing your taxes or something. Like this year, I literally just told Codex

to like go get everything I needed to do and send it to the accountant. And like there was this massive list of stuff. It had to use computers to click through and download some stuff. And it did it. It was like fine. It was perfect.

So like, I don't know. A lot of this stuff it can already do. Yeah. Okay. Give you 10x total productivity uplift. Basically, if it takes you a year to make a breakthrough now, you make a breakthrough every month. I think I would just refuse to give you a scaler on this. Like, we might already be past that in some, like, types of work. Like, let's say you're just trying to prove, you're trying to do, like, certain types of math.

Oh, sorry, but for you as AI researchers trying to make, like advance, you know, the state of AI research. It's how much our, like, AI research were sped up. Yeah. Or uplifted. Somewhere between five and ten years. Oh, really? Oh, yeah, that's far away.

Well, you think it's longer than, like, for general remote worker? Yeah. Interesting. I think you're probably, I think I'm realizing you probably have very different definitions of a fully general remote worker. I could have specified that way. Yeah, this is true. Because I mean, like, yeah, because obviously, like an AI researcher can be a remote worker.

And so, like. Yeah, no, I'm picturing, like, you know, normal white color work over the period of a month. Yeah. I think it starts to diverge a little bit past a month. They're a very competent white collar worker, but not necessarily like a super creative researcher. I would say like two years. Two years? Ten X. Okay. How about you, Byrne? I can kind of see that actually because like it really is just like, right now it's already like definitely more than 10x of like coding stuff. And so it's like if it can do it even like one or two loops of like experimental feedback, that would actually be massive already.

So 10x uplift of AI researchers within two years. If you just plug it into like a very naive model of like AI progress and how much is coming from AI researchers and they're like there's like a 10x increase in their productivity. Yeah you you have like radically accelerated pace of AI progress starting two years from now. Yeah I mean I think like this will mean that AI progress doesn't get bottlenecked on like AI research's ability to run like small experiments against barhontanect on other things. Of course, of course, but it just like happens 10x faster. For sure. Yeah. Which is a huge deal. And that also like helps the next thing which makes you 100x speed up happen sooner, et cetera.

Yeah, I'm happy to stick with longer on that one. And what's the crux? Like my capacity to absorb information and make the, like, beige and optimal decision on the next experiment. Makes sense. Yeah, I mean, I'm assuming that, like,

you can delegate some of this to the AI. So, like, the AI is becoming decent at, like, deciding, you know, it's run this experiment, just got this result, it runs, like, the next experiment. And then if it can run, like, two or three experiments in a row without, like, crashing, then, like, that is actually a big update,

like, uplift. And, okay, final question. an AI which is which dominates top human experts across every single field of work that can be done over a computer so not only AI research but all cognitive work

and not just like short horizon work but literally if it takes like three years or something they will still do better than humans so this is basically just like ASI yeah okay I would say like three or four years the fuck I mean that doesn't seem wrong I mean I would say

like AI is obviously being more, getting more attention. So it's like one of the harder things, but it's like a lot of energy is being put into it. And it's also like not one of the hardest things for AI because it involves a lot of code and math, which models are really good at. Maybe for things that involve like 3D and like spatial stuff and physical stuff, I think that will take a little longer.

So especially if it's not like, Yeah, if it's like mechanical engineering or something and it's not getting like the most attention right now, that might take a little longer. But it also just includes fields where there is relatively little data because of the nature of the field. And it has to like learn that data on the fly. So for example, it has to become superhuman at like being an engineer at TSM or something.

Oh yeah, so you would have to assume that like the onboarding, yeah, you can give the AI the same onboarding material. And yeah, then there's some like, something has to be solved about like sort of longer horizon learning or yeah yeah I'd say five to ten basically yeah it's is you think uh automating AI research is like ASI complete or something yeah I think so yeah I think there's so many things in the world which like even if you have some sort of memory system external to the model and even if like context length grows a little bit like there are just fundamentally things like even if you could research the information or write

notes yourself, like you'd need more than a bit of context field today. Yeah, I mean, I kind of agree in like the five-year range, at least for like the stuff that, like, labs are focusing on. But I think like there's going to be a long tale of stuff, which like the AI could theoretically go out and learn about, but like no one has bothered to do it. And like the computer isn't being allocated to that. So that might take longer for like literally every single human expert. That's right. But by this I also included like the ability to learn as fast as a human, a new domain. I mean, I think that's not necessarily necessary actually. Because like

the AI will have vastly greater experience. than like any human. Right. Thanks for doing this, guys. I feel like this was a great format for getting different experts to disagree and debate and discuss things together. It was very productive. Cool.

Thanks having us. Thanks. Thanks for having us.