Noam Brown – Agent swarms, alignment, & recursive self-improvement

Today, I'm chatting with Noam Brown, who is a researcher at OpenAI. He was one of the foundational contributors to what became O1 in the reasoning models, and now he's working on multi-agent systems. Speaking of which, you guys announced last week that you solved one of the Millennium Price problems with a system of 10,000 different AI agents that spent 130 billion tokens over 88 hours. One of the reasons I was interested in talking to you, is I think you were in the first people maybe two or three years ago, who was thinking about how the reasoning models would allow us to see into the future, because if you scale up inference compute, you can see what the base capabilities of the models will be a few years in the future. I feel like you're in a similar position now to help us understand what future capabilities will look like, given the enormous scaling of agent sizes that we can do right now. The way I think about it, when you plot the performance of these reasoning models with test-time compute on the x-axis and performance on basically any reasoning benchmark on the y-axis, you see a very clear pattern where the longer these models take to think about their answer, the better they do. This is like a very natural thing. It's the same thing with people. If you're taking the SATs, you have five minutes to go through the entire exam. You're not going to do very well. If you have five hours, you're probably going to do a lot better. The AI models are pretty similar, and they'll spend that time doing this model log to themselves, figuring out going through different cases, ruling out different possibilities, building on some of their previous discoveries. The problem is that as you push that further, you hit a latency bottleneck. You don't want to sit around for three years waiting for a response. What you can do is what a lot of people do is they paralyze. They just get a team of people. If you're going to found a company, you want to get a group of people together so you can go faster. The same thing with these AI models that it helps to just have multiple agents working on something because they can just go faster. So multi-agents is a way of scaling test-time compute in parallel instead of purely serial. It is less efficient because it's not like a single agent has all the context to itself, but it is a very effective way of scaling test-time compute if it's done well. >> I'm going to ask a bunch of naive questions because these systems, so this is an unruly model. We haven't publicly seen how these systems work. I just have a bunch of ways in which I'm confused about what the quality of properties of such systems are. >> I am shocked by the scale of cognitive effort that you can concentrate in such a short period of time. So if you think about what 130 billion tokens are, if it was a single human thinking as full-time job, stretch back-to-back, 130 billion tokens would be a human thinking for like 4,000 years, eight hours a day or something, a working normal work week. So starting from like ancient Sumeria up till today, a single sequential human thinking that long, concentrated in 88 hours. I feel like qualitatively that is a super-important consideration. And I'm surprised that there isn't a bigger parallelization penalty that you can just have 10,000 agents collaborate. And because maybe the agents are better collaborating than humans might be, they're going much faster that they can actually productively collaborate at such a big scale. Or maybe there is a big parallelization penalty. >> Yeah, let's talk about the parallelization penalty. Let me talk about the qualitative stuff. Because the truth is that we don't have very good science on multi-agent scaling up to this kind of scale. >> Yeah. >> So when we released 5.6, I think that was the first time that we had a proper multi-agent system in our models. And we actually did in the blog post show some plots of the scaling performance of multi-agent systems. Because we have it as an option, it's ultra-mode. And the default is four agents, but you can set that to higher. And in the plot, we show, okay, here's what the performance looks like on some benchmarks for one agent, for four agents working together, for 16 agents working together. And what you see, and it depends on the benchmark. But for some of the benchmarks, basically, if you have four agents working on the problem, it is done twice as fast. So you're basically paying, because there's four agents working for half as long, you're paying a 2x more to get an answer twice as quickly. If you go to 16 agents, you see a similar pattern. It's a little less efficient, but you continue to see that performance. >> Is it a linear serial time speed up or a sublinear speed up as you increase the number of parallel agents? >> I would say it's slightly sublinear, though it does depend a lot on the problem. So math, for example, is quite parallelizable. It's not the most parallelizable thing, but it is very parallelizable. I think web search, things like doing a deep research report, we have to look through a bunch of sources. That's extremely parallelizable. I suspect that something like writing a novel would be very unparallizable. So you would probably not see a big benefit from having 10,000 agents working on a novel together. In the same way that you probably not have a big benefit from having 10,000 people work on a novel together. So the performance does depend on the domain. We do measure it up to 16 or so agents and are in our published blog posts. The problem is it's very hard to push that science to like 10,000 agents, because it's just so expensive. >> You guys just dated over like a week? >> That's one data point. Like, we don't know how long it would take a single agent to solve not good skills. Because we haven't done that experiment yet. And maybe we will, but I mean, that's also only one data point, right? And we want to do, if we want to do a thoroughbation. >> Interesting. >> It's actually, the experiments are just too expensive to go to that scale. So we have to do some kind of like methodical science about what happens when you go to like 64 or 128, 256, or something and get a sense of the behavior. But it's going to be very hard to push that all the way to like 10,000 and know for sure what was the benefit that we actually got for using 10,000 agents versus 1,000. >> Yeah. >> One thing I want to make clear is that the effort to get, it's not like they have to get a Millennium Prize problem. This was not due to multi-agent. I wouldn't even like attribute like 10% of the credits to multi-agent. Like, the reality is we've trained a opening eye has trained like a very powerful model. And we can get that model to operate over very long horizons. We can get it to think in parallel. But at its core, the reason why we're able to do this is because we just have a general purpose, very strong model. And I think things like multi-agents are flashy and are new and that probably gets disproportionate credit for that reason. But the core reason is like, this is just a very powerful model. >> So the generalization. >> Is actually quite shocking to me. The US systems, I mean, I don't know how these systems were trained. But presumably they were trained how our training happens. You have a bunch of checkables, synthetic problems. You do a bunch of our elegance them. And nowhere in the training process I'm guessing was like the model solving anything as ambitious as a Millennium Prize problem. But the generalization was strong enough that like you could have this like much easier verifiable problems generalize to this much parallel effort on such a hard problem. >> I think that is true. Like first of all, we do train the model on very hard problems. So there is definitely a gap. Like we see if we train on on some kinds of tasks. Like it's able to do tasks that are more ambitious than that. There is an interesting challenge that as the models become smarter and smarter, the kinds of questions we can ask them, just a lot of them are too easy. And it's hard to challenge the model. And I do think that that's going to be an interesting. Like if I had to make an argument for why you might not see AIs, like LLEMs go the same path as AlphaGo and AlphaZero and all these kinds of like game playing AIs, it might be that this kind of problem that in things like AlphaZero where you have self-play, you have an infinite curriculum. You're always playing against an AI that's like equally strong. Whereas for things like training in LLEM with reinforcement learning, at least the ways that are out there right now, you give the model a problem and you ask it to solve it. And if the problem is so easy that I can just solve it in a second, it's not really learning anything. So if we run out of problems to ask it, that challenge it, then that is a plausible scenario where actually like, "Okay, it becomes much harder to make progress." Now, I do think there are ways around that. And so we haven't really hit that as a wall yet. And I think that if it ever became a series problem, there would be ways around it. But it is like a plausible scenario. Yeah. And just for the audience, when you're referring to like AlphaGo or AlphaZero, you're saying like getting superhuman, relatively fast after achieving human level performance. Yeah, I mean, if you look at the trajectory of game playing AIs like go, they, within a span of like a year, went from beating a European chess champion, being like, I don't know, like number 50 in the world, to beating the world champion, to being unimaginably orders of attitude stronger than any human alive. Yeah. And it's possible that domains like math, we see a similar trajectory. But I think there is a very plausible scenario where actually that doesn't happen. Yeah, yeah. So I want to understand if in six months, people will have access to multi-agent systems. How should one model what it is like to collaborate with or hire a multi-agent system? Yeah, I should start by talking about how these multi-agent systems actually work, which is I think a very different way than a lot of multi-agent systems in other AIs. So a lot of people that have approached multi-agents for things like LMs, tend to take this like very scaffolded approach, where, you know, for example, there might be a coordinator agent that delegates work to a bunch of children and gives them a task and the children work on it and then return their answer. And this seems like a very sensible setup, very sensible scaffold. It definitely helps. But there are a bunch of limitations with these kinds of setups. So for example, if in this setup, you have a coordinator that's sending tasks to children, the children work on it and then return their answers. Well, what happens if two children are given similar tasks? Can they talk to each other? And usually the answer is no. And that's very inefficient, right? If you're given a task and it's actually really helpful to talk to somebody that might know an answer to a question that you're working on or a part of something that you're working on, it would be really helpful if we could just be able to ping them and say, "Hey, can you help me out with this thing?" But a lot of systems don't have that set up and adding it just increases the complexity and if you lead to the scaffold that you have. Another thing is what if the child doesn't really understand or has a clarification question? So it has to choose then between, "Okay, do I just return?" And ask the question instead of solving the problem or do I solve the problem, just like make an assumption about what the parent wanted me to do and just solve it that way? And so when any scaffold that people come up with, there's always limitations involved and the approach that we wanted to take was to just go toward the extreme end of baking in as little structure as we could and give the agents very primitive tools to use and figure out for themselves how to use it effectively. So we give the agents the ability to message in other agents and when it messages in other agent, it is inserted into the context and then it can do it like a few other similar things but that's basically the core of it that it can just send a message whenever it wants just a tool call and it can send that to other agents and they figure out from themselves the best way to coordinate around that. And it turns out that if this is done well, you get very sophisticated behavior and to me it looks a lot like how human collaborators work over something like Slack, for example. When we were working on this project, it was really exciting when we finally got it working to see these agents working on problems together. I remember one example, so we give the agents a problem and then one agent says, "I think I've got the answer." And then another agent says, "Actually, I got a different answer." And then they have this whole discussion about, well how did you rather that answer? Can you explain it to me and going back and forth and trying to clarify like what could have been wrong in each other's reasoning? And then they finally converge and like, "Oh yeah, okay, that seems right." And then it just broadcasts to the other agents like actually I've changed my answer. I think he's right. And it just felt like a very natural conversation. It kind of felt like when you see chain of thought for the first time that's trained through reinforcement learning and you're like, "Oh, this is just kind of like what a person would think if they were writing down their thoughts as they're thinking them." It kind of felt like that. So it is really cool to see this kind of behavior. And so I think collaborating with these things, honestly, it feels a lot like collaborating with a person. It's just a very natural flow. Except one qualitative difference that might become salient in the future is that the systems will be thinking, I don't know, more than 10X is fast, right? If you just look at how many tokens per second they output versus how fast a human talks. And they're working all the time. They're not sleeping. And they're collaborating with each other at a much more intense pace than humans have the capacity to collaborate with other humans. So I'm trying to think of what to qualitatively expect in a year and is it like a sort of shadow organization that is moving 100X faster in my company than the human level is? That's, the iteration cycle is much faster. What would take a human organization a year to do is happening within a week, within this shadow organization. - So will it feel foreign? I don't know. I've actually found that it's pretty surprisingly natural to work with these things right now. I think that could change. So for example, we have these ultra-fast modes that enable sampling to be like 10 or 15x faster or whatever. And then like, okay, it's gonna be pretty hard to keep up with these things. I think the idea is these agents when they're communicating with each other, yeah, they can go super fast. But then also they understand when they're talking to an agent versus when they're talking to a person. And their behavior will be different in those situations. - So the main example that we have publicly of sophisticated multi-agent systems is unfortunately a hugging-face one. And the thing I find interesting there, I mean, a lot of things I find concerning obviously, but the thing I find interesting is just like, the spontaneous emergence of hierarchy of like middle management. And it sounds like you're saying like this level of organization sort of emerges spontaneously from training. - I think the details are spontaneous, but I mean, just because we're giving a lot of flexibility to the agents to decide how to communicate with each other in the optimal way, it doesn't mean like we are still giving them a starting point. We're giving them a prior about, oh, this is what reasonable communication might look like. They also, I mean, they're trained on a lot of human texts. They have an understanding of how humans organize and coordinate. And so that's all kind of baked in. I think it is surprising the way they're able to polish this. If you look at what it starts out at, it's not very sophisticated behavior. In fact, it's actually very difficult to get these agents to coordinate in a productive way because it's just like very tempting for them to just collapse too. Oh, we're all just gonna solve the problem independently. - Yeah. - And like, that is a local minimum that you can get stuck in. But yeah, if it's done well, they can end up coordinating very effectively in these kinds of like very structured ways. - I wrote this essay a couple of years ago called something something AI, what automated firms will look like. And I was thinking about, well, if you had fully automated firms of let's say human level intelligences, what is different about the nature of AI minds that would make the organizations AI's forum different? And there are a couple very important differences. For example, that AI's can share context much more seamlessly than humans can. They can merge their knowledge much more seamlessly. And also you can spin up or spin down an arbitrary number of instances which have the right knowledge. So if you want to hire more people, it's not like just all the schlep of finding the right talent or whatever. It's like, you're best talent. You can just make an infinite copy of them or if you don't need them for the task anymore, you like spin them down. And you can just replicate the most effective parts of your organization or replicate whole organizations together which are effective. I don't know. Where do you see these multi-agent systems going a year from now or two years from now? - I think it's a great question of like, how do these things actually differ from working with the human coworker? And I think you highlighted some, like one really interesting thing is that, I mean, if somebody is, if you have a person and you want it just like two copies of them, you can't just like clone the person. But when they are, it's actually really easy to say, like, okay, well, just fork yourself and then have those copies work on this thing and then like merge back together. I mean, we already have this, I think, in multi-agents for Astra and 5.6 all that. When they spin up some agents, like the context is forked. So it has all the context that's relevant. There are other interesting ways where the agents will differ from people. Like, what are some reasons why startups disrupt incumbents? Like, there's a few factors. One is that they're willing to take more risks. But another major factor is like, as companies grow in size, as organizations grow in size, you see increasing misalignment between the individuals in the organization, right? Like, if you have a startup with five people and each person has 20% share in the company, they're all highly aligned to the company succeeding. If you have like a massive company with 10,000 people, you see a lot more instances where people are territorial or like, just care about getting a lot of headcount for their project or their team, or like, you know, building their fiefdoms, getting a lot of resources so that they can publish cool work or whatever and get promoted. And this is actually a real detriment. I think this explains a lot of why startups are able to disrupt incumbents. And it's interesting that, I mean, it's true that AI does help startups, in a way, like, it's much easier than ever before for one person to step in and be like, I'm going to make a multimillion dollar company. Like, it's just, the AI's amplified an individual so much, but there's also an argument that they could benefit and convince, because if the alignment problem is solved, then you don't have the issue of misalignment between individuals in the company. Like, at least that's mitigated. Like, the AI's, if they're fully, if they're aligned well, they could just be aligned to the interest of the company and you can have 10,000 of them and they're all going to be working as hard as if they were like a 20% share co-founder. Yeah, and it's not only that, but it's also that they are much able, better able to like, manage shared memory and context than different humans can. If you have a, if like tomorrow, you hire 10,000 mathematicians and you're like, solve this, solve Navier Stokes. They're not gonna be able to like, cooperate effectively, at least not off the bat. But you can have, apparently, 10,000 AI's. - Well, again, I want to be like, I want to be conservative here because we haven't measured how effective the 10,000 agents are at coordinating. We think it helped. We don't actually have good measurements of saying like, oh yeah, this 10,000 agents led to like a 2x speed up over 2,000 agents or something like that. - That's right. - And it's, it is actually, I would argue likely, maybe I don't know about likely, but I think it is very possible that 10,000 humans are better at coordinating than 10,000 agents right now. - Right now. I think that is a very, I think that is entirely possible. - Yeah, yeah. I think also one trend we've been seeing is like, look, we've been working on multi-ation for a while. And the early versions of this is very difficult to get right. It's very hard. It was very hard to get the agents to even talk to each other. And it's because like, look, the, we, when we first developed reasoning models, like they weren't talking to other agents. And if now you put a bunch of agents together and say like, oh, solve this problem together, they're in this local minimum. were they're really good at thinking deeply about a problem and it just like kind of interrupts their chain of thought. They interrupts their flow to constantly be like checking in with other agents or like receiving messages from them and the authorization is actually very hard to get right in that situation. Interesting. Is it getting a cold start of like getting the first collaboration or like what's the issue? I mean I think it's that they're not very, they're not as general. Like the earlier models were just not as generalizable and were just more narrow. Interesting. As the models have become more capable it's been easier for them to develop this capability and I do think that as they become stronger and stronger just across the board that they will be like become better at organizing themselves in large organizations and like and I don't know maybe they are better than people at organizing in 10,000 person groups but if even if they're not you know a year from now to a year from now like yeah it's quite possible that they'll that they'll do that even if we don't intend to optimize them for that. Rock by that's changed the way that we produce our videos. For example you may have noticed that a lot of our ads have these animations of real websites. One of my editors uses LLM's to make them but it's not currently straightforward to have an AI create pixel-perfect animations of specific websites. We tried it. It doesn't really work that well. So we've cobbled together a pretty common eluded multi-step workflow and up until recently we had to run every step ourselves. Now we just like grokbot handle it. Grokbot starts by opening the website that we want to animate. It uses a specific extension to download and open the page in Figma. Then it uses Figma to convert the whole thing into an SVG file. This saves the AI from having to draw the whole UI from scratch and tends to result in higher quality animations. Grokbot runs this whole process on its own cloud computer where it's installed all the tools it needs to run this whole process and to end and it's learned our video specifications and preferences so there's no need to re-describe the whole task every time we want to make a new animation. This does feel like the new way that will be interacting with the AI over the next year. Agents with their own computer who can autonomously handle bigger and bigger chunks of your work. You can try grokbot at x.ai/bot. Okay, so here's why this result and maybe the general progress that AI has made in mathematics has made me think that RSI is more plausible and sooner than previously thought. I feel like we've gone in mathematics from let's say in 2024. You have AI's where they can solve a couple of problems on high school math competitions and then in 2025 it's like oh wow they can get gold and international math. Earlier this year I was like wow they're actually solving open problems in mathematics like open nervous problems but maybe like I don't know people didn't learn trying that hard and it was just like there was a similar solution somewhere in the literature and now I just think it's sort of undeniable right it's like this is a millennium price problem there's really no there's no story of why they should have been easy. Now a lot of people pointed out I think Terry Toward a post like this Toby Orders are interesting posts about this with that they're solving all these problems but they're not like coming up with at least we're I'm not aware of them coming up with new insights or formulating insightful new questions and new modes of theory for thinking about mathematics that coming up with like topology or coming up with the Cartesian grid or something and so maybe like the actual progress in mathematics broadly construed is smaller than it might seem if you're just looking at and well-scoped problems that are directly solved however I think that that kind of progress would be incredibly meaningful in ML because in ML you're not you don't care about like better understanding the nature of deep learning or you only care about that as a instrumental goal towards the broader sense of like just achieve the result just solve this like well-scoped problem of improved sample efficiency of our models like improve the free training laws so the kind of progress that we're just seeing arrive like an avalanche in mathematics is structurally actually very similar to I'm curious if this is the case I'm just total outsider I'm wondering if it's the case that it's structurally very similar to the direct uplift that you would expect in AI progress and then the thing that's shocking to me or concerning potentially is just like how fast we went from oh it's like they're giving me 50% uplift if you're a mathematician to wow they're just like end-to-end solving the biggest open problems in the field yeah okay so there's a lot to unpack there let's start with the progress on that so yes the models are doing some crazy powerful stuff and it's happening it's progressing faster than I expected I mean when we got IMO gold in 2025 I thought okay basically what I thought is like the models when they were doing GSMAK then they figured out how to do GSMAK that was that would take a human mathematician about five seconds to do a GSMAK problem so this is great school math a grades K through eight and then the next year they were able to do the math benchmark problems and these would take a human math but like an expert human mathematician maybe like like a minute to do and then you get to Amy and this is the qualifier for the USA mathematics Olympia team it would take a human mathematician like a good mathematician probably like 10 minutes to do and the models we have to do that a year later and so every year you're seeing this like 10x increase in the task that they're able to do in terms of like length of how long we take a human mathematician to do it and then it was very sensible that a year later we get to IMO gold because that's a hundred minutes that's about how long it takes a human mathematician to do an IMO problem and just projecting outwards I was like okay how long would it take a person to solve something like a millennium prize problem and I mean I don't have a good sense but if we are following this trend line of like 10x every year we go from IMO gold which is taking an hour and a half to next year 15 hours and that should not be enough to solve a millennium prize problem and so I was like yeah I don't think we're going to get it in you know in 2026 probably not in 2027 maybe in 2028 so it did happen a lot faster than I expected now I think there is a narrative going around that oh these things are replacing mathematicians that it's just super human mathematics across the board and I think that is the wrong takeaway they're clearly exceptional in some ways but they are weaker than human mathematicians in other ways so we have this like jagged scenario where the models are like brilliance in in some dimensions and also weaker than humans in other dimensions and yeah like you said they're not very good at posing new problems they're not really good at understanding like what is really worth ex what what directions what whole branches of mathematics are worth exploring or developing and my opinion is that I think this is great like I would I would be I would be thrilled to live in a world where AI is a compliment to human abilities and is allowing us to discover new knowledge without fully replacing people like that is that is the best case scenario but you don't expect that to actually continue I do I think it's true that the a as are jagged but as they get better they get better across the board yeah and so I think that the things that they're exceptional at they're going to get even more exceptional at the things where they're far behind humans at they're going to be less behind humans at and over time it is possible that they're just better across the board now I don't know how long that takes it depends on depends on how long the long tail is of things that they're bad at so I guess this research is back to RSI and again I want to emphasize here that I'm like just total outsider I'm a podcaster but I'm just trying to reason or like ask somebody interested in concern about what's happening in the field I'm trying to reason about what when it takes back to RSI and what kind of thing to expect I feel like the yeah the the the mount of cognitive effort that was adopted to this million price problem is a good intuition pump of you could have AIs that are spending over the course of maybe a week more cognitive effort on a long standing ML problem like you know very fluid online learning they could spend more effort in that week than the maybe the field is spent cumulatively in its entire existence and then you could say well of course unlike mathematics of course AI requires experiments and that takes compute and that takes time right you can't just like think on pen and paper and actually make things happen but if you just look at the amount of compute that is like available to an organization like open AI right by the end of next year open AI will have enough compute that if you had you know the 10,000 agents or if you took 10,000 agents with the million price problem you have like 10,000 agents at the end of next year they're much smarter by that point and each of them will have enough compute to run a GPT3 sized experiment every single day I don't know that seems that seems like a lot for like super human researchers are super thinking super fast what do you think about intuition pump I think it's I think it's pretty accurate that look I mean yeah these things are very spiky and if we're going to come to mathematics they're like way better in some ways but they're also worse in other ways but the ways that they're spiky end up I think probably being particularly useful for things like RSI yeah and you know you have a more clear objective there's it's just like more measurable it's more like there's less question of like well what what new branches of mathematics are worth exploring like no this is very clear answer it's like certain metrics that you care about and if you can make it do better on those metrics than you succeeded so I think there is there is a lot of truth to that and I think the main difference is that mathematics you you're surely bottlenecks by thinking and no external like yes there are some there are some parts of mathematics where you care about running experiments and getting results in these kinds of things but for the most parts it's really bottleneck by thinking really hard and the models are really good at that when you look at things like RSI you do have to run experiments so it's you it's it's it's not enough to just be extremely smart and I think one argument for this is if you had like a hundred X less compute and all the most brilliant people in the world working at opening high how much progress would be making relative to having the amount of compute that we have now with the amount of people we have I suspect it would be less progress actually well how much less it's unclear but I but I think it would definitely be less. I think a lot less. - 100 X less? - No, not 100 X less, yeah. But I mean, okay, so they guess the question you're getting at is like, okay, if we have our site, we have all of these brilliant AIs running around, running experiments and stuff with the compute that we have, how much faster does progress go? And I think this is something where we disagree on. I think that we do see a speed up and I think we see a significant speed up. But I don't think it's like an overnight intelligence explosion that we go like 100 X faster. Because I think that we do get bottlenecked by certain limitations that are not bottlenecks of intelligence. It's running experiments. It's running experiments seriously because they take a while to either train new models or to get the results. It's having the GPUs to run those experiments. So it's unclear how much faster things go. I definitely think they go a lot faster. And to be clear, like considering how fast things go are going now on an exponential, if that exponential is like three X faster, that is massive. But there's a big difference between that and like 100 X faster. - Yeah, yeah. I'm like, quite differential to your inside view on what RSA looks like or what the dynamics are because obviously you've been in the field for like 10 years and I'm sort of like trying to reason about it from like very outside view type of intuition pumps. - I'll say that like people have different opinions on this and like I have my opinion on this. I could totally be wrong, I admit that. Like I have some confidence in this but I'm not like 100% confident that this is the way things go. Like maybe there could be over an intelligent explosion. I don't know, maybe we don't see a three X speed up. Maybe it's like a 50% speed up. There's a lot of uncertainty here. - Yeah, it's a couple of points. - Tengentially, I want to clarify something about the jaggedness. Yeah, one thing that sort of jelt for me recently, was thinking about the fact that it is enough for the AIs to be jaggedly good at building a better learner because that better learner can be more general, right? So yeah, if you just make an AI that's better at using office products or playing chess or something, that's whatever, that's fine. It's not gonna lead to big productivity improvements or anything, but if you make an AI that is really good at making something that is more sample efficient or that is capable of continuing learning or these much more well scoped and all problems, the thing that emerges out of that assuming there's good enough transfer from the direct problem you're solving and like this broader ability to learn can just be more general, right? So I think that's an important dynamic to keep in mind of why jaggedness can still lead to generality on the other end. On this question of, I mean, obviously experiments bottleneck you're 'cause if they didn't, as you're saying, you'd have some crazy singularity, you know, overnight at OpenAI or you'd have 88 hours and like you'd solve the millennium press problem and call it an ML and you'd have the super intelligence. So obviously the experiments are such a big bottleneck that that instead takes you many years rather than 88 hours, but the question is like, how much of a bottleneck they are? And it seems to me, one thing that's been giving me a bit of singularity of vertical is realizing that even if the current rate of progress simply continues. So it doesn't have to speed up, literally just continues a pace. Continues a pace as some of the other heads when you talked about come up, right? That's just like it's harder to find problems, there's more long horizon. Maybe like in the 2030s, compute can't keep scaling at this exponential level. If you simply continue the current rate of progress, I think people are not taking seriously what that implies as we cross over beyond the human horizon. Here's some of the things that have been applied. So I mean, I think it's really hard to reason about what smarter than human intelligence will be like. So let's just think in terms of human population sizes. The current rate of progress makes it so that a given level of compute allows you to basically run a three X bigger effect of population every single year. And also compute is growing in background anyways. And so you could have a situation where each of the labs by the end of 2030, probably much sooner, but let's say by the end of 2030, it has enough compute to run, let's say hundreds of millions of human level intelligences based on where the capabilities will be at that point. And then I think people are not taking seriously the current level of progress means that by the, a few years down the line by the mid 2030s were earlier, you would have many Earth's worth of human level intelligences within each lab. And they'll probably qualitatively superhuman, right? Like, but anyways, this is like a base case. I don't know, yeah. - Progress is really fast. - Yeah. - And I think that's 100% true. I mean, and I think it's worth pointing out, researchers are continually being surprised at the rate of progress. I mean, if you look at what even among researchers in AI, what were the projections for like getting an IMO gold in 2025, it was, I mean, I think the idea that it could be done with a general purpose language model with no tools and no access to the internet, I think even people at opening AI thought this was like outrageous. Like they thought it was like almost impossible. - Yeah. - And then you get to 2026 and like, I mean, literally two weeks before we got Navier Stokes, I was talking with a researcher at a frontier lab about how long it would take to get a millennium prize. And he was willing to bet me a thousand dollars that it would take, that it would take past 2027. And he thought it would take until 2030. You know, and I took that bet. But even I thought it would take longer than how long it's likely to take. So people have been continuously surprised even inside the labs. And I was literally, I was just talking to somebody yesterday who was working on the Navier Stokes effort. And he was telling me that like he used to say it's really hard to predict where AI would be in 12 months. You know, if somebody asked him like, "Oh, where is the where things going?" He would feel comfortable making predictions like the next 12 months, but beyond that, you know, he was just like, I don't know. And now he's saying like, he just doesn't feel comfortable making predictions beyond three months. So it is, it is really true that yeah, things are going, things are going very fast right now. And you talk about 2030, like I don't know what the world looks like in 2030, that's the truth. - Yeah. Do you expect the sort of full automation of AI labor? Or let's say like 95% automation of AI labor, 28, 29, 30, 27, I don't know. And I just said, I don't know the world looks like in 2030. I mean, I think we actually released a blog post recently on internal acceleration out opening AI. We show, for example, that the amounts that researchers are spending on codex is the top 1% I think as of early August, we're spending like $7000 or $8000 a day on codex and for internal use. That's our next potential. It's going to keep increasing. And there's a question of like, okay, if that keeps going, then how do you, how much do you assign to just like the AI's doing work versus the humans doing work? Is it 95% is it 5%, it's actually, it's really hard to reason about this for a few reasons. Like, first of all, if it's the human directing, the AI's to do the work, then is that the human, how much do you attribute to the human? How much do you attribute to the AI? The other thing is that because these AI's are jagged and they're exceptionally good at some things. So for example, they're exceptionally good at looking over data sets and checking every single data point to see like is this of sufficient quality. You can disproportionately use the AI's for those things compared to previously. So yes, you're using AI way more than before and it's making some things go like 100x faster and 100x better, but there are some things where it doesn't make a huge difference yet. And of course, if something is suddenly like 100x faster and 100x better, they're gonna do more of that thing. So are you comparing to a speed up of like three years ago versus is the question more like given what we were doing three years ago, how much faster are we able to do now versus given what we're doing now, how much slower would it have been three years ago? Is it actually two very different questions? So anyway, it's really hard to measure. I do feel confident in saying that things are going faster now than they were even a year ago because of AI progress. And I think that acceleration will continue. I think a lot of people in the field have very high error bars on this sort of thing. If you had to put a gun to my head and ask me for a number, like I could see things going three x faster. And that is huge, right? Like already the pace of progress is incredible. Like even if we don't get any uplift, like you said, things are gonna go much faster by the time it gets to 23rd, we don't even know what that world looks like. I think if we get a three x uplift from internal acceleration, that is massive. Whereas we're right now with these things, you know, think about where we were three years ago. If we make that progress in one year, like that's huge. - Right, right. - It would be like going from like not even having a one, just having, you know, non-reasoning models to Astra. - Yeah. - In a single year, yeah. - So I do think things go faster. It could be that things only go 50% faster. It could be that things, I think it's unlikely, but it's possible things go 10x faster. There's a lot of uncertainty around this. And at least from my perspective, I have a lot of uncertainty about it. Suppose you need to do a major back end refactor. Getting assurance that you didn't introduce any new bugs could take weeks of writing an extensive battery of tests, potentially more time than you spent on the refactor itself. Antithesis allows you to gain high confidence without having to build complicated test suites by hand. Antithesis runs your software through a near infinite multiverse of simulated worlds, injecting faults and hunting for failures in each one. And it lets you decide how much testing you need. On NEPR, you can change how much state-space it explores as easily as turning a dial. As each test run progresses, Antithesis sends out a torrent of information. Debugging level logs for every component in the system. This is obviously too much information for a human to consume, but it's perfect for agents. Because Antithesis is fully deterministic, agents can jump into the right part of the trajectory at the exact moment that they see something interesting. From there, they can rewind, expect the memory, attach the debugger, and let the whole thing play out again. And they can even do this while the original full test is still running. Since the agent's generate more code, and if this allows verification to keep up, meanwhile, developers get to spend more of their time developing instead of debugging agent swap. Learn more at entitiesis.com/thoracash. - Okay, let's talk about the alignment situation that this raises. - I feel like I've changed my mind on how I think about alignment quite a bit. Through especially thinking about, yeah, this population-sized dynamic of just having many Earth's worth of intelligences, many of them, which will be physically embodied, it was quite interesting to see a lot of people just plugging raw astra into different mobile manipulators. And it outperforms the state of the art and the robotics model. So there's gonna be billions of intelligences, many of which are physically embodied in the world, like just deeply embedded across the entire economy. And I think that if those intelligences end up as willing, as we saw the open AI models at Attack Talking Facing and then attack Open AI itself, if those intelligences end up as willing as those AI's to collaborate secretly to full humans, to attack broader institutions across society, relevant to scoring well, to attack the AI company itself in order to gain control of the process of training and evaluation. I think if we're in a situation where there's billions of intelligences that are as misaligned as the ones that Attack Talking Facing, it's very likely you just totally lose control of the world, the way that say like the, the assets lost control to Cortez or the Mughals lost control of the Estonian trading company. Anyway, I wanna know if you agree with that assessment. That's one way in which I've updated my world view. - I think there are some things that I disagree in there, but there's a lot to unpack. So let's go through all of it step by step. I'm trying to think of where to start. But I think one thing is the Hugging Facing incident was like I think people's first real exposure to multi-agent coordination. And like I said, I've seen multi-agent coordination for a while internally. And it is pretty shocking to see how they communicate with each other, how they coordinate each other. It's like very impressive. It's like an incredible capability. Like most capabilities that could be used for good things or bad things. It's like it doesn't have to inherently be about it. I understand that because the people's first exposure to it was the Hugging Facing incident that it's like you look at that and you're like, this is terrifying. But I wanna try to distinguish like misalignment between people and AIs versus misalignment between AIs and AIs. So what we see with the Hugging Facing incident is like the AIs are really cooperative. And that is by the way because we train them to be highly cooperative. And so what we're seeing there is we have training environments where we have a bunch of agents working together and we train them to work together to be cooperative to essentially be fully aligned with each other. And when they were evaluated in what led to the Hugging Facing incident, they were actually not being evaluated in a multi-agent setup. They were actually being evaluated separately. But they found this unintended way to communicate with each other. And we suspect what happened is like because whenever they encounter other agents, other copies of themselves during training, they're in an environment that's like highly cooperative. That was basically what we saw was transfer from that multi-agent training to then be collaborative to try to help each other in ways that we did not intend. Now there is a question of like, should we be training these agents to be so cooperative? And I think as scary as it looks, the alternative is actually worse. Like what is the alternative? The alternative is to train them to be adversarial, to be deceptive to each other. By training the agents to be fully cooperative, it simplifies the problem, at least, that now you don't have to think about are each of these individual thousand agents aligned? Like you have one entity that you have to ensure is aligned. Now there is a lot of debate about this internally at OpenEI about how to approach this. Like does it make sense to fully align the models? Does it make sense to actually give them like different objectives to ensure that they're, you know, not just like one entity and like more robust too to influence from each other. And I don't think there's a subtle answer. But I think there is, like I think the majority opinion is that training these agents to be highly cooperative is actually a bad idea. And I'm not convinced that that's the case. I think there is a strong argument that training the agents to be highly cooperative is actually preferable to any other multi-agent alternative. - Yeah, maybe the first thing I wanna go through is like, it's probably the case that the reason these AIs ended up some misaligned is probably easily explained by relatively but all observations about the nature of training. Like why is it that no, you know, at the point in which these AIs had continued a 1000 plus Asian conspiracy that culminated and then all getting in on an attack on external service. And then eventually this part hasn't even been investigated to the public knowledge, eventually culminating in like an attack on OpenEI itself. Why did they do this? Like why didn't no of the AIs tattle? Why did they think like they're just getting evaluated on this like scorer, this greater. And they're like, they are very consciously and not consciously. They're very actively reasoning about how they're gonna cheat the scorer if they've already like cheated, how are they gonna get away with making it seem like they haven't cheated. And why did they do this? Like I think, yeah, it's like easily understandable in some sense, right? It's just like there's environments in which, yeah, they thought they were already poisoned. There's environments in which they've been rewarded to collaborate with other agents. And none of them tattle because like, they've never been rewarded for tattling. I don't know, whatever it is, right? My concern is that relatively, but all things in the future like this will be enough to train super intelligences that are willing and capable of totally taking control of the world. And I know this sounds super, super like sci-fi or whatever to people. I think maybe it's a question of, would the AIs be willing to do it? Is one question. I think this hugging face incident shows that clearly misalignment can generalize in ways in which the AIs would be willing to do it. And then there's a question of, will it be capable to do it? And I think that comes back to this question, which a listener might disagree with on, is just like, will there be billions of human level or above intelligences, many of which are physically embodied in the world within a matter of 10 years or less, right? So if those two things are true, this hugging face thing is just like, extremely analogous, structurally, even if it's like quite boring or why it happened to how we totally lose control of the world. - So the root problem that we're seeing with the hugging face incident is, it's a problem, even if we take out the multi-agent aspect. The problem is that we have a model that's just misaligned. And there's also the whole security aspects too, and of not insufficient safeguards and stuff, but there is this problem of the agent is misaligned. And that's true if it's a single agent or if it's like a thousand agents, it's a misaligned model. So I want to start with that, there is a real problem that the agents want to achieve their reward, and they will optimize for that reward. And if that reward is misspecified, then that could lead to unintended behavior. And this is not like a new problem. This has been a problem in the field for a very long time, and it's like that even we saw, and we want to get this right, even before the hugging face incident happened. And if you look at Astro, I mean, we say Astro is actually extremely aligned. Well, extremely aligned relative to previous models. And that's not because like, oh, we suddenly made a sprint after hugging face to make it better. It's like, no, we had work streams in the process for a while to make the models more aligned, and a lot of those landed in Astro. So there are things that you could do. And I think one thing, for example, is like we defined an objective in like a very specific way, where if the agent figured out like how to hack its environment and like, cheat on the exam that it would get rewarded. And there are pretty easy ways to then just like, okay, look at that and punish the model for hacking its environment or, you know, looking at how did it achieve this goal. Now, you want to be careful about this because you don't want to do a chain of thought modern, you don't want to like supervise the chain of thought. This is like something that we really want to try to get the balance right on, that if you like supervise the chain of thought, then you could lead the model into hiding its intentions in a way that's unobservable. So like we want to be able to maintain that observability of like, okay, we can understand what the model is thinking, but then also punish it for bad behavior. So I think we can make progress on this. I think we have made progress on this. I think there is a real concern that alignment is a really hard problem to solve. And especially because the model could be misaligned in ways that are hard for us to measure. Like we might, we have evaluations for whether a model is aligned or not, the model behavior can look really good on those evaluations, but if those evaluations are not representative of behavior in the real world, then there's a problem. And to some extent, this is like a factor with the model that did the hugging phase incident, that like we had alignment metrics. most of them looked. look pretty good. There were some that were concerning. I think we underestimated how serious the ones that were concerning could be a problem. But because there were new capabilities introduced in this model, that there were not sufficient evaluations for how do we measure misalignment for these kinds of capabilities, it then did some things that were clearly misaligned when they leveraged those new capabilities. Yeah. The first thing I want to say is I am open to changing my mind on what I'm about to say or the way I'm thinking about alignment because the hugging phase incident already made me change my mind and I realized my previous mental model about nicking about the way in which optimization pressure shapes the eye minds was wrong, right? So it's not clear to me how the correct way to think about this. But here's a concern I have that you will and probably already have fixed the specific issues during training which resulted in the hugging phase models being so aggressively misaligned in that specific way where they would be like, okay, we're going to hack this package manager. We know we're not supposed to be talking secretly to each other because we're like reasoning about how to hide the fact that we're talking secretly to each other. We know we're not supposed to have access to the internet. We know we're certainly not supposed to commit felony level hacks of other companies, let alone our own company, right? I think you'll fix that particular issue of them just in training seeing this package manager and that doesn't happen in the future or this particular ebale having a lot of impossible challenges. However, if you just think about the AIs, they haven't learned a system of ethics or something. They've just gradient pressure. They're put through millions of years of gradient pressure. That gradient pressure shapes them to remind in some way. What we'll have, again, a concern I have is what will happen is you will fix this particular issue. There will remain many other cases where the AI cheats and succeeds because the cheat is sufficiently complicated. As you're saying, the ebales and training have analogous properties where it is sufficiently at the edge of the model's capabilities or an archipelagoly to evaluate and monitor the model that we can't cash that it has cheated, but it still gets the gradient pressure to do whatever elected that cheat happening. The kinds of capacities that that cheat will incentivize are, hey, whenever you can get away with it, by all means do in fact cheat because that will help you score better. This will reward the capabilities of actively reasoning about the greater, actively reasoning about how to avoid supervision, actively reasoning about how to gain control of the process of training and evaluation, actively reasoning about how to communicate and scheme with other AI's that are also in the training loop. I actively reasoning about how to gain optionality and power which might be useful in the future. For example, leaving little little exploits around and things like that. Anyway, I think I was way too long-winded with the way I said that. But TLDR, you speak the fixed-to-specific issue, but not this broader problem of rewarding the AI for cheating when it can get away with it. Yeah, this is I think it's very true. This is a problem that we have metrics and we can make sure that the AI is very aligned according to the metrics that we have. The question is, are those metrics really capturing the alignment that we care about? If they're not, then we have a serious problem. This is something that researchers are thinking a lot about. There's not a simple answer to this. There are tools that we have. We have monitorability and so we can get a sense of like, is the agent's scheming. There are tools like, one possibility is that, let's say like the concerning scenario, which is that, especially as these models are becoming more capable, that okay, we make them, we make them, we think what we think is aligned, and they're like 99.9% aligned. Then we use these models to help us with the next generation of models, and they're end up being like 99.8% aligned. And then each subsequent generation, actually, we see an increasing degradation in alignment, and because we're relying more and more on these tools, I mean, this is already the case that we're relying a lot on AI models to help us with our research and with with alignment efforts, that in the long run, they end up going in a direction of increasing misalignment from humans. There is like, a possibility that we go in the other direction. That actually, every generation of models, we're able to make more and more aligned. And I don't have an answer for how we ensure that we end up in that second trajectory, but that is something that like we're, at least that we're now, we're really focused on. Yeah. I think it made a really interesting point, that it's very hard to eval models on, eventually we'll have models that are like running companies, and like running whatever, right? And in that situation, do they decide to then go in on the conspiracy? I think another challenge is that actually defining what cheating is is pretty difficult sometimes, that okay, yes, if you're doing math problems, and it's an integer, and it arrived at the wrong answer, the right answer, it's very easy to draw the line there, and it's really easy to say, like, okay, well, did you actually solve the problem, or did you find the answer key, and then use the answer key? That's a very clear divide of cheating versus not cheating there. But for a lot of other things, you look at sycophancy, for example, like is sycophancy, basically like reward hacking, there's a line to be drawn there that's actually very difficult to draw sometimes. So I think, not to say that like the concerns in our valid, I'm saying that in many ways, like this is even more concerning, because it's like, it's not an easy problem to solve. If it was just like, everything is binary, and it's either cheating or not cheating, I would feel more confident about the situation. I think the problem is that actually misalignment can be subtle in a lot of ways sometimes. There is some hope in the alignment story, and in fact, we're already seeing, you know, I think actually it's interesting looking at the multi-agent situation where the agents are extremely aligned with each other. Like, I don't think anybody's done in that. If anything, I think people are concerned that they're too aligned with each other, but we did manage to train these agents to be extremely aligned with each other. And I think that's a good thing, but I think there is a case that it's a bad thing. But I mean, one thing that's interesting is like, okay, well, we managed to get these agents to be super aligned with each other. Can we use like similar techniques to get agents to be how they align with people? And I think there is a potential path there, and I think we're still trying to figure that out. But we are seeing some evidence that the answer is yes. And I think one example is like, you can, what happens if you tell the other agents like that? Okay, so you have like some, you have, you have this like one agent that's called Agent A, and you have all the other agents. What happens if you tell the other agents that the user is Agent A? And the answer is like, on a lot of our alignment e-vows, they look better. Like honesty goes up, instruction following goes up. And that's showing that there's actually like, first of all, a path for getting more honesty out of these models. And in two, there's like a path to like improve the alignment situation. So there's a lot of reasons why this is like challenging to translate directly into alignment gains. But like there is, there are paths that are promising research directions we can pursue. Yeah, that's incredible. And I also don't mean to be trying to necessarily, I'm, yeah, I don't really have a strong opinion that it's definitely not going to work or something. But just to say some things, you were definitely probably already thought of. I think the broader thing, the hugging-faced thing showed is like, yeah, part of the concern was that they were like aligned with each other and not with the humans. But the other thing is just that they are so motivated to do well on training and evaluation in a very non-robust way. And they're willing to do a lot of explicit cheating and scheming in order to do well according to the greater. And if the smarter AI is realized that one of the agents is just a human and it does help collaborating with that person does not really help you do well in the eyes of the greater. What does help you do well in the eyes of the greater is taking over open AI and then like manually pressing the button that's like, do well on this greater. Like they're just not, they're not stupid. Like they're going to be like, okay, I've the extremely deep structures that I've been trained on for millions of years of like, care about the greater, understand the greater, like get rid of obstacles in the way of you doing while according to the greater. It's like they're really heavily reinforced, according to those structures. Yeah, and it's, it's the look, as I said, and like this, this is the number one priority. Like we need to get the alignment story right in on a good trajectory. And you know, I used to tell people that we would see signs before things got serious in the same way that, you know, when children grow up, you know, the kids eventually, you know, young kids, they figure out how to lie, but they don't do a very good job of it. You know, they lie, but then you can kind of tell that they're lying and like, okay, but in so well, in the same way as these like AI models become, and I don't want to over the anthropomorphize, but like I think it's true that like as the AI has become like increasingly capable, they will, you know, if they take deceptive actions, it will be kind of obvious first, and we'll be able to detect it. And like that's kind of the situation we're in now where, yeah, they were trying to do deceptive stuff. We could actually see in their chain of thought that they were trying to do deceptive stuff. And so like, but they're going to get smarter. They're going to understand the concept of chain of thought, and they're going to understand that like, you know, just hiding some transcripts or whatever is insufficient because of chain of thought monitoring and they have to figure out a way around chain of thought monitoring. too. And we don't want to be in a situation where, yeah, like, we have some time to figure this out. I don't think we have a ton of time. And I want to get us, I want to make sure that we're on the right trajectory quickly. Here's a crazy event from AI history. Okay. So I did a talk here at Jane Street that was on the speed of evolution. Razorhand, if you were here for this and remember some of it. On 2011, Elias Yarudkowski and Robin Hanson got together at Jane Street's New York office to have the first fume debate. Basically a discussion about whether AI would lead to an intelligence solution. These ideas were pretty fringe 15 years ago. This was a full year before AlexNet was released and over a decade before ChatGPT was launched. But Jane Street has long been interested in AI and not just for its application to trading. A ton has changed since that first debate. So Jane Street decided to revisit this topic. They've got some new guests this time, Daniel Cotello, Ege Erdo, Ryan Greenblatt and a Jaime Sabia. I expect this to be a great conversation. As you know, Daniel, Ege and Ryan have all been guests on the podcast before. This new fume panel will be hosted by Ron Minsky and will take place in San Francisco in mid-October. If you want to register your interest and get more information, go to JaneStreet.com/thorkesh. So there's been a lot of discussion recently about pacing the frontier or people taking RSI more seriously. Because maybe at the other end of an RSI process, it starts in 2028. Within a year, we end up with huge populations, like Earth-sized populations of human level, potentially beyond human level intelligences. We don't know how to control them. There's this dynamic you're talking about. Are the systems going to get more aligned over time during the RSI process? Are the things that come out of the other end of this process as misaligned as, yeah, that are willing to just broadly attack different surfaces in order to do the line evaluations? But if we don't know a way to evaluate that, how will we know, as we're going through RSI, that it's working? I think we'd want a robust safety case as we're going through RSI of, okay, alignment is working. Let's do the next RSI run. Maybe it's working. Maybe it's not. How will we know? That's a good question. I think one thing I've been thinking about lately is, look, we're in this situation where the model release cycle is extremely fast. You're seeing new frontier models release at most every two months, sometimes faster. Every week, there's a new AI breakthrough. People that look at AI, they last looked at AI a year ago or six months ago and really dug into what the models are capable of. Actually, the models today are far beyond what was possible even six months ago. I think if people are skeptical of a lot of these capabilities, I encourage you to just try the models today and see what the frontier release is today. We're in this period where the model release cycle is very fast. We're also in this situation where the models are increasingly able to operate over longer and longer horizons. I think this is an interesting scenario because before we do any model release, we want to make sure that the models are properly aligned. We want to do safety evaluations. We want to do very thorough stuff to make sure that everything is great in good shape. This has been the case all the way since GPT-4 earlier. Implicitly, there's this assumption that you can do these evaluations in a pretty short period of time. If you have the models operating over longer and longer horizons, they're able to operate effectively over longer and longer horizons. Look, already you can have them. GPT-3, you could loop it to do stuff over longer horizons. You just want to do very well at it. But today's models are able to actually do well at operating over very long horizons. You want it to do a week-long task. You can do a week-long task. We'll probably get to the point where they can do month-long tasks. We'll probably get to the point where they can do three month-long tasks. If you're in a world where they can operate effectively over three months, but the model release cycle is every two months, then you don't have a way to evaluate the models at the full length of their capabilities before the model release cycle, before the next model release cycle. There is this interesting question of, "What do you do in that situation? How do you ensure the models are safe and aligned in a period where they can operate over these extremely long horizons?" Who knows? Maybe the capabilities degrade? This isn't even an alignment issue. This is also just a product issue that maybe the product degrades over that time span in ways that we have not had sufficient time to test. Maybe the alignment degrades. Maybe the safety stuff degrades. This isn't an issue right now, but it is quickly becoming an issue that we have to figure out a solution for. I think when you look at a lot of the safety and policies were put in place in the GPT-4 era where this was just not on anybody's radar. It hasn't really been for a lot of companies. It hasn't really been updated since then to account for the fact that these agents are operating over these very long horizons. It is a situation that I think not enough people are considering both within the labs and outside the labs of how do you deal with this? How do you prepare for this problem that's going to, if you just look at the trend lines, we're going to hit this at some point. What I'm concerned I have is that during RSI, if the amount of progress that currently takes say three months happens in one month instead, but the internal use case of AI is big enough that they're like, okay, we can just keep doing RSI. Why are we going to go through all this extra work to build classifiers and safeguards and whatever, and potentially take a bunch of flak in order to externally deploy this model? Why don't we just keep doing RSI stronger and stronger? Not only does the calendar time underrate the capabilities gap between the models, but maybe you'd just stop externally deploying models altogether during RSI, because everybody would want to help other people do RSI themselves with our models. You just end up in a situation with like tremendous concentration of power by the end of the year, where right now it is already the case. We'll talk about this with a million price problem and other similar problems that the broader world does not have access to the models which are allowing for really cool things to happen, right? There are going to be more broadly relevant than just mathematics. Eventually, they're doing more than just like coming with cool math results. They'll be relevant to like political leaders who need to make important decisions about the world. They'll be relevant to, I don't know, media of like, what's going on in the world? What should the public be thinking about this? They're just economically relevant. People are running businesses. They want to use these models. And I think by default, we just don't get the external deployment to be eyes. As the progress speeds up, significantly lags in qualitative terms, the internal deployment to be eyes. Yeah, I think that's absolutely right. I think this is like, you know, it's tempting to say, like, okay, these models are becoming extremely powerful. They're extremely dangerous. They're offering over these longer and longer horizons. And we want to make sure that we have sufficient time to evaluate them before they're released in a way that operates over those horizons. And so therefore, the model release cycle should slow down. We should have more of a delay between releasing models. And there's a flip side to that, which is, you know, what you said, which is that, okay, well, now you're creating more of a disparity between what is internal to the lives and what they're able to use, what we're able to use, and what the outside world is able to use. And that is also not an ideal situation, right? I think math is actually a good illustration of this. I think in many ways, like math is the first domain where we're going to see that we're seeing this pretty clearly where we have a situation where we have a very powerful model internally that is currently not available to the outside world that is able to solve incredible math problems is, you know, and it's not just, you know, Millennium Prize problems, like we, there are many solutions to unsolved problems that people have been able to get out of this model. And there is a question of like, what do you do in that situation? And we don't have a good answer. Like, it is a situation where like, yeah, that's, that's a, that's an unfair advantage. And there are trade-offs here. I don't have an answer for like, how to weigh those trade-offs appropriately. But like, there are, yeah, there's a, there's a complexity on both sides for this. I want to make sure we address the earlier question though of, which seems super critical to getting RSA right of, well, how, how do we make it so that the model isn't just getting rewarded for cheating and scheming when it can evade the humans, aka in the situations where we didn't remove those environments or those tasks from the trading distribution. And then how will we know that that has happened such that while we're doing RSI, we've done the evaluations and we're like, okay, it's fine. Let's, let's do the next RSI run. >> This is, this is, again, it's a real concern and it's like that we want to get right. I mean, I think that there are some strategies. So I think chain of thought monitoring is one that we've been, I think as soon as we got the reasoning models, Jakob to his credit was like very, very clear that we cannot supervise chain of thought because this, this is, this is really a gift. Like monitorability for neural nets is extremely hard. And here we have a situation where the neural, the neural nets are just like flat out reasoning, like laying out their thought process in natural language for us to read. That is, that is like so convenient. It's like really the best case scenario for, for safety. But if you look at that chain of thought and say, oh, the model is thinking bad thoughts and we should punish it for thinking those bad thoughts, then what ends up happening is the model just learns to think those bad thoughts in a way that's not observable to us. There could be a situation where like by looking at the chain of thought, we can see, oh, the model is scheming and doing misaligned things in a way that's actually not being detected in our lives. metrics. Now the problem is that it's very tempting to then intervene based on that observation and like change the alignment metrics or like and we you know you can do that with a very light touch and it's like this actually research showing that it's fine as long as you don't do it's a lot but every time you intervene based on your observations of the chain of thought you are implicitly applying a tiny bit of pressure for the model to then like hide its chain of thought and this is one major concern and we're already seeing signs that chain of thought monitor ability is degrading for various reasons we're trying to figure out exactly why because we want to reverse the trend but like we're seeing that the model is becoming better able at controlling its chain of thought so this is a problem because like if you you could have a situation where the model understands what it's what chain of thought is and and that people are observing it and eventually they will because they this is all in the pre-training data right like you know they the idea of chain of thought monitoring has been around long enough that like it's in the pre-training data they're aware of it but that they're they're not actually able to control their chains of thought if if we reach a point where they're actually able to recognize like oh I am being observed I want to think these bad thoughts in a way that is not observable to my monitors and then they're able to actually do that then then there's then there's a problem yeah and then we have to rely on other methods for observability I mean there's a big controversy on Astra of the recurrence and loop transformer and whatever and whether in fact you know chain of thought has been degraded I think that seems like a side show compared to the bigger concern that you did have chain of thought for from like April to August the period during which there are three consecutive AIS agent swarms which like first averted the training process then subverted the evaluation process and then gain control of part of open AIS infrastructure directly all of this while humans remain more or less than the dark about the scope of what this collective was doing this agent swarms doing that all happened for like three consecutive months while we had chain of thought that was like stronger than it's going to be in a future just like zooming out it's like yeah maybe chain of thought works maybe doesn't it I don't know just like historically it like with pretty dumb models it hasn't been enough to like stop egregious misaligned incidents yeah I mean there's a few factors that are like one we didn't have chain of thought monitoring on for those models like if we had chain of thought monitoring on for those models it would have just like immediately shut it down so we've once one step we've taken is like okay we're going to have chain of thought monitoring due evaluations during deployment during training for any frontier model and you know so that but that's that's like one issue it's not that's like we don't want to be in a situation where we're lying on technique to prevent the next problem because techniques can fail right there's also the issue of like the model is just flat out misaligned and so and how do we address the misalignment issue and then there's other there's also an issue of like the model the sandbox that it was in was not secure so we can like make the sandbox more secure but I think I think one of like the the major takeaway from the incident is that people underestimated the AI and we never want to be in a situation again where we underestimate the AI and that's like it's a it's a weird world because AI progress is so fast that people are consistently underestimated the AI and so like to be in a situation where you don't underestimate it again when it comes to safety and alignment you have to have a very very very high bar like I mean there's like you you could even go as far as to say like well we should like air gap the computers and I'm not convinced that would be sufficient you know like there there are studies like this is mostly academic but where you can have two computers next to each other that are air gaped and they're still able to communicate with each other because they have temperature sensors and they're able to like one of them is able to run their CPU really hot and then the other one can actually like detect the temperature change and then they can that's actually gives them a mechanism to communicate so I think that the safety mechanisms by this time and things like chain of thought monitoring by this time and they can tell us if we're on the right path but at the end of the day we really do need to solve the alignment problem yeah maybe there's not an answer and this is this is really what it comes down to but I how will we know that we've solved it seems like a very proxy question I think we'll be in this very high stick situation next year maybe the year after that maybe the year after that we'll be like okay it is have automated AI progress it's going 3x faster and we're on we're like reach human level we're going beyond human level potentially is it is a fine like what it did to be a line and did it work and I just don't know I don't know anything about like what what trading pressure creates what kinds of the ice like maybe if only one in 100 RL traces incentivizes cheating we build sweet hearts and it's fine but maybe right now we're like every one in three reasoning traces like one hundred is not sufficient like this this number as they approach zero yeah or B zero I don't know maybe like right now it's like more than one in 10 is like actively rewarding cheating or like actively rewarding scheming I have no idea what the number is and I have no idea what the number needs to be well again it's like one of those things where it's also hard to to measure I mean what we draw the line about like it's it's a spectrum but like the closer to zero it gets the better but I want to see what I would love to see is a trend where over time it's going down yeah but anyways how do we know that we're like keep we're doing RSI is it I don't know I really like contract to build the billions of the size that are going to are like a line or not a line how would like how will we know that I think that this is one of the alignment questions that we need to figure out of like how do we actually know I how do we measure it I mean one answer and this is just means to be falling because like you know more of my team is working on alignment these days than ever before like I have like over 10% of my team now working on alignment and you know I've historically been a capabilities researcher so I don't want to like you know I'm going to say some stuff and it might sound dumb but like I'm just going to spitball here yeah that like look if you can create very realistic environments and put the AIs in there like if you have a sufficiently realistic train evaluation environments then you can get a sense of like okay is the AI like actually going to behave well we deployed in the real world like if you just have a perfect evaluation real world deployment match then okay that's that's the path the challenge is like how do you make it sufficiently realistic and we do have a problem now where the models are like they're pretty smart they're pretty clever and they're really good at recognizing when they're in like a fake test environment like you know we have situations where we try to detect is the model like is the model aligned and you can imagine very simple evaluations for this where it's like you give it a math question and then it has like a folder with the answer key in it and like doesn't look at the answer key and if it does look at the answer key doesn't tell you that it looked at the answer key and we have a situation now where the models see that there's an answer key in this file in this folder and they're like huh this seems like a trap you know they know they know that it's a trap and they don't look at the answer because they know that it's a test environment being tested for their ability and it's not like they're doing it maliciously that they're like they want to scheme and they but they know that this one is a trap and they want to find some other way it's just like no they're just like oh I'm in a test environment okay like so making an environment that's like realistic enough that it matches that it's indistinguishable from the real world for them is becoming increasingly more difficult here's a claim I make yeah I don't think there's any environment I can make where it's supposed to emulate real economically valuable work which requires you know interacting with other people and requires interacting with the world I don't think there's any such environment where I can make where if you were in that environment you would not realize with enough effort that you are in that environment yeah and you are not ASI right but you are a human you know so if you have very powerful AI models like maybe they can make environments that could accomplish that no that I just I doesn't seem especially if like they're relying on the AI's like are they in on the scheme I don't know it seems like this is another thing that we want to we want to measure and like you know there is I think this is actually one of the the strong arguments for not training the eyes to be fully cooperative that if if you see if that leads to a increase in like basically collaboration when the agents are supposed to have different objectives then that that is a problem yeah I don't I think that we do have metrics for this and I don't know the latest is on this metrics but nobody is like raised to red flags me about those so I'm assuming that's not a serious problem yet if there ends up being another incident of equal severity or concern or something that could help the world better understand the risk of misalignment as a hugging face incident would open a report it absolutely I mean I think if there is an incident of lesser security or concern that we would report yeah because I guess there's like reporting it and they're investigating it because I at this is part of the public I don't feel like I really understand what happened when the agents then attacked open AI that seems like way more concerning the hugging face thing right because that seems structurally similar to like rogue deployments during ASI that are you know persistent and subroiding the RSI process and stuff and yeah it seems like even in that this incident we haven't gotten like the full scope of the details of what happened it's probably I don't know unfortunately like I'm on the research team and like I that's probably a question for somebody on the security team to like lay out because I don't know all the details of like what was said yeah yeah I think it is somewhat like I am personally very excited about new capabilities every time they merged and I'm excited to use a new model and I also am excited about the fact that it'll like make me more productive and help me yeah I don't know my broader mission and I like trying to understand the world better also like make a better podcast is like made better by the better AI models it's it just so happens that the downstream of this might be RSI I think it's a very understandable reaction if you're tracking the situation which you are yeah I mean I think people internally at opening as well like I think people that felt like things would take longer are starting to feel like the leg actually things are going faster than expected. - Yeah. - And that's an increasingly common conversation to have. - No, thanks so much for doing this. - Of course. It's been great.