METR’s Joel Becker on exponential Time Horizon Evals, Threat Models, and the Limits of AI Productivity

21:29Joel Becker:I do want to comment on this whole thing about how you are, the threat assessment side is in your name. And typically I expect, let's say EA affiliated companies or organizations to be on the Eliezer side of the world where they're banging the drum about danger. Whereas here you're actually saying, actually, it's like a pretty balanced, like we care about AI safety, but also we're not there yet. And we are actually the watchdogs looking out for it. And I would say you stand out as someone not funded by the labs where, let's say ARC, is it ARC? Or some other groups that also do threat evaluations before model releases.

22:06Joel Becker:They would typically be funded by OpenAI or some other big lab. META came out of ARC, so I think. But now you're a separately funded organizationist. And as far as I know, it's like a big deal that you're not funded by the big lab. Yeah, I think it's vital to have this independent source of expertise. I can bang that drum forever. Yeah. Yeah. The other thing also is just like this concept of capability explosion, which is a word that you use. That's also something I wrestle with. If you believe in emergence, you believe in multiple capabilities fusing together to produce generalized capabilities that you may not be able to detect.

22:37Joel Becker:It's hard to predict based on trend lines. It should be discontinuous in some sense. And I don't know that going, oh, the N minus one model was fine. Therefore, the N model is probably fine. It's really hard to tell. So the thing that gives me comfort is yesterday I was at the OpenAI livestream and even Sam Altman was like, yeah, I just let Codex just YOLO dangerous permissions, whatever, on my computer and I don't approve the model anymore. Like it just does whatever it wants to do on my laptop. And I think, I guess the guard is every model lab leader dogfooding. If it screws up their personal permissions, then they have the skin in the game is what I'm saying.

23:17On the continuity arguments, I'm not sure what I think. I agree that it's flimsy or like this, there's only so many models, so many data points on this time horizon trend. How much should we expect it to be continuous to keep going like this? I'm not sure. Maybe an intuition that something might be discontinuous because models are providing so much effective labor in improving the next generation of models. Maybe that's a reasonable thing to think. On the other hand, I've been pretty surprised so far about the degree to which it's continuous. And that gives me some faith that it might continue to be continuous in future.

23:46Seems ambiguous to me. we have break points in physics right as i'm curious if it doesn't seem like when you it's funny it's like when you think about water right what does it boil yeah at this exact temperature maybe we do know but i feel like we don't really know and i feel like with models i don't know if there seems to be the same thing because it's all just like compounding of the same thing if that makes sense it's just like scaling the same thing over and over yep but yeah maybe we will see it But I'm curious like what you would need to see to feel that is here. Because even if you look at the Opus 4.5, that's clearly out of trend.

24:20And so you were saying four months instead of seven months. But if then the next month is so maybe it should not be four months, it should be two months. Would that make you change your mind about whether or not the months thing even makes sense? Or like we maybe we pass some base level after which it accelerates and we'll keep going. I don't know. I feel like you must be having this discussion internally. me. In some sense, the thing that would really concern me is if AR &D was fully automated inside of some lab, that would totally seem like the conditions are there for potentially a capabilities explosion.

24:51If I saw a time horizon of a year, I would still find it ambiguous, I think, at the moment, whether that was the case, because for things to be fully automated, 90 % automated isn't enough. You need some full loop to be closed, and perhaps we're missing some sort of task that points to that missing 10%. So I think it's a tricky issue. I think I can't give a number. But yeah, my intuition for where water boils is at some point where this loop is fully closed. There are interesting debates about what exactly that loop is. So some people talk about software-only intelligence explosions, which means even holding hardware fixed, we could get to the point where just from models improving themselves, they then be smarter in this next step to create even better models with even fewer resources, this sort of thing.

25:31And this could lead to some extreme takeoff. Or maybe that fizzles out some more quickly. And instead you need, in addition to the software only capabilities, you need chip design, or maybe you even need chip production. And that's this larger loop that can close. If you think that, I think you maybe should still think that closing the chip production and software only and chip design loop is potentially very stabilizing and concerning. But yeah, tricky issue.

25:54Joel Becker:I think that is the actual paperclip factory. If you incentivize a model to go build its own compute and it would just build whatever it needs and it will turn the planet into chips. I don't think it can do it. We will stop it before that. Question mark? I don't know if we have the power. There's no off button like this. I think it's super hard to foresee. But a model that had those kind of capabilities, it's hard to rule out there would be something like a capabilities explosion. And who knows what happens after that point. Yeah. Okay. So there's a bunch of other benchmarks that actually directly track this, right?

26:26Joel Becker:OpenAI has like PaperBench, I think, which directly tracks its capability to reproduce papers. And I think there's a lot of other than Rebench, There's a lot of other sort of similar sort of ML self-improvement benchmarks. Dave directly, like Yakun from ODI has like directly prioritized, like we will have an automated AI researcher. I did a podcast with Yitay from Gemini, who is also basically he's plugging his own training logs into Gemini to improve his own code. And I'm like, at some point, you don't need to be here. I think this year. I'm not speaking for everyone at Meta. I'm a relatively longer timelines, quote unquote, person at Meta.

27:03We have Nicola, my colleague, who helped out with AI 2027, who's on the shortest timeline's end. This is not a META view.

27:10Joel Becker:Which is officially AI 2028 now. One year past, we move it back to you. Yeah, I think my view would be not that Nicola's view is necessarily different, just so I'm not speaking for other people at META. A paper bench, let's say, perfectly measures not only reproducing papers, in fact, producing novel research papers. That's just a part of this R &D production process. There's also like your GPUs are constantly failing. Can you get someone to go to the data center and fix them in the appropriate way? Can you call up the water company when the cooling breaks down, et cetera, et cetera, et cetera?

27:40Not aware of benchmarks tracking that in particular. My point is more there's this very long tail of things potentially involved in R &D that would perhaps need to be fully automated in order to lead to capabilities explosion. I expect we're measuring in some ways only a small proportion of only a small proportion of of those capabilities and so I expect the capabilities needed for the full loop to close to to come somewhat later yeah that's a controversial view I don't think so I think

28:07Joel Becker:that's a reasonable take something I do that does surprise me in terms of when I'm talking to capabilities researchers is that you guys don't have an enumeration of the capabilities that matter. I think you implicitly do in your choices that you make, but I think it's almost important. Like I always imagine like the wagon wheel and this is like the terminology. I don't know who came up with this term, but like that, that here's like the 10 things we care about and here's where everything is on those 10 benchmarks. And I feel like capabilities tracking is just tracking, okay, what's that list? And then where are we on that list?

28:39Joel Becker:And I think I almost feel like this need to reduce everything to a single number is actively working against that because it reduces any form of nuance of it's insufficient here like the calling a data center thing so like we're fine and it's like actually we should just not invest anything in that area because that's the danger zone yeah i think i could not agree more that time horizon is for instance but many other single numbers is one number and that's like collapsing an enormous amount of really important detail I don't know how to come up with that list of 10. And I challenge you if you're able to come up with that list of 10.

29:14I'm working on it for code. I'll be very interested to see it for code. My intuition is that we'll come up with a list of 10 and it will turn out that there's a secret 11th thing that we thought was important, but it was difficult to pre-specify ahead of time. And now it seems obvious that even ahead of time, if we'd have that foresight, that would have been helpful to add.

29:29Joel Becker:I think that the security community does this by versioning year by year, right? So this year, the top 10 are blah, and it will just publicize it to everybody. so everyone knows what top 10 is. And next year we'll have a different top 10. It obviously is stochastic and we should update our assumptions, but it's broadly useful to have that list as a public service. You also had this research on the slowing AI improvements based on AI compute. And you mentioned that it's like, in a way you could tie the AI time horizon to like the growth in compute. Can you say more about that? It's in a way unintuitive because the compute growth is not always tied to like how much every single model compute needs.

30:09It's kind of like a broader market thing. Yeah. Yeah. How did you get the two together and then some of the findings that you had? Yeah. Maybe for a second, let's take Time Horizon very literally. We don't have the qualms about it that we've just been discussing. It makes sense to continue extrapolating it into the future. What are some important forces that might cause it to rise more quickly? Some of the things we've just been talking about, automated R &D versus go more slowly. One of the most obvious forces that might cause it to go more slowly is if input's slow. One important input is compute.

30:37I think we all have the intuition that to some extent, if compute growth slows, which we expect it to at some point in the not so distant future, then capabilities will slow. But by how much? It's a big question. The suggestion in this paper is that if you think that algorithmic progress, that is coming up with the transformer, coming up with RLHF, all of this stuff, better learning rate schedules is itself a function of compute because you need compute to discover it. the transformer, the gains from transformers show up much better with scale. If you don't, if you don't put in those resources, you'll never find out that this is, this is the superior algorithm.

31:13You need to run a ton of experiments. Each of these experiments can be quite compute expensive. Not to say that no labor is involved. Obviously people are working on this, but if you think it's ultimately bottlenecked by compute, then algorithmic progress too slows down, right? If compute growth slows down. So then if you think about time horizon or whatever your favorite measure of AI capabilities is being a function of algorithms in some sense and compute in another sense. And both of them, both of those components half when compute halves trivially because compute is halving and algorithmic progress halves because computers is this important input and compute halves, then you might expect time horizon growth to half.

31:49And then some of these major capabilities milestones that we might be interested in would be significantly delayed. I think there are so many caveats to that picture. I think there clearly are some types, at least, of algorithmic innovations that did not require a lot of compute to go about creating, some that took a lot more compute inputs. If you expect that no compute inputs are required, we could just survey researchers for the best ideas and then immediately put those into training the frontier models, then there'd be no slowdown of algorithmic progress from compute growth slowdown. And of course, all of this is counteracted by the possibility of capabilities explosions or AIs providing, even short of capabilities explosions, AIs providing significant labor at making AIs better.

32:33But just analyzing the compute force on its own, it might need to significant slowdowns depending on the degree to which it makes sense to call algorithmic progress, basically determined by computes versus not needing compute to come about. Do you think of compute on a per lab basis? Because there's kind of one way you can model this out is the improvements slow down. Not every company is able to stay in business and then their compute gets recycled back into the other labs, which then grow compute again. There's almost like benefit to like the heterogeneous distribution of like researchers and compute.

33:07But I'm curious, like how much you care about like just a broader compute computers out there for people versus the big labs of more and more compute. Yeah. So for the paper, we use OpenAI data and OpenAI projections. I think this applies more broadly, but we use that as a kind of case study. I think the argument I just laid out goes through if you're not interested in compute at all, and you just talk about dollars, what are the dollars going into? Going into models, will algorithmic progress slow if dollars that goes into them slows? The whole argument works. And that works, I think, at an industry level or at a lab level, so on and so forth.

33:38I agree things like certain labs going out of business or labs consolidating or these kind of industrial organization things would be very important. I'm laying out an extremely simple picture. Real picture is not extreme, but that's the basic.

33:51Joel Becker:We have examples of XAI has been said to be distilling from Claude, right? So like people kind of share compute in indirect ways, let's call it. I think it's also very interesting. I'm just kind of curious, like what OpenAI numbers did you have? Is this like the 500 billion for Stargate or something else? This is from their previous tax returns, the amount they've spent on R &D compute and then from a from information reports earlier this year some projections that open ai have for how much they'll spend on compute r &d in the future converting that from dollars back into flops yeah it's interesting because like back to flops sorry and obviously all the labs but particularly open ai in the last three months have basically thrown 10 billion dollars each to every single compute provider on the planet to develop alternatives to their current approach which is very interesting.

34:41Joel Becker:But I also say don't discount meta compute spend, don't discount XAI compute spend, and don't discount deepmine compute spend, all of which you have basically zero visibility, right? If you're looking at a single company, maybe that's authoritative, but then the total spend could be a lot higher. It's interesting. I do think, I do also observe that like people like Dylan from SemiAnalysis do tend to very strongly tie model progress with compute clusters coming online which is like the people on the model sort of api side don't see it but this is all downstream of our like 10 000 gpu cluster just came online and it takes six months to do it and therefore grok 5 will be here and like it's pretty mathematically like deterministic there yeah it's right to me yeah it's fascinating yeah they must yeah because from the lab side they must see something in the early checkpoints to like go ahead and keep investing 18 months from now Because I wonder what the time gap is between finishing a good pre-training run and going live.

35:40That's probably like nine months, 12 months, something like that.

35:43Joel Becker:I think Mr. Hall is actually pretty open about this. The plans for Mistal 3 and 4, I think they've been pretty open about the number of GPUs and the direct timeline from coming online to when they ship the model. It's pretty set. I don't have a clear timeline in mind, but I would say 46 months. But yeah, the competition is very tight. And one of those things is it's also very interesting to see when labs throw away models because they failed. Like their run was like came behind someone else's run that was better. Then they were like, oh, we can't release this anymore. Yeah, release it. Yeah. Yeah.

36:19That's the biggest risk with the prediction markets on model performance, actually. Just to tie back. I'm always. Failed runs? Yeah. Yeah, it's okay. I think in December, there was like the who's going to have the best model by end of 2025. I think there was like a lot of activity like in the last few months while the GPT 5.1 model came out. And it's okay. Then I guess Gemini, because they just threw that out. It means that Gemini is coming out next week. And so trade that.

36:44Joel Becker:Do we want to talk about manifold? Yeah, you were like the most profitable manifold markets trader. I mean, there's obviously like a lot of talk about insider trading on like these markets, especially in AI. I've seen it with a lot of the embargo news that we get. I'm like, man, people are trading like a million dollars on this market. It's like there's thousands of people that know the actual information. How, if you didn't have insider trading information, how would you think about modeling these things out? And do you think it's like a worthwhile thing? Like, for example, who's going to have the best model in three months?

37:16Do you think that's a prediction market where you can build some sort of strategy, alpha? or I guess the naive prior without any extra information is just in 2025 for what percentage of time did which model providers have the top model as measured by Time Horizon? You could do it for any old benchmark. I think that's something like 5 % XAI, 50 % opening I, 45 % anthropic. Don't shoot me if I'm incorrect. I think it's not the case that a DeepMind model was at the frontier of Time Horizon at any point in 2025. But yeah, yeah. Different things, different measurements. Yeah, maybe that's the same prior that you want to apply.

37:53XAI was coming online at the beginning of the year. So maybe naively you want to raise XAI a bit. Yeah. I'm always curious like when I see people betting on these things that are obviously not, there's no real basis to. I was just going to work on it. What's your secret

38:08Joel Becker:for a minute for markets alpha? Yeah, I see. So the secret, if you read this article about how I became the number one most profitable trader on Manafort, which sounds very nice and impressive. Like I must be so good at predicting things. But actually it mostly comes down to this one market where Manifold had opened up a charity program and the market is on how much is going to be donated through this charity program by the end of its first month. Okay, and the market opens, or I first see it five days in, and it's giving a kind of linear projection of how much has been donated so far. Let's assume that per day amount keeps getting donated every day until the end of the month.

38:42As a person who gives money to charity sometimes, I noticed that you can manipulate this market in a way, right? By getting more to charity and so moving it more up. And so I think the strategy was to a ton of mana, this fake currency that's used on Manfold, into the option that was above the linear projection. People keep betting against you because it doesn't look like that's happening. I haven't actually done any donations yet. Eventually, they cotton on to what's happening, that someone's going to make this donation to move it over the edge. And they're betting on that. And then I did it again into the next category.

39:12Once people had started betting on that category above the linear projection. And again, people bet against that and against that. And I mopped up those fake internet points. and then i think i did it once more as a bluff the bluff failed but the previous two worked out and then i ended up donating i can't remember exactly how much it was not so much something like five thousand dollars oh it's all for a good cause yeah yeah to give i think and won lots of faking net points on the market and so became the number one most profitable trader slightly legitimately there's nothing about that that was outside of the rules exactly this is called also i used to have less respect for my forecasting abilities yeah this is called prediction markets

39:47Joel Becker:with high agency as you actually go, the future is what you make it. So I think like, to me, the broader lesson is the classic difference between manifold markets and polymarket is that polymarket is only real money, right? And so is the whole fake internet points thing a worthwhile pursuit or a waste of time? Should people actually want to use real dollars? And maybe that's one question. The other question is obviously like prediction market ethics, which like I think always just it's going to indirectly come to assassination markets. It's just, even if it's like you, there's you banned the word like will someone die like some other proxy to will someone die will happen to be an assassination market yeah i'm good friends with the manifold markets co-founders i love them very much i do my view on the social value of prediction markets which was always the dream right it would be nice to have calibrated probabilities on events that matter this country going to war with that country think things that's things that really matter to people it'd be nice to have high quality information but when i look at real examples that have come out in the past year it doesn't seem to me like those examples are so socially valuable i'm not sure about assassination markets in particular i'm sure that i'm sure those would be outruled hopefully but i think gambling like behaviors are socially costly and the value of higher quality information is is real but is it worth that disbenefit of of people trading away their money it's not it's not so clear to me yeah yeah price discovery has a cost and sometimes that is gambling and that is the stock market like it funds a lot of corporate america yeah a lot of the stock markets like i used to be one of those big firms playing against other big firms it doesn't have the same versus versus sports betting markets take an example on the other extreme has this very different character you might imagine that at least one direction that prediction markets could go is this sort of big players playing against retail and that maybe has a more worrying dynamic i'm not i don't get such so closely in touch with the space but at least something like that you can imagine being concerning i think at a large enough scale it becomes profitable for some of the companies to do it they can get the markets on it i think now it's still small enough numbers compared to the rest yeah which company is the best ai model by the end of january 28 million dollars of trading volume oh my god wow so why are people trading 20 like people it's just crazy but i think there's some i would i was having dinner with somebody this weekend i think it's totally possible to i'm not going to do it i think it's important that meter employees not be not be making bets on prediction markets like that but i think it's totally possible in principle to have a guess for the answer to these kind of questions oh but other people know exactly gemini three score on frontier math benchmark by january 31st it's like some people at google already know what the number is yeah you know what i mean it's i think if you believe in the benefits of price discovery then this is this is legitimate yeah they actively encourage insider trading and this is a way for insider information to pseudonymously leak out and as long as like They bear the consequences of leaking information.

42:35Joel Becker:Whoever is traceable to that thing. And people, I think, have been fired for trading on insider information. That's okay. The only step up from that is the government coming in and saying, this is actually illegal, put you in jail for that. But I think for now it's self-policing. Retroactively, you can't really do that. All right. If you have any embargo news, press. We do work with people on embargo and we don't trade. Yeah, we would have made a lot more money trading on embargo news than we've made on anything else. What else? What are other interesting model evaluation trajectories or anything that you're not doing at Meter that you've maybe seen other people do that you find interesting or like you would like more people to do?

43:16Yeah, one project that I think is interesting is AI Village that I think possibly both of you would have come across. These are these very open-ended goals given to a village of agents and they try to accomplish them. I think set up a merchandise shop is maybe one of them, organize an event in a park, build a human subjects experiment this sort of thing. I think I have a number of questions about exactly what I should learn. They're using old models as well as new models in this quote-unquote village. The models are relying a lot on vision capabilities, which we spoke about models not being so capable of today, this sort of thing.

43:46But the vibe of models trying to achieve open-ended things instead of benchmark like tasks, the vibe that's a bit more like that vending machine bench in some ways, seems like a very interesting direction to me for the science to go, or seems like something that comes with a lot of cons, but attacks some of the cons of benchmarks in a pretty interesting way. I think seeing the ways in which these models trip up, seeing the ways in which they're in which they're derpy is an important source of information. I'd be interested in more work like that coming about. I think that's one of them. Another is transcripts as an extremely interesting source of information.

44:20This is the models taking actions and then seeing outputs and then using those outputs to commit the next action and so on and so forth on benchmark style tasks or even more interesting on in the wild deployments like you might find on your own code code usage codex usage curse usage etc that has the con of being less experimental less less clean and scientific in some way it's it's more selected quote-unquote like the tasks that you get ais to do are obviously the tasks you expect it has some chance of succeeding in so you're not just giving any sort of task if i see the models doing something extremely impressive or or potentially unsafe in some sense subverting user preferences it's not clear how often that kind of behavior would happen given in the previous history, but it's a massive data source.

45:02There's a huge amount of information there. And I'd love people to be working more on that sort of thing. As we mentioned, there are a lot of problems with Time Horizon, our developer productivity work. I think it has been important evidence. I think it's moved the field forwards, but it's far from perfect. I think there are lots of other directions there that look very interesting to me. Maybe one that I'll call out there is this difference between whether models pass unit tests, whether they succeed by SWE bench-like scoring, kind of meter-like scoring, benchmark-style scoring, versus whether their solution would be merged into main.

45:36That is, whether the solution adds tests where it should or doesn't, whether it follows existing patterns in the code base, whether it makes sure that its changes speak to other parts of the code base in inappropriate ways. That seems very interesting to me. I think model capabilities probably are lagging behind there somewhat versus that which you might see on SWE bench like scoring i keep going on but these are the novel research things that you were

45:58Joel Becker:these are novel research things yeah just to comment on the average thing first that you mentioned a lot of stuff i even want to double click on the transcript stuff the air village ties back to our one of our highlights of last year which was known brown's conversation that he's actively working on multi-agents that are cooperative instead of competitive and the basic idea that we can do more as a team than we can do individually or like the the agents are the friends we made along the way. And I think that's great. And I think on the DeepMind side, the way they phrase it is literally have an open-endedness team, which I think is a topic that re-merges once a year.

46:32Joel Becker:I just like, yeah, I think it's unclear what open-endedness does for us. And this is a core divide in terms of studying these things as life forms, potentially new artificial life forms versus tools that serve us. and maybe their open-endedness means that there is no goal and if you're just trying to eval this as what does it do for me that's completely wrong and you will never get anywhere with that because they are just living their lives as artificial life forms in some sense that the gold standard evaluation that's that i would like to do if i was looking to learn most about the questions i'm most interested in the degree to which um aIs might automate or accelerate air and d i'd quite like to just give the ai a bunch of affordances type into the ai automate aro and d go and see what it does and i suspect that wouldn't work today even with all the affordances because it would fall over on its face and in in when working with resources handling handling resource use in ways it's not so capable of today it would struggle at some types of long horizon tasks etc etc etc and in some ways i think benchmarks have a face difficulties in capturing this sort of thing and And AI Village seems like a, or AI Village style things, these more open-ended goals, seeing how models pursue open-ended goals.

47:46They give some color to this sort of thing, to seeing models fall on their face. We think that this will become these more open-ended goals, more and more important over time. I agree to some extent, like you're going to provide them in the extreme case that I just mentioned, you're going to provide them documentation about how this part of the company works and that part of the company works and so on and so forth. It's not purely open-ended, but it's pretty open-ended. It's more open-ended than the kinds of problems that we're giving them today. Yeah, and if models are excellent when Spex uses them, given some detailed issue description and some very clearly specced thing on what they're supposed to do, that's interesting.

48:20But it's a very different thing, I think, from being able to automate AR &D. I'm interested in how far we are away from that. And in some ways, this speaks more directly to that sort of thing. So we had the terminal bench guys on the podcast. How do you think about the harness benchmarking in a way? because if you look at their leaderboards, the same model with different harnesses, there's like 10 percentage points of difference. Does that seem interesting? I don't know how you build the harness in meter or whether or not you always pick the best harness or compare them. Yeah, let's say how we pick harnesses at meter.

48:50This is not what I work on in particular. I'm not an expert, but roughly we build harnesses to get models to be as performant as possible on a dev set of tasks, some held out set of tasks. And then we use those same harnesses trying to make sure they're not overfit for our main suite of tasks. On the one hand, I do have the intuition that there's a lot of juice in scaffolding. It's easy to overstate how much juice there is because of this overfit problem. Or if we were building a scaffold to do as well as possible on our test tasks, then it would do much better than the scaffold that was built only on our dev tasks.

49:23And in some sense, that would feel illegitimate or not interesting, or you wouldn't expect that to generalize some other set of tasks, potentially. On the other hand, a lot of work has gone in meter to building scaffolds that are as make models as performed as possible because we are interested in up-abounding the capabilities of models when thinking about whether these models might or might not be dangerous i do have faith these scaffolds are a lot better than the first thing that people might try because so much effort has gone into them yeah it's interesting because i do want to overfit as a customer of the models like you do want to overfit to your task specifically and i think sometimes people underestimate maybe how much value you can get out of it but yeah i think if you have a kind of mechanical workflow or something that you're imagining automating it you imagine automating that workflow and there's some place where more sort of stochastic intelligence would be nice inside of that like deciding where to route customers to on customer calls something like that there i feel like that makes a lot of sense but for this sort of more general in particular thinking about helpfulness and software engineering thing i'm not sure i have that same yeah take an example it's like i work in typescript yeah if I build a better linter that is private to me or a better test suite, like a better play replacement, like in theory, I'm overfitting the model to perform better, right?

50:34It doesn't really matter to me. I'm not trying to report on the model performance. I'm trying to build the best thing. Yeah. But if you have a model to build the linter. No, I agree. I agree. I think that's like the question of, okay, should I just wait for the next model? You know what I mean? It's like, at what point should I be building the better scaffold? Like no one brown, it's all scaffolding. it's going to get washed away. Yeah. But on a realistic...

50:57Joel Becker:He would say that. On a realistic schedule is what am I supposed to do this week? Yes, those simultaneously can be true. That all scheduling will be washed away and that scheduling today is valuable. Totally. Within model generation, it's valuable and across model generations, it's not so valuable. Yeah, I'm, I'd say at best, an acceptable software engineer. Intentionally not investing in engineering skills because the areas are getting so good. Maybe that's the wrong decision. Yeah, I agree. If you expect, as I think you should, capabilities to keep going up and up forces difficult trade-offs about how you spend time today, because maybe it won't be so helpful in six months' time.

51:29Take a sabbatical. If you live in Europe, you can just take six months off or something. Yeah, but then in six months, you might want to take another sabbatical for the next week. Perfect.

51:38Joel Becker:Cool. Just to wrap up, what do we expect out of Meter in 2026? What does success look like in 2030? I don't know if there's like a sort of broader vision. And then maybe on a personal side, we can talk about the karaoke stuff. But I've got Meter as well. Yeah, so from Meta, I think you're going to see more, hopefully, high quality capabilities evidence kind of thing you saw in the past with Time Horizon and the developer productivity work along the lines of what we've been describing. Some of these future research directions. We also have some monitoring research directions that I'm not so expert in that is thinking about if we can successfully apply safeguards to models attempting dangerous tasks.

52:14Joel Becker:There's a whole line of work there. Is that an interpretability dimension or what kind of safeguards? Usually this is black box, not white box in my understanding in current work. So not using interpretability, but you can imagine in principle doing something more white box. And then this risk assessment work that is taking into account how capable we think models are, what their propensities are, whether we can track using safeguards, the kinds of things that the models are doing. Do we think that these models pose large scale harms? You can expect to see much more of that in 2026. Maybe now is a good time to say that we are hiring.

52:45Yes. on my team we're hiring for research engineers research scientists people from startup backgrounds from from ml backgrounds of course i'm originally from sort of economics or quantitative genomics backgrounds so we are accepting a pretty wide range of people who get stuff done the kind of stuff that you've seen in in past meter work as well as a director of operations i think on the

53:04Joel Becker:meter jobs page you can find out more yeah everyone i know is hiring a director of operations including myself and i feel like that's probably the one agent that everyone wants that we cannot have Yeah, whenever people have a hiring pitch, I always try to push for, okay, the average candidate comes in, you reject them. Why? What's the thing you're looking for? So there are lots of different shapes of people we can look for. So different stuff for different folks. One thing is good kind of basic research intuitions, like checking your data. We don't work on free training at Meta, but if you're working on free training, you should look at the corpus to get some sense of what's going into the models.

53:36Even working on this uplift RCT, that was pretty important. Really having a shape of these issues in your head. I think people who are communicating in writing with sort of a lot of transparency, not overstating their results. My hope is that your sense of meter work in the past is that it's trying to be level-headed, not to understate, not to overstate what the science says. That's important internally. It's important externally. And then I think productivity or something that there are a lot of people with great talents who are not going to work quite as well in a scrappier environment working on frontier science.

54:07And that's the thing we do.

54:08Joel Becker:I just want to prime people for what are the valuable skills in this new age? Because I think the more people articulate what the positive directions are, what it's hard to hire for, that's what we guide our audience towards improving themselves. And I think that's important. Oh, yeah. Did you have a karaoke question or? I don't know. Are you going to sing on podcast? I've never done it. I've been. Can't help falling in love with you. What is this karaoke thing that you organize? This is like a music. You're a musician. A musician might be exaggerating it, but I hit instruments and noises come out.

54:41So I've hosted a couple of these live band karaoke events that is like getting a group of friends together and people accompanied by a band singing karaoke to an audience of 50, 100, 200 people. It's great fun. I think people should be doing more of this. I look forward to seeing you both at the next one.

54:55Joel Becker:I will do that at one of your events. Yeah, it's one of those things where it's weird because I used to be in a cappella a lot. Oh, wow. And I just think it's like a dying form. I just watched this video that was really good about like 2010's wave of acapella from like Pitch Perfect. Glee is a Pitch Perfect to that's that group. Pentatonix. Pentatonix, exactly. That's where it died. And it's very interesting to see how it's like dying as an art form in general and like how new formats have taken over. And I don't know, it's weird for humans also, because like now I'm also, let's call it more interested in synthetic song generation or DJing, anything like that, that the human voice is actually more commoditized.

55:35Joel Becker:Like it doesn't really matter who sings it. I don't know. I feel like there's a kind of transcendence to singing in person that the AI generator songs are not providing me. That's good. That's good. Yeah. I do think that like we humans always want that. But I'm not sure humans in the year 3000 will want that. It's one of those weird things. Thank you for coming on. It's great to have you as a human in person here. Thank you so much for having me as a human.

56:09Joel Becker:Yeah, someday we'll interview AI versions.