When AI Improves Itself | Richard Socher (Recursive)

29:32It's called ProGen. Ali Madani is the first author of that paper. That was back in the day when I was the chief scientist at Salesforce still. And he's since started ProFluent. They've now closed like multi-billion dollar contracts with Eli Lilly at ProFluent, his company, because they've created new kinds of proteins that are, for instance, even better than CRISPR-Cas9. and gene editing and being even more specific and targeted for changing certain genes inside living people potentially and creating new kinds of therapies from that. And so proteins being such an important piece of all the building blocks of life and disease and health, making them programmable will unlock very, very obviously many, many exciting use cases.

30:24And I think you're starting to see this sort of in this recent trial that is making a lot of progress where they basically created a different drug for every different patient in the trial. And this is a first for the FDA too. And more will happen there. It's actually unfortunate how hard it has become in the US and certainly in Europe to run clinical trials. And so a lot of folks are now moving to either China or Australia for their clinical trials. interesting enough in china it's cheaper it's faster but you also have to worry a little bit whether your ip gets sort of sucked into the ether and is gone and in australia they had a clever move where they actually decentralized clinical trials and every hospital can run its own clinical trial so all of a sudden you get competition instead of having one centralized um uh sort of decider on on which clinical trials to run and how to sort them and uh and all of that and so yeah anyway there's not enough people in Australia, so it would be great to get that kind of system happening in the U.S.

31:23too. But yeah, clinical trials will be more and more efficient over time. We'll collect more data and NDI will be able to automate more and more of that. And when you think about the drug discovery and creation life cycle from initial intuition to being available that take, 10 or 15 years, what are we talking about here in terms of accelerating discovery? what realistically what portion of the process does it shave off it's a good question and sort of touches upon uh what some people call the hard takeoff too where some people think once we have RSI and and generally with AI there will be this really hard takeoff and then everything will just happen very quickly and as bullish and excited as I am about AI I'm not a believer in this crazy hard takeoff I think yes things will accelerate but there are certain things that will just require time because of physics and constraints in the real world, such as like long-term trials that you want to know whether people have some issue like three years after the, you know, they stopped taking the drug and things like that.

32:28And so there will be some delays, but the biggest difference is that the whole biomarket and somewhat contrarian take that we have at AIX Ventures too, a lot of folks think that like bio is just a terrible space to invest in because in the past, a lot of drug companies kind of spent eight, 10 years. They finally get, you know, they have to be public because there's not enough late stage bio investors. So they go public with one drug or maybe two drugs in late stage trials, like stage three, and then the stage three trial fails. And then the whole company is dead. Now, what we're seeing, the difference is like we now have companies that instead of having one molecule or drug after eight years in late stage trials, They actually, within six to 18 months, have multiple different drugs in late stage two trials already.

33:21And by the time they'll go public, it'll be with eight plus different drugs that are then also much more likely to succeed because we have better predictive models. So clearly we are living in a moment when AI has become quite controversial, whether that's the job question or the data center question. The number one thing the industry keeps saying as a way to justify why AI is a great thing is AI is going to cure cancer. What is your sense of the reality of that claim and what it's going to take to get there? There's a lot to unpack there. Maybe at a very high level, I think if you care about the outputs of an industry or a company, then you love AI.

34:09If you get paid hourly, you probably hate AI. And so AI in the positive instantiation of this future is a huge force towards more entrepreneurial thinking. If you're an entrepreneur, generally you kind of love AI because it's making your things more efficient. You just get more done. You have to like, you have an unlimited list of things to do if you're a startup founder or just running a company. And to have any, I do many of those things for you just means you can do a lot more. But if you're like basically being told you're training your replacement and you're like all your data is being collected hourly, then you know at some point those hours will end and then the I will just do the thing you just taught it how to do.

34:53And so it's understandable that people, if they have this very unentrepreneurial mindset of just getting paid by the hour and they don't own any equity in creating that IP, then they're understandably unhappy. And then you can go one level deeper and think about, well, what is the impact on jobs? And my theory here, after thinking about this for quite some time, is largely dependent on the elasticity of the demand of the product when its prices go down. And so concretely, for instance, illustrators. Illustrators hate AI. The world needs a certain amount of illustrations. because of AI, you cannot charge 200 bucks anymore for one illustration.

35:39So now any little blog post has illustrations. If your goal was just to see more illustrations in the world that are specific to a text, you love AI. If you got paid 200 bucks for one illustration and now it's worth two cents maybe, you hate it, right? And so the problem was that the demand for illustrations didn't go 1 ,000x when the price went down by 1 ,000x. didn't grow because you just don't need that many illustrations in the world. Now in coding, it was a very different world. Actually, as coding got cheaper and cheaper, you had this famous Jevons paradox that everyone's talking about now.

36:15I think I was the first, at least I didn't see it online for a while. It's like an interesting fact from history. And you actually like the thing got cheaper and cheaper, but we actually used more and more of it. And for coding, that will definitely be the case. And so we're seeing actually more demand for programmers now because they're so much more productive when they use AI. and anyone ultimately could have like dozens of apps on their phone that are unique to that person that are modified in some way and very special. And there's so many other ideas that people didn't explore because it was like, maybe the market wasn't that big, but now that you can just create an app really quickly, why not?

36:49And so I think that is another aspect of jobs. And so go back to cancer. Yes, I do believe actually AI will play a big role in curing multiple cancers. We are seeing trials now where AI is being used to make a specific cocktail of drugs, create specific RNA sequences and so on for the types of cancer. And, you know, each cancer often is also not one like homogenous thing. It has different types of sub cancers in it and so on. And you need to specialize treatments for each person and for the various different forms of the different cancers that you can have. And so all of that is much, much more feasible to be done with AI.

37:31And we're seeing it. Is that precisely the point that cancer is just extremely complex and ultimately a system problem? Exactly. That AI is uniquely equipped to solve? Exactly. So, yeah, it will take some time. And obviously, like, even if AI, let's say, had the perfect molecule and like, okay, for this type of cancer, this is the molecule that came up. Like, and I came up with it, you'd have to still run it through many clinical trials. It'll still take years to come out. So everything in biology just takes longer than it does in software. What else are you optimistic about in that field? Your predictions for the next decade, rare disease cures, organs designed for individual patients, pollution eating synthetic cells.

38:16You cover some of this in the book. What are you most excited about in terms of what may come first? Yeah, I'm excited about all of these things. I think we can design bacteria that eat microplastics. And once there's no more plastics, they just die. I think that would be extremely helpful for the oceans and so on. Obviously, you have to be very, very careful that they don't somehow mutate into eating other things and so on. So when you mess with the environment at large scales, it's important that humans have done that many times. And sometimes it worked out pretty well. Many other cases, maybe not so much, like forests are a good example.

38:49People deal with forests too much. They don't let small forest fires happen. And then they get even bigger because the small ones didn't clear out the underbrush and so on. Everything is a system. Everything is a complex system. And we need AI more and more to do some of that engineering better than we've done in the past. And so I'm excited for at all the different levels. You know, when you look at like how to balance plasma and tokomaks for nuclear fusion, that's already a machine learning control problem. I think we'll have a better handle on that. So sort of at the lowest level of physics, clearly there are more and more materials, more efficient solar cells and solar panels that we can design with AI.

39:29There's companies I've invested in that do that, better batteries, better materials. So we don't need only lithium. We can try to build batteries with more abundant molecules that are easier to get and mine of less pollution. We, especially again in biology, seeing a lot of things. I think it sounds like science fiction and I understand. So the famous saying of like, if you want to know why something doesn't work, ask the experts. I think that that was true in natural language processing and neural nets. And I think it is currently also true for longevity and cancer and other kinds of research for neural nets applied to biology and medicine.

40:08I do think we'll see, we'll make more progress than the most skeptical people think. But we also want to have a heart takeoff again because things do require careful experimentation in medicine. And I'm personally excited for all of these things. I think if you're mostly interested in like making humanity more productive and more efficient and create more outputs and grow, then you're going to love AI. But I also, in some ways, it becomes a philosophical question. And I think you've already, we already observe many sub-civilizations or like, you know, subgroups of people, I mean, and cultures that have essentially off-ramped from progress.

40:50Like if you're living on some beautiful island in Greece, you don't really think about AI. You don't have to think about AI and you just enjoy life. You go fishing and, you know, sometimes there's a storm and things are bad, but most of the time, like the weather is good, the fish are abundant and you just kind of live your life. And so I think there will be different groups of people who will want to off-ramp from civilization progress, right? There's already, you know, people prefer to live way deep in the countryside and never go into the big city and so on. And I think we'll have more of that.

41:23And in some ways, I personally love progress. I think scientific progress especially is what helped humans solve most of the hard problems that were in our biosphere. David Deutsch has a whole section in his book, The Beginning of Infinity, which I highly recommend people read too, where he talks about, you know, how there are all these different material problems. And we came up with solutions thanks to science and better explanations and better research. And I'm personally all for that. But, you know, some people will not want to participate in that world anymore. And I think AI is such an accelerant that it makes that question even more pertinent for people.

42:02That's fascinating. I mean, just to keep going down that path before we go back to our little tour of frontier science, how would that manifest? I mean, so we would end up with groups of people that would deliberately opt to just not participate in progress. I guess progress has been sort of jagged throughout humanity in different regions, obviously. But as it spreads and as the world keeps going more global, those people make a political decision to organize around a principle of non-participation in AI. Yeah, I mean. Like is that city-state, that kind of stuff? Yeah, I mean, like a sort of example that I'd love to visit actually is Bhutan.

42:44Bhutan decided we will not measure our gross domestic product based on money, but based on happiness. And happiness mostly for people who want to keep it simple and have a simple life, not want to build startups and so on. I'm pretty sure those folks aren't quite as happy in Bhutan. But like overall, Bhutan is just very green and like it cares about the environment and cares about like a specific subset of religions. And like and people are more often content in keeping things the way they are rather than trying to like progress in various different ways. This is why this book and this conversation today, from my perspective, is so important.

43:24And I think the AI industry has done a terrible PR job in general. so if you and you know others can clearly articulate why is good that may hopefully unlock some of this debate yeah it's really interesting because clearly people use the technology it's like if no one used chat gvt or cloud code like there wouldn't be a problem people clearly like it it's just that the people get a lot of use out of it are not quite as vocal and you are there are negative things there's also some amount of moral panic about chatbot friends, similar fashion to how novels used to be a really bad thing. Like there's all kinds of stories of older people saying, oh, these novels are ruining the youth.

44:07They're now living in these dream worlds and are distracting themselves from the real world. And like, you know, like the Leiden des Jungwerters is a very famous book in Germany, actually led to some suicides, is really sad. And like now it's like the book every German kid has to read in high school. And it's just like a high form of literature in Germany. And then comic books were really bad and computer games were really bad. And like, you know, there are various sort of levels of that. And currently the chatbots are really bad, but there are also clearly a lot of people get a ton of value out of these chatbots.

44:37And now all of a sudden you make access to like medical advice cheaper, legal advice cheaper, and sometimes also emotional advice cheaper. But you don't hear many people like, or the many people that clearly exist, who are like hundreds of millions of users of these technologies talk about how much this helped them not commit suicide or something or not be very sad and dysfunctional and so on. So I do think you're right. Like in some ways, not just AI, but I feel like the future as a whole needs better marketing. All right. Going back to our tour, because I want to make sure we cover some of the fascinating parts of the book.

45:13So we talked about drug discovery. We talked about computational biology. Another fun example or domain that you mentioned is economics. with a fun stat where you said economists failed to predict 148 of the last 150 recessions. And so your team, while you were at Salesforce, built an AI economist that basically operated on a simulated society. And you came up with policy recommendations that were better than the state of the art, quote, end of quote. Walk us through that. Yeah, economics is a really interesting field that unfortunately doesn't have obvious benchmarks the way computer science and many other sciences have, where you just say, if you do better in this benchmark, you clearly have the better ideas, the better algorithms, and we should all learn and study those.

46:06when we submitted these papers and two-level reinforcement learning systems to nature and science, they just desk rejected them. In one case, some random like ethicist who had no idea about AI, it was just like desk rejected. I'm not even going to read the full paper because AI for economics with reinforcement learning is just a weird thing. And so it was just like gone. And so because of that, economics often becomes just a political field. And if you're in one economics department and has a political certain slant and direction they want to see the world move into, you just have to write papers that make sense for that political ideology.

46:46And so that unfortunately makes it very hard to do more objective research. And so we try to create the simulations, a very, very simple simulation where you have a bunch of agents, you know, like this is from 2018. The agents were much, much simpler back then. They just had a certain utility function. They had certain hours in the day that they would be willing to work. They were sampled from certain priors that you may make assumptions about. Not everyone wants to work 14-hour days, but some people basically make all these assumptions. And then you let these agents collect resources, build houses.

47:21They can block other agents from those resources to try to build monopolies and become even wealthier. And then you had a sort of meta agent that looked at all of these other agents and basically chose how to tax and subsidize different groups of agents. And in that fairly simple simulation, you could essentially give it an overall reward. Like in our case, we said, let's maybe start with equality times productivity. You want the economy to grow, but you also don't want one agent to have access to everything and everyone else is really poor. And so obviously you don't want just equality and you don't just want productivity.

47:59So you have a combination of these two multiplicatively. And now if you agree that that's a good reward, you could have politicians say, well, I'm going to do this and that to help, for instance, the middle class or like to do this and that. But if we had a much larger scale up simulation, you could then run their one proposal through billions and billions of years of simulations and of taxation and subsidization to say, well, will that proposal really result in that outcome that you say you have, the goal that you have? Or maybe, probably, if you simulate billions and billions of years of different tax years, maybe there are better ways.

48:40And what we found is that the agents will try to avoid taxes by dumping a bunch of stuff before or making like a bunch of gains just after the tax year and so on. And the funny thing is that paper, basically the baselines that the field uses, one very famous formula is called the Sayes formula in economics. And basically it's beautiful math. And it shows that provably it's the optimal taxation scheme, but it's the optimal taxation scheme in a one-step economy where you make one economic decision and then no other decision again. And so we showed that this very complex RL system basically recovers that thing and does come up with the same solution.

49:20But now you can actually deal with the fact that economics is a temporal sequence of many different decisions and you can learn and adapt and there are counter adaptations from the agents to certain taxes and subsidy schemes that are trying to play things and then you can still simulate it. And so my hope is eventually that that paper will have kind of a GPT-3 moment where someone actually scales it up, builds a really realistic simulation and then we could have AI give us feedback. Obviously, you don't want to let the I make those decisions without any human oversight, but at least have some economic policy suggestions on how to most objectively try to achieve the goals we want to set.

49:58And of course, humans then have to really formalize kind of what is the goal of our society. And in many ways, these are very deep questions that philosophy and political philosophy have asked many times. Socialism, capitalism, like maybe social market economies where some regulation in like healthcare, but maybe not in other areas and you want competition. You can actually define once like what your real goals are. So I think hopefully over the years, this kind of system will help us run economics much better and make it a much more objective science. Do you think that's realistic that we could model all of the economy, you know, with all its nuances?

50:40There is an emerging space around simulation of worlds and, you know, a couple of exciting companies in the space. But at the same time, the economy is a lot of rational decision, but a lot of irrational stuff. It's very human. There's fears, there's greed. Can all of this be modeled by AI? All models are wrong. Some are useful. I think we can make those models more and more useful and they'll be less and less wrong. I think we've seen surprising results where you can prompt an LM and say, you are now, you know, a 43 year old, like from this region, blah, blah, blah, blah, blah. Give them all kinds of sort of prompts on what they're supposed to act like.

51:25and then after having trained on tens of trillions of tokens on the internet you can say similar things to what people might say from that setting and so I do think these models will get better and better the fidelity of the simulations will get higher and once they cross a certain threshold then the recommendations from such a simulation with an AI could become more useful I don't think this is very feasible in the United States for a very very long time it's just so much identity politics and special interest groups and how, you know, super PACs and so on are like get funded. That is very, very unlikely to be used.

52:01My hunch is like Singapore or China will probably be more likely to try to use those ideas. Say, hey, we all agree or we at least make it very clear that this is our objective function. And then, you know, we're going to really try our best to set the various taxes and subsidies and so on in a way that really achieves that objective function. Okay, great. So we talked about drug discovery, computational biology. That was the economics aspect. You talk about astronomy. You talk about neuroscience. So again, I strongly encourage people to read the book and hear all the stories and all the nuances.

52:39Let's talk about the European machine itself. You alluded to four stages. And maybe as we get into that question, there is also the question of the quality of the data that is fed in all those machines. Because if you train AI on a lot of AI data, don't you inherit all the biases and the assumptions and all the stuff that is just wrong, that is spread out through all of human history? Yes, I think AI often is only as good as the people, the data, the systems, the infrastructure, the rewards that we give it. And we have to be very careful about how we design and filter out all of those things. I think we have more and more control over it, but it is still surprising how poorly engineered some of the environments are and some of the sandboxes are that Frontier Labs use for AI.

53:41So let's get into the machine itself. So you got four core pillars. Walk us through the first one. so uh yeah the four pillars i briefly alluded to them earlier um where like the first one is just large language models essentially to try to ingest the world's knowledge uh into uh the eureka machine and uh i think the interesting bit here actually is that in some ways uh there's this weird cycle that happened uh that don't talk about it in the book as much uh but it sort of lived through this now, which is the few large closed labs, Anthropic and OpenAI, took almost everything they could from the open internet, trained a model, but then the Chinese open source companies basically siphoned a lot of that knowledge out of those closed source models by distilling it, but then they open sourced the model back into the open domain, so now the knowledge is back in the open internet where it started.

54:46And so I think it's very clear that that first pillar of just like having access to all the world's information, being able to reason through all these different concepts and the crazy large commentatorial space of existing knowledge is the first pillar. So pillar two is a model of reality itself. So what does that mean? If you think about how limited human perception and the current set of human knowledge is and how we could actually expand that, you have to look at scientific measurements, right? We cannot observe gravitational waves. We cannot observe sort of gamma rays, but we can build tools and scientific machines that measure these things for us.

55:33And so that is a clear second pillar that is different to human knowledge that in some cases hasn't been fully sort of described in human language. And in some cases might be very complicated to describe in human language. Like we can already say, oh, like a neural network predicted this word because of these 5 million parameters. But it's like, okay, well, you just list them all out, but you don't gain an intuition because the system is so complex. Similar to how no one can really say, why did you move this muscle fiber in your pinky when you try to move the steering wheel? No one has access to that in their brain.

56:08And even if they did, it would just be like because of this very complex system. And so that is basically the ability of an AI to take in all of these measurements and try to start actually like digesting it and taking real knowledge, extracting it from scientific measurements. And PLO1 sounds like it already exists. Does PLO2 exist? How do you teach a machine the rules of the universe? So one, you'd have to really collaborate with a lot of different sciences to put together this kind of foundational model of physics, chemistry, biology, and larger and larger systems, and then have many universities and labs work together to bring all of that into one model.

56:56So I think, you know, we've had sort of the projection of the humanities knowledge onto the internet, but there's just lots of things that just don't make sense to put up on the internet. And so those things are still hidden. No one, people have many, many companies are now working on quote unquote foundational models. Some of them now rebranded them as world models when there's a similar technology below where they basically try to ingest as much information about one domain. and then they build like a first example of a virtual cell that is particularly good at estimating like particular gene variants or something, but not lots of other aspects of a virtual cell.

57:32A virtual cell is a good example of like a goalpost where many different teams would have to come together and bring all of that data into one unified model. No one in its full glory, that doesn't exist yet. Okay. And then pillar three, again we're describing the the the four pillars of the york machine what you call the york machine which is this super powered ai scientific discovery machine so pillar three is simulation so that that goes a little bit to what we were discussing about economics so would you create different simulations for different domains or one simulation for everything in a perfect world we'd create one crazy simulation for everything but it they're obviously sort of different levels of abstraction and sometimes for most aspects you actually get away with not having to simulate all the quantum details of a very complex like subatomic particle you can just say all right these are the molecules and then you you know how in chemistry those molecules will work together based on the valence shells blah blah and then in biology you sometimes just can abstract from like oh i don't even care about that molecule i'll just say this is overall this protein and that protein and has like connects to a cell at that level.

58:46So as you try to build it all together, but then you have to have sort of computational efficiencies and abstractions that humanity has been good at building and computer science is particularly good as a field in building. Like no one has to program in zeros and ones anymore. They can now program in English and a lot of the abstractions can be ignored. I think similarly in these physical simulations, we can ignore more and more levels down. But sometimes like there is sort of quantum biology and there are maybe some effects that we didn't realize and we're oversimplified. And those might come out from a model where you're like from one large simulation in which the AI can then try to experiment.

59:23And that level three or pillar three exists in bits and pieces? In many small bits and pieces, right? The simplest example is like a simulation of Go or chess. And that's like, okay, we have it, it's easy. An interesting new one that many people are working towards now is a virtual cell. If we had that, I mean, a virtual cell is so complex, like a real human cell is so complex. We're very far away from that. But I can see how with enough people coming together with enough funding, we can eventually get to a fairly useful model of a virtual cell. Okay, great. And then pillar four is the real world.

59:59That's right. At some point, like especially in biology, but in all other fields, you have to engineer a system. You have to really put it together to see if you missed anything in your simulation, any confounding variables and so on. you have to really run experiments in the real world. And obviously in the smaller case of physics, chemistry, and biology, you can do that in a lab. At some point you have to, you know, build real machines and really get out there and build satellites and whatnot and take measurements of, you know, the universe and all kinds of scales. And so, you know, there I think we, like, it makes sense for us to put more and more resources behind that as AI has gotten really, really good in the first three pillars.

1:00:38So there's some sequence to it. And is the future a concept of self-driving robotic labs? And if so, how far away are we? You know, I love that there are like first efforts in this, like periodic labs is a great example of that. I love that we're starting to think about this. Personally, from an investing perspective, I feel like it's a little bit early, but in like two to three years, I think it'll be right on time. We'll have figured out a lot of the software, will have gotten, will be really good at, you know, the LMs, the scientific sort of data and like connecting that also potentially to LMs, like building even more high fidelity simulations.

1:01:19And then we can ask the AI to ask, like to come up with really good, expensive experiments that can take sometimes hours or days or weeks to really run through. We'll have better organoids or like tiny cell systems where you can, you know, base the human derived stem cells, a parallel bio disease, for instance, for human lymph nodes, like immune cells. And then you can experiment with those cells more quickly. And that is done with robotics already. So few real examples of that exist. I think there's a chemputer too in chemistry that can put together a small set of molecules. their first examples of this with parallel bio with like organoids and doing clinical trials of that by the way that alone that company alone has already gotten FDA approval to skip certain animal trials so you save many many millions over the next few years and lives of animals that are just bred to then be tested upon and then dissected and evaluated and this like if you love animals you can also love ai because ai is now already not just will eventually but through this one company parallel bio it's already saving animal lives uh that are just again bred for being uh like tested upon and so i think there's like tons of really uh amazing um work that is very targeted to build these out in more and more generality is kind of what is required to then allow the ai and the agent swarms to sit on all of these four pillars more efficiently yeah and To finish the tour, so there is like this agent swarm.

1:02:56So what do the agents do? Do they decide which experiment to run? Or is a human still deciding what the machine runs? And do the agents measure what's coming out? Or is it a human measuring it? How does that work? The agents will ideally work on as much of the scientific process as possible, similar to how scientific communities do it. And in many cases, the evolution of science and culture and even biology has aspects of open-endedness, which is very inspiring for us at Recursive also, and are actually basically exploring interestingly different ideas, highly in parallel, that then can then be recombined.

1:03:43combined. And so these open-ended processes have led in biology to everything from, you know, our fingers, eyes, and brains. In technology, there are lots of examples where, and Jeff Kloon, one of our co-founders at Recursive, talks about this a lot, how you can't get a microwave if you just say, make this pot faster in heating up my food, right? And you just like, you add all kinds of pressure and so on, but you had to work on radar technology and realize like some like chocolate bar in your pocket was melting as you work in radar to then eventually get to a microwave to then like like warm up your food faster and so there are like these different paths and recombinations of different research ideas that can be coming together and we can model that better and better with with agent swarms it sounds like an incredibly compute hungry and data hungry machine given the complexity of what it is that we're trying to model, especially as we think about cross-domain pollination?

1:04:44Do we have enough compute? Do we have enough data? I mean, Ilya said we've reached big data. Does the machine need to create its own data? What are the constraints? Indeed, compute is the biggest constraint. I think in the future, more and more humanity, and already we see this inside different companies, will have to decide Like what problem is worth solving? How much compute do we give to solving that problem? And then there will be new kinds of scaling laws where we give enough compute to really solve different kinds of problems. And yes, I think there's like the majority of the public Internet has been digested by a lot of these labs.

1:05:26But there's always new data. There's always new things that happen in the news. That's why, you know, U.com, we work with a lot of like Neo Labs and other labs also to just give them constantly new search results when they ask about something that just happened last week and wasn't yet part of any training data set. All right. So thanks for that. So a lot of those ideas are embedded in your new startup called Recursive. Tell us about the company. Yeah, so Recursive started with the goal of building Recursive self-improving superintelligence to automate knowledge discovery and scientific discovery.

1:06:03And it came, actually, the eight co-founders came together and we all in one form or another came to the same realization, but actually from very different directions. Tim Rock-Teschel and Jeff Klune, for instance, came very much from this open-endedness research direction of evolutionary algorithms and so on. And I came very much from this idea of, well, we automated feature engineering to have word vectors. We have neural nets. Then we automated architecture engineering by just having one unified architecture. What's the next level of automation? It's like the actual ideation and implementation validation of general ideas in all of AI research.

1:06:39And that's like clearly and obviously the next level to unlock a new set of capabilities. And when you think about the automation of science and then you apply the automation of AI research to, you know, AI like itself, before you know it, you're in this recursive self-improvement loop. And we believe that that will be a great unlock to then apply that kind of intelligence to all kinds of other scientific discoveries. And you guys raised a massive round of 650 million. And interestingly, to the compute discussion that we were just having, I read that you committed 410 million, basically most of what you've raised to a single compute deal with Amazon.

1:07:18So that goes to show the fundamental importance of compute. Yeah, we raised in the end like 670-ish. And yeah, that will probably be one of the smallest compute deals that will happen in our future. And so what can we expect from the company? What is it that you guys are going to release first? By when? There will be a couple of interesting things coming up. I can guarantee you they will happen this year. We are, and I struggle with this sometimes, the Neolab category. I don't love it because I think a lot of them will not succeed. We are a real company, not an academic lab. We are building real products.

1:07:58We're talking to real customers. And we're very excited of taking this technology and making it useful for real companies. I can't share the details yet of what we're going to release, but I think it'll be exciting. And we already know from some first conversations that it is exciting. But is that going to be horizontal or focused on a specific vertical along the lines of what we discussed? I would like there will be different, there is a sequence to it and some things will be general, but then obviously at some point it'll be more and more specific. I think we did publish a blog post that gives you a little bit of a glimpse of things we're thinking about that are essentially milestones towards full recursive self-improvement.

1:08:34that showed that, for instance, when a lot of people use AI or do some auto-research on small models, our system, the sort of first instantiation of this Eureka machine and a very narrow domain can already outperform months and sometimes years of human endeavor on particular problems. We also showed that they can build new CUDA kernels, which is very useful for faster inference, which is very useful for all large hyperscalers and people who provide tokens and run models. And we're very excited to keep pushing those. And we've heard very positive feedback from folks who are using these kernels now and have, like at NVIDIA, folks had created these benchmarks, the sole exec bench is a particular example there.

1:09:21So yeah, those are all just simple examples of artifacts that this Eureka machine can produce when it comes to AI research on the path to full RSI. And as an aside, I cannot resist asking the question why are most Neolabs not real companies? I mean, they're just like ideas of like, we want to explore, you know, this particular idea. And that particular idea is like, you know, it's one of the many useful artifacts our Eureka machine could also produce, but it's not really a product. Like, you know, if you just want to try to think about how humans interact with AI in the future, that's not quite like anything very concrete.

1:10:02And so you have to be very careful about, you know, what are founding teams and so on that have not just done amazing research, but also shipped real products. All right. To end, I want to talk about intelligence and super intelligence and where all of this is leading. So the book ends by asking a huge question. How far can intelligence go? And you propose your own definition. So maybe talk to this. Yeah, that one will take us more than the time we have left. I feel like it's almost like a new book. I had to wrap it up at that point, the book. And so one, I'm surprised no one has really defined intelligence in all of its complexity really well.

1:10:43Neither in terms of the very foundational building blocks, which I currently think are prediction, action and goals and a combination of those three. Those are sort of the three principal components. Just like sort of energy has one unit. We don't yet know what this is a unit of intelligence, something I'm thinking about a lot right now. We don't yet then have a proper definition, sort of physics inspired, you know, in physics, we have kinetic and potential energy. But then it also makes sense to study chemical energy and mechanical energy and electrical energy and different forms. And some are like still pure science fields and others are very much engineering fields.

1:11:22And so I think a similar thing has happened in AI where we have visual intelligence, language intelligence, physical intelligence and robotics. And I define these 10 different spaces of intelligence and each space basically has many different dimensions. And I'll just give you this one example on visual intelligence. We have AI can only, sorry, humans can only see in a specific part of the electromagnetic frequency spectrum. But you can go much beyond humans when you think about what are the bounds of visual intelligence? How far could an AI or any kind of intelligent life form or entity in the universe push visual intelligence?

1:12:03And then you get into very interesting sort of often physics inspired kind of like thoughts and loops. For example, you can see everything from gamma rays to gravitational waves. So very different, like the whole spectrum of electromagnetic frequencies. You can try to have not just two eyes, but you can have millions and billions of different sensors all throughout. But then how far could they go? Well, at some point you have communication bounds of the speed of light. And you have each sensor has sort of a speed of light cone around what it can see. And you quickly get into these thoughts around bounds.

1:12:44And what you then realize is that, boy, are we far away from the true upper bounds of any of the spaces of intelligence. And there's still so much further that AI can go in research. And so when people think, oh, you know, this set of algorithms or the field of AI is sort of like, it's a bubble is going to burst. I mean, maybe like energy, right? The cost, the unit cost of intelligence may fluctuate depending on a bunch of factors, but we can still go so much further as a field and as a civilization and pushing that field forward. All right, Richard, this has been another fascinating conversation and I could keep you for another couple of hours.

1:13:26But I know you have actually a couple of companies to run. So thank you for spending time with us. The book, again, is called The Eureka Machine. It comes out on September 22nd. That's right. And where else can people follow your work? On Twitter, xRichardsRosher. And Recursive.com. That's right, Recursive.com. Wonderful. Thank you so much. We appreciate it. Thanks for having me and wonderful questions. Great chatting with you always. Hi, it's Matt Turk again. Thanks for listening to this episode of the Mad Podcast. If you enjoyed it, we'd be very grateful if you would consider subscribing if you haven't already or leaving a positive review or comment on whichever platform you're watching this or listening to this episode from.

1:14:09This really helps us build a podcast and get great guests. Thanks and see you at the next episode.