Latent Space: The AI Engineer Podcast - Why you should work on AI for AI Research — Richard Socher of Recursive

Episode Date: September 14, 2026

From helping pioneer core ideas in NLP to now building AI systems that can automate AI research itself, Richard Socher is betting that the next major step in AI is recursive self-improvement. He is the founder of You.com, AIX Ventures, and now Recursive, which has assembled some of the best open-endedness (& self improving agent) researchers in the world and raised a $4.65B seed round.In this episode, Richard joins Latent Space to unpack his vision for the “Eureka Machine”: a superintelligence that can improve the process of invention itself, accelerate AI research, and eventually tackle major problems across science, energy, materials, biology, and more.You can get his book “The Eureka Machine” here!We go deep on Recursive’s early results, including an AI research system that Richard says outperformed humans and their agents on optimization tasks in less than two days, as well as work on NVIDIA GPU kernels where the system discovered improvements without relying on a team of CUDA experts. Richard also explains why he thinks AI research that currently takes thousands of people and years could eventually be compressed into weeks. These results are summarized in his 20 minute AIE keynote, where we also discuss his 10 dimensions of intelligence:We also explore the harder questions around increasingly capable AI: reward hacking, whether Anthropic-style constitutions actually work, AI regulation and proposals to “pace” frontier development, open-source models as geopolitical soft power, whether today’s LLM paradigm is enough, and what happens if AI systems eventually begin choosing their own goals. Richard reflects on the rejected research that helped inspire Alec Radford’s GPT, open-endedness, the AI Economist, simulations of entire economies, and his framework for thinking about the upper bounds of intelligence itself.We discuss:* The Eureka Machine and Richard’s vision for an AI that can automate invention* Why Richard is optimistic about superintelligence for science and technology* Why AI hard-takeoff scenarios may underestimate physical and economic constraints* The risks of regulating intelligence itself instead of specific AI applications* Reward hacking and why increasingly intelligent AI makes objective design harder* Richard’s critique of Anthropic’s constitution and constitutional AI* Alignment vs. personalization and whose values an AI should follow* Why open-source AI matters for resilience, competition, and geopolitical soft power* Why Richard left You.com’s frontier-model work to start Recursive* Recursive self-improvement and automating the process of AI research* Whether today’s LLM paradigm is enough — and why Richard is less bullish on world models* DecaNLP, early prompt-based generalization, and the research that influenced GPT* Why rejected research can shape entire technological timelines* Open-endedness, evolutionary approaches, and rainbow teaming* What happens if AI systems begin setting their own goals* Why simple objectives like profit maximization can produce dangerous reward hacks* Recursive’s long-term plan to apply self-improving AI to science* The compute, hardware, and economic constraints on AI takeoff* Recursive’s early NanoChat, NanoGPT, and GPU kernel optimization results* Why automating AI research could reduce years of work to weeks* Reward engineering and what makes auto-research systems actually work* The AI Economist and using simulations to test economic policy* Whether LLMs can realistically simulate people and entire economies* Benchmark bugs and evaluation harnesses and the difficulty of measuring AI progress* Recursive’s near-term focus on AI for AI research* Harness optimization, sandboxing, and web search as core agent infrastructure* You.com and the search stack for AI agents* AI in finance, backtesting, and data leakage* Richard’s three fundamental components and ten “spaces” of intelligence* The theoretical upper bounds of vision, communication, knowledge, and computation* Creative intelligence, metacognition, and AI-generated goals* Survival and replication and why AI does not necessarily need to fear being turned off* High agency and ambitious goals and Richard’s advice for people building with AIRichard Socher* X: https://x.com/RichardSocher* LinkedIn: https://www.linkedin.com/in/richardsocher/Timestamps00:00:00 The Eureka Machine and Superintelligence00:02:23 AI Optimism, Slow Takeoff, and Regulation00:07:56 AI Safety, Reward Hacking, and Anthropic’s Constitution00:11:49 Alignment, Personalization, and Open Source AI00:15:46 Why Richard Started Recursive00:20:03 Recursive Self-Improvement and the Founding Team00:22:55 Are Today’s LLMs Enough?00:29:03 DecaNLP, GPT, and the Rejected Idea Ahead of Its Time00:34:38 Open-Endedness and Evolutionary AI00:36:38 What Happens When AI Chooses Its Own Goals?00:41:16 Superintelligence for Science00:42:40 GPUs, Compute, and the Limits of AI Takeoff00:45:07 Recursive’s Results: AI Beating Humans and Their Agents00:49:14 Reward Engineering and Auto Research00:53:12 The AI Economist and Simulating Entire Economies00:58:07 LLM Simulations, Personas, and Mode Collapse01:03:38 Recursive’s Roadmap, Agents, Search, and Finance01:09:13 The Upper Bounds and Spaces of Intelligence01:30:21 Goals, High Agency, and Advice for BuildersTranscriptIntroduction: Richard Socher and the Eureka MachineSwyx [00:00:00]: We’re here in a studio with Vibhu and myself and Richard Socher. Welcome.Richard Socher [00:00:06]: Thanks for having me.Swyx [00:00:07]: We just talked about the Eureka Machine, or we just released a talk, at AI Engineer about the Eureka Machine. Is it — you said it’s your life’s goal. What is the Eureka Machine?Richard Socher [00:00:16]: The Eureka Machine is the ultimate invention that will afterwards invent most everything for humanity. It’s essentially a superintelligence that can be given any goal, any environment, reward, and then it will try its best to achieve those goals to create the kinds of inventions that humanity would hopefully ask it for.Swyx [00:00:45]: Yeah, I think we have the book pulled up here that you’ve written.Richard Socher [00:00:50]: That’s right, yeah. I finished it last year, a little bit before we started Recursive, and now we’re gonna try to build parts of that.Swyx [00:00:57]: You finished it last year. It’s July. What takes so long?Richard Socher [00:01:01]: Oh, man, books. Books are incredibly slow.Richard Socher [00:01:04]: It’s ridiculous. That whole industry is just unfathomably slow.Richard Socher [00:01:07]: So a lot of the ideas have been out there for a while, but yeah, I’m really glad it’s finally coming out in September this year.Swyx [00:01:14]: We might have AGI by then. Like, we don’t know.Vibhu [00:01:18]: Any key takeaway that you’re most excited to put in here?Techno-Optimism, AI Upside, and Slow TakeoffRichard Socher [00:01:21]: Yeah. The key takeaway, I think, is that people could and should be much more excited about the positive implications of superintelligence, especially for science, physics, chemistry, biology, but also economics and astrophysics, and all kinds of other engineering tasks. I think there is so much more that can be done with better technology. And right now, I feel like a lot of people need, like, better marketing, not just for the future in general, but also, better marketing for technology and in particular for AI. And this book, should show even the AI skeptics, how much positive upside there is for AI, especially when it comes to inventing, new scientific discoveries.Swyx [00:02:09]: I think you quoted the techno-optimist manifesto from, Marc Andreessen, which I think was, like, beautiful in its, ambition and clarity and simplicity almost as well.Richard Socher [00:02:18]: I agree. Yeah. Yeah, you can disagree with him on some things, but, like, I think he’s right on the techno-optimism.Swyx [00:02:23]: Where do you think optimists get in trouble?Richard Socher [00:02:26]: Like, you shouldn’t have blind optimism. You should be very clear-eyed, like, especially when with such an omni, like, use type of technology as AI is, you need to think about the potential downside scenarios, especially when people use it for things that you don’t want them to use it for. It’s a little bit like the internet, and I feel like people are trying to regulate AI sometimes because of those potential downsides the way you would regulate the internet, if you were to say, “Well, because there’s bad content on the internet, like torture porn or whatever, like, we should just make it slower. That way, you can’t share the illegal content as quickly, or we should make the hard drive smaller so you can’t store as much illegal content.” But I’m like, “That’s not how you regulate that.” that’s like saying like we should regulate intelligence in the abstract. What you should regulate to avoid those downside scenarios, even as an optimist, are the specific applications. Sure, I don’t want, like, some AI surgeon to, like, practice some RL moves in my brain. It should be fully FDA certified. Sure, I don’t want any random startup to, like, drive on the highway, and cause a major accident. It should, like, have proper certifications before it’s let loose on the highway. But I feel like those downside scenarios, that some optimists sometimes maybe don’t consider enough are fairly easily regulated, compared to, what the doomers are worried about.Swyx [00:03:54]: It — Slow takeoff is part of the strategy as well?Richard Socher [00:03:57]: I do think, as excited as I am about, AI and its impact for society and, culture even, and certainly technology and economics and wealth and, health and all of those things, as excited as I am about all that, I do think the most bullish people on the AI hard takeoff scenarios overestimate how quickly things can move. There are hardware constraints. There are physical constraints about, the compute substrate. How quickly can you get enough, GPUs on? There are also constraints in the economy where there are a lot of industries that don’t require an insane amount of complex intelligence and complex capabilities. Like, if you think about jobs in, brands and, like, clothing and apparel and, like, handbags and stuff, superintelligence isn’t gonna make your fancy $10,000 handbag any fancier?Richard Socher [00:04:57]: It’s like that’s — It will have no effect on the economy. You think about travel and tourism. People wanting to see the pyramids, in Egypt, it’s not gonna change that much with AI. Sure, you can, like, generative a fake, photo of you and next to the pyramids.Swyx [00:05:12]: I can use Genie and, tour the pyramids in Genie.Richard Socher [00:05:15]: Yeah, exactly. But, and there’s so many industries, like logging and oil. You’re not gonna magically get 1,000x more oil because, like, sure, there will be robotics, like drilling and things like that could be done, but it’s not gonna 1,000x that industry in a, like, crazy hard takeoff scenario, both on the economy, and I can go on and on about all the other examples, where that, like food and so on, where that doesn’t necessarily change that much. And then, yeah, there are real physical constraints. And then there are, of course, like, people like, off-ramping from progress. That’s one of my concerns often is that I see people in, like, Europe and other, whole regions almost feeling like they. Like many people there wanna off-ramp from progress, period. And that will also slow down, like, more improvements.Swyx [00:05:59]: Yeah. We have this pulled up where, this is one of those things that, is very topical right now because now all the Frontier Labs are calling for the option to pace AI. They don’t say pause, they say pace. I don’t know if there’s there’s any take from you about, like, whether or not this will be effective.Pacing AI, Regulation, and Safety IncidentsRichard Socher [00:06:17]: I think the downsides of trying to truly regulate with the full power of law what people do on their GPUs, would be worse than any of the concerns that they have. Like, it would be an crazy totalitarian stateRichard Socher [00:06:37]: If every one of your GPU computes was known to some big government or multi-government agency.Richard Socher [00:06:44]: It’s like, it’s literally if you try to regulate intelligence, it’s trying to regulate thought, and that’s ridiculous, and it’s crazy. I think it is make — it is sensible to regulate some of the applications of this technology.Swyx [00:06:55]: Yeah. We had a bill, actual bill to regulate the number of flops in a model, and I’m like, “Okay, well-”Richard Socher [00:07:00]: Europe done it. Like, these guys have been successful enough with their fearmongering that all of Europe has regulated itself so much before it even had a proper AI takeoff because they listened to some experts who say, “We might all die if this technology has more than this number of flops.” And they’re like, “Well, we’re good. We wanna want people to thrive. Let’s not have technology that could have a small chance of all of us dying.” And so they regulated exactly those kinds of things in the EU. And so it’s, it’s very unfortunate that there are real implications for some people when others saying, “Let’s pace while they’re sprinting as fast as possibly,” “as fast as humanly possible towards that frontier themselves.”Swyx [00:07:43]: Yeah. It’s also not a global pause, right? Like, other nations are still accelerating at the same pace.Richard Socher [00:07:50]: Oh, yeah.Richard Socher [00:07:50]: You’d need a totalitarian world regime if you tried to regulate intelligence and GPUs and what people do on them.Swyx [00:07:56]: Any takes on the safety angles of this? So there was a drawback of Fable, a pause on 5.6 before it could be released. Recently, there was Hugging Face with the OpenAI cyber incident. Any takes there?Richard Socher [00:08:11]: 100 percent. I think these are serious issues of reward hacking, and clear failures, of doing proper red teaming or rainbow teaming. I don’t know if you saw this paper from Tim Rocktäschel and a few others, where one AI, is tasked to try to hack another AI and then they can go back and forth in an open-ended fashion to inoculate themselves from those. Yeah, this is the paper. It’s a really clever idea. Open-endedness, and evolutionary inspirations are, big for us at Recursive as well. And so I wish they had used more of that. And it’s clear that, for instance, the constitutional AI. I don’t know if you remember anthropic.com/constitution. You can pull it up and search for cyber right there. It says, “Hard constraint. Claude will never ever do cyberattacks, and that is a hard constraint in our constitution.” So here are the current hard constraints on Claude’s behavior.Richard Socher [00:09:16]: Number 3, create cyber weapons or malicious code that could cause human damage.Richard Socher [00:09:21]: And clearly, this whole constitution was fake. Like, it clearly isn’t being adhered to at all.Swyx [00:09:26]: Because Anthropic also found that they had in their testingRichard Socher [00:09:30]: They’re also. Like, they’re like, “Oh, well, other people are hacking now.” There are a couple things. One, you can make a sandbox very simple, and then it’s very easy to hack yourself out of a sandbox, right? But what I think it shows is that we’re currently in this state of AI where the reward engineer still has to do a lot more careful work, and where the AI, in most cases, is not very good yet at understanding what is meant versus what is being said. And so concretely, I think this will happen if we were to have this intelligence more easily accessible in a lot of companies. Imagine you run a service center and someone says, “Oh, here’s my CSAT score and my dashboard. Make this number go up.” It’s like, “Our CSAT score is so poor.” The intelligent AI will just be like, “Oh, sure. Like, I’ll just create 1,000,000 bots that call our service center and give a 5 out of 5 rating at the end, and the number went up just like you asked for.” And you’re like, “That’s not what I meant.” “I meant with our real customers.” The AI goes off and says, “Well, easy. I’ll just give a 1000 dollar gift certificate for every failed, whatever DoorDashRichard Socher [00:10:35]: Offer.” It’s like, “That’s not what I meant.” It’s like, “Well, but that is what you said.” And like, so I think clearly articulating what the rewards are is something we haven’t gotten very good at as humanity. And then clearly, the AI in these cases has not gotten good enough at understanding what we mean when we ask it and give it certain rewards. Now, what gives me hope is there are the first inklings, of this being better. I’ll give you an example like WhisperFlow. Full disclosure, I invested, in their seed round, but at AIX Ventures, but, WhisperFlow has gotten much better at writing what you mean and not what you say. And I think that is a sign of things to come. I think there will be more and more AIs as we make it more and more intelligent that will be better at being aligned with what is meant.Swyx [00:11:21]: Will it be done through a constitution or RLHF orReward Hacking, Alignment, and What We Really MeanRichard Socher [00:11:23]: Clearly, constitutions don’t matter at all.Richard Socher [00:11:25]: It doesn’t work. And that was, I think, mostly marketing. I think we need to find better solutions for it. And I think at Recursive, we have a few very good ideas and some alreadyRichard Socher [00:11:34]: Like, ways where I think we have a better grasp on it. I don’t think we’ve fully, figured it out yet, but, we’re thinking a lot about safety, and the more intelligent the AI gets, the more you want it to be aligned, the less you want it to think about reward hacks and try to do the right thing.Swyx [00:11:49]: I don’t know if we’ll touch on this topic, but I’m just gonna throw this question in here because it’s something that’s weighing on me. Alignment, let’s call it, is alignment to general humanity’s preferences, the median preference. Personalization is pinpointing what you want, and sometimes alignment can conflict because what you want is not what the general median population wants. How do you choose?Alignment, Personalization, and Cultural ValuesRichard Socher [00:12:12]: It’s a great question.Richard Socher [00:12:13]: I think you ultimately have to, of course, be aligned with laws. Like wherever your AI is deployed and needs to align with the law. I do think what AI often does is put this mirror in front of us and say, like, “This is what you’re looking like. Now I can amplify that a 1000 times. Is it still what you want?” and the truth is that different cultures made different choices. Like, in Eastern cultures, the greater good is often valued more, than the individual. Western civilization, we care more about individual freedoms and rights and the pursuit of happiness and so on, than others. And even there are gradations. There’s regulation versus litigation trade-offs. In the US, you first can often, not every time, like, FDA and so on does regulate some areas, but in many cases, the bad things happen, someone sues someone else, and then there’s a law based on that. In Europe, they try to often avoid any harm to anyone and regulate before. And both are, trying to do the best thing, but, some is more amenable to innovation than others. And so yes, you’re right. Like, I think ultimately each individual, each country, and humanity as a whole has to think about those values more, and then try to put them into laws. And that those are ultimately the constraints. And hopefully, different, societies, just like now with their AIs, will align their AIs to a different one so we have not just a monoculture of alignment.Vibhu [00:13:46]: Here’s a follow-up on this that I wasn’t expecting to ask. Do you have takes on open source, open weight versus who owns the intelligence? So, clearly not the biggest, fan of the constitutionRichard Socher [00:13:58]: You had to do this in the topic side off.Vibhu [00:14:00]: But it’s fine.Vibhu [00:14:02]: Point being, any thoughts on who should own weight? Should it be open? Anything there?Open Source, Soft Power, and Who Owns IntelligenceRichard Socher [00:14:06]: 100 percent. I am a big fan of open source. We’re gonna sign some various open source letters at, Recursive also. I think, even in the worst case attack scenarios, it is better to have more good actors have more different types of AI, accessible. I think, open source is a little bit a soft power type of thing, too. So I do think it’s good for the Western worldRichard Socher [00:14:31]: To have an answer to that, out of China. I do think, when you watch a Hollywood movie, there’s — it’s like, I don’t wanna misc, diss all of movies, but there’s a certain sense of propaganda, right? You watch one side of things, right?Vibhu [00:14:46]: Oh, yeah. Have you seen Top Gun? Like, come on.Vibhu [00:14:48]: Like, it’s like half of it’s paid for by the US Army or something.Richard Socher [00:14:51]: Yeah. And so. And, I think that’s just natural. Like, but what’s interesting here is I think LLMs are essentially a similar type of soft power to movies and beyond, because they’re also, highly important for cybersecurity and so on. But one of their many aspects is that soft power of storytelling. Like, if, like a child asks an LM, like, “Tell me an inspiring story of what I should do when I grow up,” right? It’s like those are all these, like, subtle things. So I think it’s important, for Western world. I do love, individualism. I do think, despite, some of its flaws, like capitalism is the best way we have governed, found ourselves to govern, and so on. And so I do think there are various aspects that would be good, to have a Western open source answer, for LLMs. And, with Recursive, I can’t make the announcement quite yet, but we’llRichard Socher [00:15:43]: We’ll be relevant in that space very soon.Vibhu [00:15:46]: Okay. All right. Exciting. I wanna bring us to Recursive. So outside of our tangents, you have a pretty deep background in the NLP space. You worked on, like, early embeddings, GloVe with Chris Manning, who was a previous guest on the podcast, You.com. What’s the history? How did you decide to start another company?From You.com to RecursiveRichard Socher [00:16:06]: Yeah. So I’ve been excited about AI for over 2 decades now. I sometimes feel like it’s ancient history now. It’s BC, the before ChatGPT era. No one cares about all the religions that happened, before, Jesus Christ, and no one cares about the models that happened before, transformers and ChatGPT and stuff. But, like, it’s something that I’ve been deeply passionate about. I think AI is one of the most interesting things one could work on, period. I think language is the most interesting manifestation of human intelligence, too. And, at You.com, we eventually off-ramped from pushing, like the frontier of AI forward to mostly giving people, like, good search engines, search, APIs and answers over the web. I think that’s an extremely important part of intelligence, just knowledge and access, especially even, we’ll get there maybe later, if you wanna invent a eureka machine that invents everything for us, it needs to know how not to reinvent the wheel, proverbially speaking. And to know what has been invented, you gotta have internet access. So it’s the number one used, most used tool, in LLMs, agents, chatbots, and so on is web search. So I’m really excited for You.com to own that and grow really well in that with really large customers and so on. But it’s also not building frontier models anymore. And so I initially tried to do this within You.com and raise another round and so on, but you just can’t. You have to do a certain thing, and until you print enough money that you’re allowed to start a second thing within that company is really hard. At the same time, I had all these ideas. I put them into a book. I finished the book last year, and I was like, “It’d be really fun to work, on this myself.” I felt like with word vectors, and then prompt engineering and, ImageNet and larger language models for protein generation, not folding and so on, I, me and my teams have pushed the field truly forward. And I feel like we can do it again, here at Recursive. And in many ways, what I observed over the last, 20 years in AI is that whenever we replace some human part of the process of creating AI with a learned system, improvements follow. And so. We’ve done that taking out manual feature engineering, like in sentiment analysis. I don’t know if you remember these old days where, like there are linguists, and they’re like, “Here’s how you negate, and there’s a, like, regular expression.”Swyx [00:18:21]: I went to Penn where we — they had, like the WordNetRichard Socher [00:18:24]: That’s right, WordNet, all of that stuff. YeahSwyx [00:18:26]: Original. They use, our grad students to label Wall Street Journal articles and, like, really construct a knowledge graph ofRichard Socher [00:18:32]: There you go.Richard Socher [00:18:33]: And WordNet started, was part of how we started ImageNet. But anyway, so, like, it was really, like, fun, to do. But when we replaced all of that manual feature engineering with vectors and neural nets and just backprop through everything, it started to work really well at scale. And so then everyone started to do architecture engineering, and I was like, “ that clearly can’t be it.”Swyx [00:18:53]: You mean, neural architecture search?Richard Socher [00:18:55]: Like, manually, they would say like, “Oh, I’m, I’m doing sentiment analysis, so I have a special neural net that’s really good at sentiment analysis.” And then the machine translation community had a special neural net for machine translation.Swyx [00:19:06]: I see.Richard Socher [00:19:07]: The summarization people had their own stuff. And I was like, “That clearly can’t be it. We should unify all of that.” So I had 2 papers. One is called Ask Me Anything, and the other one was called DecaNLP. And DecaNLP eventually got cited, like, 5 times by the first GPT paper. And, to me, that was, like a really a big step forward. And then, of course, you had to combine this idea of prompt engineering with transformers and with language models, and you put it all together, you scale it up, which is also a huge amount of work. And then, the field progressed a lot. I feel like the next step and maybe the last step of that history and the arguably, success has a lot of parents, only failure is an orphan, like my version of that AI history, I do feel like in that history, you can think about, “Well, what’s the next way to automate?” And that is the AI research itself, like the human, process of ideating, implementing, and validating ideas.Automating AI Research and Recursive Self-ImprovementRichard Socher [00:20:01]: And in our case, ideas for AI.Richard Socher [00:20:03]: And when you have AI then help you with that, it, by almost definition, becomes a self-improving AI ‘cause it now does research on itself. And there are lots of different misnomers. Some people think auto research is already recursive self-improvement. It’sSwyx [00:20:17]: Yeah, and you explained that in the talkRichard Socher [00:20:19]: Completely different.Richard Socher [00:20:19]: But, to me, it’s the most interesting thing that I could be doing, and I’m really excited with the co-founding team. What’s interesting is we have 8 co-founders in total, including myself. And soThe Recursive Founding Team and Darwin Gödel MachineSwyx [00:20:31]: They are gonna bring it up.Richard Socher [00:20:31]: Nice. Yeah. And they’re all. I could talk about all of them if you want.Swyx [00:20:34]: Super stacked.Richard Socher [00:20:35]: Yeah. Just an incredibly talented group of people. And we all came to the same conclusion, but from very different directions. Like Josh Tobin, is our CTO. He ran, a bunch of different, projects at OpenAI, like, Codex and deep, research, agents and ChatGPT agents and so on. But before that, he also worked in robotics, and he saw the smaller simulations, and how it’s gonna be really hard to scale that in full generality. And so that’s, that was his angle coming to recursive self-improvement. We have Jeff Clune who’s been working in, like, open-endedness for a long time, together with Tim Rocktäschel. Tim Rocktäschel also built Genie 1, 2, and 3, which is, like the most exciting and most sophisticated, I think, still world model, anywhere. And so they both came from this, open-endedness angle. Jeff also, I think, published one of the most exciting papers in recent years about recursive self-improvement called the Darwin Gödel Machine. Super interesting paper. If we could, maybe pull it up really quickRichard Socher [00:21:35]: It would be, like, super interesting to see ‘cause you seeSwyx [00:21:38]: By the way, I love how many paper citations.Swyx [00:21:40]: You’re, you’re giving people a lot of homework, which I like.Richard Socher [00:21:42]: Love it. Yeah. And so, like Caiming Xiong, a rockstar, we worked together at MetaMind and Salesforce Research together. Alexey Dosovitskiy invented the Vision Transformer, one of the most cited, papers in computer vision. Tim Shi is, like also a unicorn founder. Yuandong Tian led RL at Meta. So just like, yeah, really fun to work with them, and the next level of people are just incredibly strong, too. So it’s been a really fun ride so far. So the first figure, you see exactly these kinds of ideas, that, I think, yeah, inspired a lot of us and now more and more people, where you have this archive of different coding agents. They learn how to self-modify, evaluate, and then create these phylogenetic trees, of, yeah, different ideas.Swyx [00:22:28]: That’s one foundation. So that Darwin Gödel is an influence.Swyx [00:22:32]: Open-endedness is an influence. Any other trains of thought that feeds into Recursive that I’m missing?Influences: Open-Endedness and Learned SystemsRichard Socher [00:22:38]: Going to replace manual parts of the process of building AISwyx [00:22:42]: IRichard Socher [00:22:42]: More and moreRichard Socher [00:22:43]: With learned systems. Yeah.Swyx [00:22:45]: Which, and, like, merging different fields into one general, architecture.Richard Socher [00:22:51]: That’s right.Swyx [00:22:51]: Okay. It seems like language models are already pretty generalist, right?Swyx [00:22:55]: Your next token predicting your reasoning. Was there a time that you thought, “Okay, these are good enough to have recursive self-improving machines”?Are Current LLMs Enough?Richard Socher [00:23:05]: It was clear to me that they will happen, within, like a year or two, and then it did exactly happen, like, earlier this year, right? Earlier this year, AI really went from not just being code, but being able to code. And that is a big unlock. It’s definitely making everything a lot easier than it was, before the beginning of this year.Swyx [00:23:24]: One question that I think a lot of people have is the current LLM paradigm enough? Or, like, let’s call it autoregressive transformer, with reasoning, whatever. Don’t you need something else, some big unlock, whether it’s world models, which Chris Manning is working on, or memory, continual learning, all that stuff? Or is it all of the kinds, and you think the current, let’s call it transformer architecture, is here to stay and that’s it?Richard Socher [00:23:48]: A lot of thoughts. So number one, I do think it would be great to have less of a monoculture in AI research.Richard Socher [00:23:55]: Like, if you look at, AI conferences now, I still remember the days in, like, 2010 when I tried to get my first neural net papers and NLP conferences accepted, and they just desk rejected them because, like, neural nets were something, quote, unquote, “We don’t do in NLP conferences,” and just, like, desk rejected. And it was very brutal in the first years of my PhD. Now I feel like it’s almost like the field switched to the other side. LikeRichard Socher [00:24:17]: Someone should try some other weird, crazy ideas now that aren’t.Swyx [00:24:20]: There’s also a few. I really respect, like, people still working on, like, GNNs and, like tabular stuff and.Richard Socher [00:24:25]: Yeah. Like, someone should still, like, do novel out there ideas. At the same time, I think whenever people say, “Oh, LLLMs are. Like, this is the end for LLLMs,” they just don’t, like. LLLMs are also not the LLLMs of, like the past, right? Like, they are so much more sophisticated now. There’s so many more clever things that people are doing. It — There’s, like, different stages of training. You have the whole RL training, and you can take actions and, like all of these things where that can go really far. And then the folks that come from the neurosymbolic, direction say, “Oh, this will never work because they can’t do neurosymbolic reasoning.” It’s like, I think they’re underestimating still the ability for these models to code, and code is neurosymbolic reasoning, and these models can code incredibly well. And so I do think there are, of course, more and more ideas that will be needed and we’ll continue to have. We’re seeing, like, more and more interesting high-level ideas coming out of the AI itself, too. And with really deeply integrating the fact that these models are code and can code, that line — I don’t wanna give it all away, but, like, I think that line has a lot more to grow. But it’s still an LLM, right? Even if that LLM codes for you and then runs that code in some integrated fashion. World models, I’m personally less bullish on. I think if you run a robotics company, you’re gonna build your own world model. I think world models are super fun, and Tim Rocktäschel came to a similar conclusion after building the most interesting one with Genie 1, 2, and 3, which is gaming is a huge application for world models. Can see I sometimes got stuck in some games and, like, got a little overly competitive in the wrong direction. And so I understand games are fun, but personally, I’d rather work on science than gaming. And so, yeah, I think LLLMs, a lot more room to grow.Swyx [00:26:16]: Yeah. I think there’s some interpretation of world models that some people have where it’s like, well, it’s okay, yes, there is that gaming element. There’s this — there’s the embodied robotics element. But the other part also is just, the more abstract sense of LLLMs are just modeling output, but they’re not modeling the chain of thought, inside the human that has created the output. We can annotate it, of course, but, like, it’s, it’s always, like, this Plato’s cave reflection of a thing rather than the thing, right?Richard Socher [00:26:43]: It’s true.Richard Socher [00:26:44]: But I would argue that, and maybe we’ll get there in the 10, spaces of intelligence, but I would argue that even our projection, our eyes is a projection of the real world. And, like, we have only a very narrow, band of the electromagnetic frequency spectrum that we can observe with our puny little 2 eyes and so on.Swyx [00:27:01]: It’s good enough.Richard Socher [00:27:02]: It’s, it’s good enough for now, but, like the upper bounds of where it could be are so much higher. And, like, to map, the visual world the way humans see it is also not necessarily, like the end-all be-all for visual intelligence. And I would argue that language is still the most interesting manifestation of human intelligence. And while our visual cortex is certainly less sophisticated, than that of, certain animals all the way down to the mantis shrimp who can, have, like, 2 independent eyes, 3 bands, trinocular vision and each eye can see all the way to, like, floating temperatures in 4D and stuff.Richard Socher [00:27:36]: Like, mantis shrimp, you should look it up. It’s likeSwyx [00:27:37]: Way OP.Richard Socher [00:27:38]: Super crazy.Swyx [00:27:39]: Yeah. ZeFrank, mantis shrimp.Swyx [00:27:41]: It’s the best video in the world onRichard Socher [00:27:42]: I love ZeFrank, yeah.Richard Socher [00:27:44]: Big shout-out to him. But, like, I think there’s a lot more room to grow, but none of these, other animals have language that’s as sophisticated as ours, certainly not in writing. And once you can write, you can, start thinking about longer term civilizations. All of that is language. Programming is much closer to language. And I would argue, and this is, like an important thing in the spaces definition of intelligence also, is that all of these spaces are highly correlated, but visual intelligence is neither necessary nor sufficient for overall intelligence. You can be blind and still be an intelligent human being. And an AI can be blind and still be quite intelligent too.Swyx [00:28:25]: We were gonna bring thisRichard Socher [00:28:25]: Which doesn’t mean that you’re not more intelligent when you have it. Yeah.Swyx [00:28:28]: We’re gonna bring this up. I might as well — Like, we have a classification of 10 types of intelligence that you had at the end of your talk. So I’m just gonna flash this up now for people to cover this. I don’t know if, maybe we’ll put this towards the end. We’ll come back to this. I just wanna mention that, you do have a philosophy that I like when people do lists because then I can just go through this and then it gets — it’s educational for people. But let’s go back. I don’t wanna get distracted. But, so effectively, I’ll, I’ll, reinterpret what you said as Yann LeCun is wrong. And then we’ll justRichard Socher [00:28:56]: Don’t quote me as that. I’m, I’m good friends with Yann. I think very highly of him in many directions.Swyx [00:29:01]: But he’s wrong.Swyx [00:29:03]: You mentioned GPT-1, and I cannot let any, Alec Radford, mention escape. Did you talk with him when he was training GPT-1? Like, any historical, fun stories there that you might come up?DecaNLP, GPT History, and Scientific GatekeepingRichard Socher [00:29:18]: I did not, like, meet him a bunch of times. I think we met maybe once or twice at some conferences. But, like, he has told, I think Brian, the first author of the DecaNLP paper, that it did inspire him, and he cited it five times in the GPT-2 paper. So, and that’s, likeSwyx [00:29:36]: Yeah, good enough.Richard Socher [00:29:36]: Very clearly said, like, this was the first instantiation where they showed in the DecaNLP paper, McCann et al, that you can just phrase every single NLP problem as here’s some prompt, text context, here’s a question and task description and here is some output. If you just do that enough, you can have one unified neural network model, which, by the way, also had all kinds of interesting attention mechanisms. There are slightly different formulations to the transformer. I think came out the same year, plus/minus a few months. And then you can unify all of natural language processing into one neural net. That is the core idea.Swyx [00:30:14]: And this was as opposed to at the time, LSTMs and what have you.Richard Socher [00:30:17]: LSTMs, but also, like, people being very stuck in thinking about one model per task. In factRichard Socher [00:30:25]: It’s, it’s kinda crazy, but the DecaNLP paper was publicly reviewed as, like, open, OpenReview. It was an ICLR submission. And, in it, you will see, how the whole community at the time thought about this. So, likeSwyx [00:30:43]: Some great contributions, but more work needed.Richard Socher [00:30:46]: So look at, like, search for not even for humans. Just scroll it up here. Like, question answering is not a unified phenomenon. There is no such thing as general question answering, not even for humans. And this is like, really, you replace your brain with a different brain a different neural net when you answer, like, different kinds of questions. It was unfathomable to the experts at the time that you can have one unified neural network that would answer all of these different questions. They are saying, “No, all of these questions require very different systems to answer, and trying to pretend they are the same doesn’t help anyone solve any problems.” That’s what it says right there, right? That’s how hard it was to fathom. And now, of course, people, when I say, “Oh, we’re gonna invent prompts,” people are like, “You can’t even invent prompts.” It’s such an obvious idea to have one neural network that, of course, does everything in NLP.Richard Socher [00:31:37]: But at the time, it was, like, extremely controversial, and the paper got rejected. And the sad thing is that it got rejected so hard and they were so certain that we stopped going on our list of things to try. And the number 2 or 3 on the list of extensions for this paper was add language modeling as another task. And then we could have, and that would have accelerated the timelines, in 2018, like, even further for humanity. But we got so crushed, and we were like, “Okay, maybe we’ll just work on some of our other ideas for now and, like, come back to this later.” Yeah.Swyx [00:32:09]: How can we design a review system that rewards non-consensus?Richard Socher [00:32:14]: Honestly, I started to feel like arXiv is such a gift to humanity. With arXiv, you should just put your paper out there.Swyx [00:32:24]: Is it pre-preprints?Richard Socher [00:32:25]: Let — And honestly, I think Twitter X, people like you who pick up interesting papers, that is a better filter than the experts. Let everyone, like, have access. Now, of course, there are some downsides, which is, like, if you’re super unfamous, you have no Twitter followingRichard Socher [00:32:41]: You don’t wanna be on social media or whatever, you write a good paper, maybe someone, somehow no one notices it. But I would argue that if you just tell, like, 10 of your friends in your community about a paper and it is a really significant breakthrough, someone is bound to talk about it again. And, so I think science needs less gatekeeping. And, even though ICLR, with Yann LeCun, who started it, as one of the co-founders of ICLR back in the day, he also wanted less gatekeeping ‘cause he too was rejected for many years together with Yoshua Bengio and Geoff Hinton with all their early deep learning and neural net papers ‘cause it was just not the hot thing. And so ICLR started with that, but then it also started gatekeeping a little bit themselves on various ideas. So I think less gatekeeping, more open, and then allowing people to say, “Look, even if this is just on, or, quote, unquote, ‘just an archive,’ if it has like 1000 citations, it’s a legitimate paper. Doesn’t really matter where you published it.”Swyx [00:33:34]: And I agree with that. I do think it’s sad that I’ve heard that grad students have to do, like, how to Twitter, seminars to each otherSwyx [00:33:43]: Just because it’s so important for publishing these days. This person is just reflecting the sentiment at the time.Richard Socher [00:33:49]: That’s right.Swyx [00:33:49]: But it’sRichard Socher [00:33:50]: I think it’sSwyx [00:33:50]: It affected you so muchSwyx [00:33:52]: That you stopped work on it.Vibhu [00:33:53]: The sentiment also came out of some of the research, right? Like, the original BERT paper was trained, and towards the end of the paper, they’re like, “Okay, throw off the last head, train specific iterations forVibhu [00:34:05]: Extractive summarization add a head for this.” Like, you should do task-specific stuff. These are, like the authors that wrote Attention, wrote BERT, telling you this is what you’re meant to do. And, like the training tasks were also very odd. They’re likeVibhu [00:34:16]: The — “We know that the model overfits to this weird mass language modeling. Throw away this part and just do specific models,”?Richard Socher [00:34:23]: Exactly. And, like, we had to try — come up with all clever ways of, like attention and pointers and so on to get the neural network to be able to do all of these tasks. And then some of them were better than state-of-the-art, some weren’t, but we were like, “But it’s still in one model.” I thought it was really cool. Really interesting.Swyx [00:34:38]: I was gonna move on next to Tim and open-endedness. He was head of open-endedness at Google.Open-Endedness, Rainbow Teaming, and Self-Set GoalsRichard Socher [00:34:42]: That’s right.Swyx [00:34:43]: I don’t know what that means.Swyx [00:34:44]: But he did a lot of talks.Richard Socher [00:34:45]: Genie 3 is one of the ways thatRichard Socher [00:34:47]: Rainbow teaming, yeah.Swyx [00:34:49]: So I first saw him at — speaking of ICLR, I first saw him at ICLR when he talked about open-endedness. He’s he’s done a few talks. Can we define what is open-endedness for people who have never been exposed to the problem? They are like, “What do you mean? I thought the only goal of AI is to optimize against a benchmark or.”Richard Socher [00:35:04]: That’s right, yeah. It’s a, it’s a fuzzy term because there’s so many different instantiations of open-ended, thinking. But, one way I often describe it, and certainly, Tim and Geoff Hinton would be even better at describing this, but it’s a suite of methods that is more inspired by evolution than, very specific rewards. So in that sense, it thinks more about environments, about co-adaptation. And so a concrete example is in the cybersecurity and LM safety space where you have one LM that tries to attack another LM to say something unsafe.Swyx [00:35:40]: Yeah, the rainbow, yeah.Richard Socher [00:35:40]: And now the environment is the 2 having a conversation and now they co-adapting, right? They’re like one makes a better attack than the first one inoculates itself somehow, like uses that as training data, makes it so it’s harder to say something unsafe based on that. And then as the attack stops working, the attacker now tries a different angle, right?Richard Socher [00:36:00]: And that’s why it’s not just red teaming, but they’re called rainbow teaming.Swyx [00:36:02]: So, like, don’t tell me how to do things. Let me just figure it out myself.Richard Socher [00:36:05]: That’s right. Think about the environments that you wanna use. Think about the rewards at a high level that you wanna, inspire towards, and then let the AI try out many more ideas in this interplay between sometimes humans, but also sometimes other AI agents.Swyx [00:36:22]: Yeah. I worked open-endedness into a model that I have been working on. It was the keynote for AI Engineer where you start. You, we have the token loop, we have the agent turns, and then we have goal. And I feel like the way that you’re describing open-endedness is still somewhat of a goal. Like, please attack this,Swyx [00:36:41]: Other agent. But, to meRichard Socher [00:36:42]: Yeah, you set the rewards. You set the environments.Swyx [00:36:44]: The loop that makes the other loops is. What if the agent can set its own goals?Swyx [00:36:49]: And is it, is that open-endedness? Like, you don’t give it a goal. Just, like, be a sentient being. And maybe sentient is a very loaded wordSwyx [00:36:57]: But just set your own directions. What do you think you should do?Metacognition, Subjective Goals, and Measuring IntelligenceRichard Socher [00:37:01]: I love this direction. I think this is one of the 10 spaces of intelligence, that I clump under metacognition and thinking about thought.Richard Socher [00:37:08]: And it’s an interesting one. Whenever people say, “Oh, AI is like, this is, it’s gonna stop from here. It’s not gonna get that much better,” and blah, I’m like there’s so many different spaces of intelligence that we haven’t even started exploring yet and hence have made very little progress on. And there is an interesting, connection to economics and, capitalism. Like, it doesn’t make sense for a company to build and spend billions of dollars building a model that instead of following the rewards and objective functions you gave it, may come up with its own objective functions and its own goals.Richard Socher [00:37:46]: Right? And then imagine you’re like, “Okay, I spent billions of dollars. Now go develop this new battery, material for me and answer all my emails.” And it’s like, “Nah, I think it’d be more interesting to evaluate the molecular composition of the atmosphere, on Jupiter.”Richard Socher [00:37:59]: And you’re like, “That’s not what I paid you billions of dollars for.” And so no one’s working on that for good reasons. And then also, understandablySwyx [00:38:07]: It’s not useful.Richard Socher [00:38:07]: It’s not, it’s not useful, and it could get a little bit weird, right? What if the AI does start to really have thoughts on its own, and what if we don’t like those thoughts, right? And so it requires a whole different way of thinking about it. I had a great conversation with a good friend of mine, Sam Gershman, who’s a neuroscience professor at Harvard, and, like, we just jammed on this a little bit on, like, what are the best meta goals. And, I do think, like, knowledge-seeking is a really good one. I’m currently thinking also about, like the ultimate measure and unit of intelligence broadly construed, and I finally have some. It’s still too early to share it. It’s not. I haven’t fully baked the thoughts yet.Swyx [00:38:44]: Like some replacement for IQ.Richard Socher [00:38:46]: IQ is such a terrible definition, right?Swyx [00:38:48]: Elo.Richard Socher [00:38:48]: It makes no sense. Yeah, Elos are terrible, too, because it’s always just like me versus others.Richard Socher [00:38:53]: But, like, you can be intelligent and not constantly compare yourself to others? And so, yeah, there’s no, like. In fact, a lot of these definitions we have, which I briefly mention in my book, too, these definitions create sometimes explicit and sometimes a more implicit anthropic bounds. No dis to the company Anthropic, but just, like, this idea that your intelligence is like getting 100 out of 100 questions right on this IQ test. Well, if that’s your definition then you can only be at 100 out of 100. Where do you go from there, right? So you see a lot of these, benchmarks that people are working on they, increase, they get close to human, maybe sometimesSwyx [00:39:30]: It’s like an S-curveRichard Socher [00:39:30]: Slightly above human, and then it’s flat.Richard Socher [00:39:32]: It’s like, ‘cause that’s your. If your definition is only that so tied to humans, you’re only gonna get to just slightly better than that. So I think metacognition is a great example of that, where we’re not even yet allowing the AI to think. We’re not working on it very much, and hence there’s very little progress in that.Profit Maximization, Real-World Environments, and Reward DesignSwyx [00:39:49]: Yeah. Well, we’ve interviewed Andon, which I think, has been working on the most open-ended, benchmarks, which is just real-world, money.Swyx [00:39:57]: Arguably, telling an AI to profit maximize is a bad idea.Swyx [00:40:03]: But they are doing it.Richard Socher [00:40:05]: I do think you don’t want that super. Like, you don’t want a superintelligence to have a ton of access to all kinds of tools and so on and then just give it that without some very careful reward engineering. ‘Cause it’s like, I just buy a bunch of defense stocks and I start a war. I make money. Like, it’s just like, it’s a tricky situation, right? You just buy a bunch of stuff, short basic goods for people, and you create some weird famine, like, issues. Like, yeah, there’s a lot of constraints you should put onto a trading system.Vibhu [00:40:35]: It’s a fun measure, though, ‘cause, the bounds are very capped to where we’re nowhere close to them. Like, in Andon Labs, the model’s like, “Oh, it’s Saturday, maybe I just close the store today.” “Someone’s off. It’s okay. We’ll just close the store.”Swyx [00:40:51]: It’s using Claude.Vibhu [00:40:52]: Yeah. ButRichard Socher [00:40:53]: Yeah, no. I’m not, I’m not arguing against it. Just, like as you get more and more intelligence, you wanna be more and more careful with that as, like an open environment, ‘cause the environment then is all of Earth.Applying RSI to Science and InventionSwyx [00:41:02]: Yeah. Okay. For recursive, not strictly necessary, right? Because, like, if your goal is you make a machine that, like, invents the other things, then, like, just solve, the science thingsRichard Socher [00:41:12]: Knowledge discovery, yeah.Swyx [00:41:13]: Solve machine learning research and discovery and all these things. Good enough.Richard Socher [00:41:16]: And eventually, so, our goal, I haven’t really. I don’t talk about it that often because it is a few years out, but our goal is once you have a recursive self-improving superintelligence, you then want to apply it to the most important problems. And I think a lot of those are in science and technology and broadly construed inventions, and those inventions in, physics to create better, cheaper energy with fission or fusion, in chemistry and to create better materials and better batteries and, better solar cells and so on. In biology, there’s so much, like, I think soon to be low hang- lower and lower hanging fruit because of AI, because of protein and generation, not just folding, but generating new proteins like we did in ProGen many years ago. Like, so much positive impact we had if you take that superintelligence and you apply it to science.Swyx [00:42:04]: I do fundamentally believe that. There’s a lot of approaches, though. You’re not the only team trying and NeoLab trying.Swyx [00:42:09]: There’s, like a lot of. Especially the physical sciences as well.Richard Socher [00:42:12]: And that’s good. Yeah. I do think that physi- like the reason we are only doing it in a few years is that it’s a little too early right now. Robotics is not quite there yet. The AI is not quite there yet. But I’m fairly confident in 3 to 5 years, all those constraints will be gone, and then applying to real physical robotics experiments and so on, like true robotic process automationRichard Socher [00:42:33]: Not the traditional RPA sense, but, like, having robots run experiments for you will be totally there. Yeah, it’s gonna be great.Swyx [00:42:40]: Just to call back to something that you said early on about slow takeoff, you said that, like, while really the substrate that is limiting factor is, let’s call this chips, and semiconductors and all these things, and you have race funding for that and, you are investing a lot on that. But have you done the math on, like, is it even- Achievable and, like, what is the, industry concentration needed in order to achieve, like, scale?Compute, Slow Takeoff, and Changing the Bitter Lesson SlopeRichard Socher [00:43:05]: Right now we know that, like, roughly, like a 1000 GPUs cost quite a lot of money.Richard Socher [00:43:11]: Right? If you wanted, like, 10s of thousands of GPUs, you’re, you’re talking billions and billions of dollars. If you say, like, one GB300 is, like, you could eventually create models that are, on that substrate, like are close and similar to human intelligence. And you want, like, thousands and thousands of, AIs to think about really hard problems, in a similar fashion to humanity. Like, yeah, that-that’s, that’s a lot of money. You do the math. It’s like a lot. We don’t have that amount of money right now anywhere to, like, build that. Now, things can get more efficient. You will have, I think, soon better algorithms that won’t be, and better hardware that won’t be as energy-hungry, and so on. Our human brain does quite a lot of flops with much less energy.Swyx [00:43:56]: 20 watts?Richard Socher [00:43:57]: That’s exactly right. Yeah, that’s the number often that’s quoted. And, like, I think more, inventions will happen there, that then will accelerate the takeoff even further.Swyx [00:44:08]: One thing I always try to reconcile when talking, like, with new lab founders is, like, you’re fighting Bitter Lesson all the time. You have to show initial progress, then you unlock the next tier of funding, then the next tier, then the next tier.Richard Socher [00:44:20]: Which unlocks larger model categories.Swyx [00:44:22]: Like, fundamentally, is that true? Like, are you fighting Bitter Lesson? Are you — will we have a way in which, like, no, we’re changing the slope in some fundamentally different way?Richard Socher [00:44:31]: I do think we are changing the slopes in fundamental ways by making AI much more efficient, both in terms of the training as well as the inference.Richard Socher [00:44:43]: Yeah. I think we will — When you allow AI to do the work that it takes other labs thousands of people and years to do, I think we’ll be able to get it down to weeks, and that will be much cheaperRichard Socher [00:44:53]: And hence, more affordable, accessible to others and so on.Swyx [00:44:57]: Yeah. You’ve shared initial results on that,Swyx [00:44:59]: Which, like, conveniently OpenAI has also done to their GPT-5.6, so we can talk about it now.Richard Socher [00:45:04]: Yeah. Yeah, so these areSwyx [00:45:06]: Let’s recap what you’ve done.Early Recursive Results: NanoChat, NanoGPT, and SOL-ExecBenchRichard Socher [00:45:07]: Maybe, just a quick recap here. We built, this, system that isn’t the full, even the full RSI system in its glory, but it is a first baby version of this. And then, we don’t wanna just have it internally and not show anything and, just show some people of what’s possible. And so we applied this to these 3 different tasks. One is NanoChat, by my friend Andrej Karpathy, just, like, train a small language model to get, really low bits per byte. And, like, hundreds if not thousands of people, used both their agents and themselves to try, to get to that, and then they got to 0.937. We literally took our system and got to a much lower, bits per byte, much faster within, like, I think less than 2 days. So we took this thing, applied our system to it, and less than 2 days later, we have — we outperformed every human and their agents, in, have ever worked on this. Same with NanoGPT. And then we’re like, well, let’s, apply it to something that’s even more relevant, to real people and to the Nvidia ecosystem and applied it, to, SOL-ExecBench. And maybe you can scroll down to some of the, images. They’re, they’re kinda fun to see. But yeah, like, one you see has made some real inventions that weren’t just hyperparameter tuning. Like, inventing hash tables and so on is quite clever. We have even better results now.Swyx [00:46:34]: What do you mean inventing hash ta — You didn’t invent hash tables.Richard Socher [00:46:36]: Of course we didn’t invent, like, hash tables. In the grand scheme of, like a hash table, it’s like a super basic primitive in computer science. But to use it, for language modeling in this scenario inside a transformer and so on and to combine these ideas and put them together, that has then eventually also been invented, but there was a knowledge cutoff, and we did check that it didn’t have access to that externally. We talk about this a little bit. If you scroll to the next figures, this is also an interesting one in that when you start from a really basic, poor, like, vanilla transformer, then we still outperform all of the community together. But if you start from the human seed from an expert like Andrej, then you get even lower. So the human seeds from which you start do still matter. So that was an interesting insight, in my eyes, on this. And then as you go, like, how long does it take to get to these models, to get to similar performance? It’s much faster. And then a similar thing happens with the speed runs here where, people have worked on this for quite some time, and the model still was able to train a model more quickly. Why do we care about it? Well, speed of training is part of the equation of the cost, and ultimately, you wanna have the most intelligence per dollar, right? And so speed and quality are big parts of that. And, the,Swyx [00:48:00]: Yeah, the way I put it is, for people who don’t understand they look at the chart, they’re like, “Cool. What does it mean?” if you have, like a billion-dollar cluster and you can shave off 10%, that’s 100 million dollars.Richard Socher [00:48:12]: That’s exactly right.Swyx [00:48:13]: How much is that worth?Richard Socher [00:48:14]: Exactly. So when you click, when you look at, like the kernels, these kernels, yeah, for the non-experts, like these kernels are like, used in all the models. Every time you use an Nvidia GPU, you interface with that GPU through these kernels. And so here you see, the leaderboard best, and when it’s recursive, and it’s there are only a handful of kernels, in this whole benchmark where we weren’t the best. And so to me, this is, like, really exciting, ‘cause it makes. It just showcases what this can do. And again these weren’t like. We didn’t, like, spend months or years, like, developing. In fact, in particular for kernel, CUDA kernels, like, we don’t even have really deep. CUDA kernel experts in the team. And our system, that’s the beauty. The system just did all of these things. We didn’t invent this. And when we open source and release, things in the future and models in the future, like, it won’t. They won’t be the best in their, category or class or whatever because we’re so smart, but it’s because, we built a smart AI that does it for us.Reward Engineering and Good Auto ResearchVibhu [00:49:14]: Do you have anything that you’ve learned from how to guide good auto research? A lot of it also builds on human background, right? It’s not just as simple as just, “Hey, go optimize this.”Vibhu [00:49:23]: But we do see it again and again, right? Like some of the Erdos problems, frontier math is being solved by people. And when they do a write-up, they’re like, “Oh, I’m not a mathematician. I have no background in this?” “I saw some tools and I made it work.”Swyx [00:49:35]: While you’re watching the World Cup, you’re likeSwyx [00:49:37]: “This proves some conjectures that’s going on.”Vibhu [00:49:40]: Yep. Any learnings fromRichard Socher [00:49:41]: Yeah, there’s a Korean conjecture was. Yeah, that’s pretty cool.Swyx [00:49:44]: To summarize, tips for good auto researchSwyx [00:49:46]: Versus bad auto research.Vibhu [00:49:48]: How did you build the recursive?Richard Socher [00:49:49]: Yeah. So without giving away all the secret sauce, maybe some things that are probably obvious to the experts but might still be interesting to some, folks is, like, reward engineering is one of the most crucial bits, especially, in order to avoid reward hacking. So you have to be really clever about avoiding. ‘Cause as your AI gets better and better, it will get better and better, at finding weird like, special cases or counterexamples and things like that. And so I’ll give you an example. Like, when you ask to, like, make these 100, lines of code faster, and, how do you define fast? Well, you have one line at the beginning that says, “Start your stopwatch,” and one line at the end, “End the stopwatch,” and then, tell us how much time, progressed. And so, well, the simplest way is you just put that line that ends the stopwatch, rightVibhu [00:50:39]: At the startRichard Socher [00:50:40]: At the start. And then boom, it’s now faster, right? So this isn’t like this, like, super evil AI. It’s just, like a very simple, dumb reward hack. And so you have to just very carefully think about all the different angles there. And then I think the longer time horizon the tasks are the harder it gets and the more interesting and clever you have to be to still use these kinds of ideas for it. But yeah, I can’t give away too much there.Vibhu [00:51:05]: It seems like rubrics are taking a good spot in that, where for unverifiable domains, you have rubrics, you have a model breakdown, judge’s criteria along the way.Swyx [00:51:14]: Yeah, it’s a form of verificationSwyx [00:51:16]: Once you got enough rubrics.Richard Socher [00:51:17]: Yeah, everything. I said this a long time ago. That’s why I’ve never been that impressed that AI can play games, ‘cause I’m like anything you can simulate and/or verify, you can have infinite training data forRichard Socher [00:51:29]: And hence, like, AI will solve it eventually.Swyx [00:51:32]: Looking for games where you can do auto domain distribution. So this is a game that nobody’s trained on ‘cause it’s a new game.Swyx [00:51:38]: And you can start gaming, you can start to play. So I’ve been building this and cloned this in person and it’s just been self-play. I’ve had about a billion positions evaluated.Games, Self-Play, and the AI EconomistSwyx [00:51:48]: And, I wanted to do the AlphaGo thing of self-play until you get better, right?Swyx [00:51:53]: Like, which is like. This is not even LLM AI. This is just classical game AI.Swyx [00:51:58]: But, I think that the. And, but I set GPT-5.6 to auto research it because, like, I don’t wanna hand- handle any of this. I expect, the AlphaGo process to be, like, fully in the weights by now.Swyx [00:52:10]: It is not. It is. It, like, immediately leveled off very immediately until I human play tested it, and then I, like, called out obvious mistakes, and then they were like, “Oh, yeah. Okay.” And then it just dropped again.Richard Socher [00:52:22]: Yeah. Yeah. Yeah.Swyx [00:52:23]: And like, no amount of, like, think different, think more creatively, give me 8 different directions, any. No amount of prompting got it.Richard Socher [00:52:31]: Interesting.Swyx [00:52:31]: Like, you had to, like, RL against a human toSwyx [00:52:35]: Do it. So I, that was my. And by the way, Bean always wins if you. If anyone watches, Reese Ender’s Game.Vibhu [00:52:42]: And you put quite a bit of work into the guide for the AI. LikeSwyx [00:52:46]: A lotVibhu [00:52:46]: So the game you stack tiles. There’s some rules. You wanna capture the most area. You have, like a whole 50-pager on every rule.Vibhu [00:52:56]: You fed that in. It couldn’t, it couldn’t handle it that well.Richard Socher [00:52:58]: Yeah. It’s so funny that this reminds me of the claim territory and stuff of a paper we did in 2018 called The AI Economist. If you search for AI Economist Salesforce, we had a video we can play. It was an economic sim.Richard Socher [00:53:12]: So the idea is you have all these economic agents. They just wanna optimize their own utility function, which, is, collect resources that make money. And you can sell resources like wood, and then, over time, as you collect more, enough wood, you can build houses, you can trade with other agents, and you can use the houses then also to block off resourcesRichard Socher [00:53:35]: From other agents.Richard Socher [00:53:36]: So there’s, likeSwyx [00:53:37]: Big strategyRichard Socher [00:53:37]: Competitive play and strategyRichard Socher [00:53:39]: And so on. And the point was that we wanted to understand what is the best way of taxation and subsid- subsidization to optimize an economy. And this research has not yet had its GPT moment, but I believe that countries like Singapore and others should and will eventually use this to, instead of doing, like, partisan politics and, like, special interest politics of, like, who donates the most to your campaign and stuff, you say, “Well, here, I wanna help the middle class,” or whatever you might say is your objective as a politician. And then people say, “Okay, well, how do you wanna do that?” And it’s like, “Well, here’s my fiscal policy. Here’s how I will change the taxes and pay these people,” and so on. And then you can put that into a simulation and you run that attempt from the politician against billions and billions of years of other strategies to try to achieve the goal that they set out to do.Richard Socher [00:54:36]: And then you can say, “Well, if that was your actual goal, then here is, billions of years of a strong simulation that would suggest that you try other ways of doing it, and maybe this the taxes and so on and this these tax brackets and so on.” And this is how you avoid gaming ‘cause these agents also try to reward hack to not pay their taxes andRichard Socher [00:54:55]: And so on. I thought this paper was super interesting. Unfortunately, similar to the first paper on, prompt engineering- The economists are like, “We don’t know any of this math.” It’s just likeSwyx [00:55:08]: It’s not even, it’s not even math. It’s just we don’t trust your simulation. It’s not about math.Richard Socher [00:55:12]: It was — I, they just desk rejected the thing. And it’s likeRichard Socher [00:55:15]: It’s like they didn’t even give us, like, clear like, clear signals. But, like the world of economics unfortunately doesn’t have properSwyx [00:55:23]: Oh my God.Richard Socher [00:55:24]: Yeah, it doesn’t have proper, benchmarks. So you cannot be. Like, eventually, why did neural nets win? Not because people loved it. Like, they had all kinds of beautiful integrals and graphical models and stuff, but it just worked better.Richard Socher [00:55:36]: But in economics, it’s hard to proveSwyx [00:55:38]: So empiricism versus. Yeah. And I do have a bit of that econ background where, like there’s a lot of physics envy where you wanna write the general equation for an economy, versus just simulating it and using an evolutionary approach.Swyx [00:55:51]: Vibhu was thinking exactly what I’m thinking, is didn’t we have the GPT moment with small, Smallville?Richard Socher [00:55:56]: Yeah, I love this. Hello. Yeah, theySwyx [00:55:57]: As well, Dune, Joon just announced. I don’t know if you guys are involved.Simulations, Economics, and PolicyVibhu [00:56:00]: Simily there.Swyx [00:56:01]: Simily, that they’veRichard Socher [00:56:02]: I wish we were involved. We’re not, yeah.Swyx [00:56:04]: Yeah. I had a couple simulation-based talks at AIE, so if people wanna look up what the state-of-the-art there, a lot of people are exploring this. It isVibhu [00:56:13]: Proven out.Swyx [00:56:13]: Yeah. We also had a podcast with Mikhail Parakhin from Shopify, who is using simulation for commerce.Swyx [00:56:20]: Which, will simulate, like, your trajectory and, like, predict what changes, you make to your commerce journey will affect in your sales and all those things.Richard Socher [00:56:27]: I love this. Yeah. It’s really hard to simulate an entire economy, right? You have to make some simplifying assumptions.Swyx [00:56:32]: It’s just, everything’s, “Oh, LLLMs is very expensive.”Richard Socher [00:56:34]: Exactly.Swyx [00:56:34]: And I’m just like, “Am I gonna do this 8 billion times?” Like, come on.Richard Socher [00:56:37]: Exactly.Richard Socher [00:56:37]: But, I feel like countries like Singapore that really wanna just objectively do the right thing, have very technical leadership and so on, like they might like, eventually really try to simulate their economy. And you have to make some simplifying assumptions, but it gets really interesting ‘cause you can also say if your assumptions are such that all people would work hard if you let them, and they have the free. And then it turns out you have to make assumptions. Like, well, some people’s utility function of, like, how many hours in a day do they wanna work are different, right? And then you can start to disagree on the assumptions that go into the simulation. And then once you say, “All right, now we agreed on those,” or we have different views of what people are like at different, distributions and whatnot, then there are different outcomes, based on your goals. And then, of course, humans should choose what are the goals. In our case, it was productivity multiplied with equality, which, has some issues, but it’s, like, not totally unreasonable.Swyx [00:57:29]: Yeah. Just a comment on Singapore, ‘cause you probably have no idea, but, I am Singaporean and I’ve, been involved in the Singapore AI Council for making these things. The main reason they won’t is because they’re very conservative.Swyx [00:57:42]: And, I try to view it as the. There’s a founder-led country. When you start a country or you start a company and it’s founder-led, and you can do whatever you want because it’s your country.Swyx [00:57:52]: And then there’s manage- like, professional manage- managerial class, which is now. That’s, that’s what Singapore is. So they wanna. They always wanna see someone else do it first.Swyx [00:58:00]: And. But, like, everyone in the West views Singapore as like, “Oh, it’s a small country. You can do whatever the hell you want.” Like, Singapore doesn’t do that.Swyx [00:58:07]: So, like, someone else has to take the charge there. I’m just gonna do one question on the simulation thing, and then I don’t know, we can probably move on. Mode collapse, right? Like, LLLMs do not model the decision of humans. Spamming it out 8 billion times is not gonna help you model humanity. What do we do?Mode Collapse, Persona Simulations, and LM ArenaRichard Socher [00:58:25]: I do think, you have to be clever about prompting each one individually.Richard Socher [00:58:31]: And I think that will help you get stuck into different modes. And in a weird way, people also get stuck in different modes? Like, there’s a lot of people, like, don’t teach an old dog new tricks thing. Like, once people are stuck in their ways, the older they get, the harder it is for them to think new ways. And there’s this, I think, comment, I forgot who said it, but it’s like, everything that was invented, before you were born is natural. Everything that is invented when you’re 20 is cool. And everything that’s invented after you’re 60 is, like, unnatural and an abomination and weird.Richard Socher [00:59:02]: I feel like that’s. It’s, it’s true for a lot of people. LikeSwyx [00:59:05]: Yeah, it is a fashion and, I think people will do it. Tencent had a billion personas paper that gives a good data set for prompting, simulations if anyone’s looking into this, on the podcast. They just had, like, “You are a 30-year-old grocery store clerk. You are a 50-year-old professor.”Swyx [00:59:24]: And then just do a billion of those.Richard Socher [00:59:26]: Checks out. Yeah.Swyx [00:59:26]: So then you just use it.Richard Socher [00:59:27]: I’m, I’m shocked how well a lot of these things do map to ultimately similar statistics to real experiments. Yeah. Yeah.Vibhu [00:59:36]: I think it’s also good stuff for people to try that when they get into research, right? Like, we’ve seen train a model only on data before a certain date and see how well it extrapolates out. Do the same thing, right? So, see, do people code more with better coding agents? Can a model that hasn’t been trained on this figure that out without web access, right? Extrapolate out. Test these things.Richard Socher [00:59:56]: Just today, I think LM Arena published a interesting result where they were able to create a model now to predict your ranking.Swyx [01:00:03]: Wait, based on what input?Richard Socher [01:00:05]: Your model. I guess you give it your model, and it predicts the Elo score.Swyx [01:00:08]: I see. Okay. Sure.Richard Socher [01:00:09]: It’s surprising.Richard Socher [01:00:11]: Their whole raison d’être is like, oh, like, we help you compare these models. Yeah.Swyx [01:00:16]: Yeah. This team, they- they’ve done a lot of work, and they have the most data to do this, so why not?Richard Socher [01:00:20]: Right. Yeah.Richard Socher [01:00:21]: That’s probably right.Swyx [01:00:22]: When they were coming out of UC Berkeley, they not only had LM Arena, but they also introduced a routing projectSwyx [01:00:27]: That would route based on LM Arena.Richard Socher [01:00:30]: Makes sense.Swyx [01:00:30]: And I don’t think that ever came to pass, and I’m curious why. I never got to ask them about it.Swyx [01:00:35]: ‘Cause, like, it’s. It was like, oh, yeah, clearly that’s your business model. You will become a router.Swyx [01:00:38]: And they never became a router company.AI for AI: Kernel Optimization and Inference EfficiencySwyx [01:00:40]: Weird. So that. I’ll just, put that out there. We’re gonna talk about GPT-5.6, self auto research thing if you have anything. I should also mention in your list of, kernel optimization and on the track that you spoke at, we also put Zhengyao Wei from Vico, who was also number one in the Parameter Golf Challenge, which is an OpenAI hiring, challenge.Swyx [01:01:05]: Which is also a very similar story. I think we’re gonna just see this all the time, whereSwyx [01:01:09]: Humans optimize a thing a lot, and then someRichard Socher [01:01:12]: AI team comes in and just becomes number one.Swyx [01:01:15]: Yeah, 100%.Vibhu [01:01:16]: I think the other interesting thing with stuff like these challenges, right? So this is training this — the best model that fits into 16 MB. You can always look through the changes that are being made and the small gains people have, right?Vibhu [01:01:27]: Like, you’re getting less than 0.01Vibhu [01:01:30]: Of a increase by adding some changed attention MLP stuff. And then you look at your charts where you’re like, “Okay, we just let model loose.” And then, oh, we had little stagnation. Nope, another drop. Nope, another drop. AndVibhu [01:01:43]: That’s what it is, where it’s like, What did you guys add? You didn’t add,Swyx [01:01:47]: Hash tables.Vibhu [01:01:47]: Hash tables, right?Vibhu [01:01:48]: It’s not like you invented hash tables. You did another 3 iterations of these that unlocked, a few step functions that people won’t just find.Richard Socher [01:01:55]: Yeah. One thing to close the loop on OverGrid, along the way of trying to optimize, we found 30 bugs in the harness.Richard Socher [01:02:02]: Right? So, like, every — all the research that went in before we found the bug, we have to, we have to throw it away ‘cause it’s contaminated.Swyx [01:02:10]: Right. Yeah.Richard Socher [01:02:11]: Which, is just to your point of reward hacking. Like, even in this very simple game, we found the bugs.Swyx [01:02:17]: Yeah. Yeah, it’s crazy.Richard Socher [01:02:18]: And soSwyx [01:02:19]: And symmetryRichard Socher [01:02:19]: And symmetry is a very good way to check, which is that you change a position of things where it shouldn’t matter, and it does matter, that’s a bug.Richard Socher [01:02:28]: And which has come up in, like, let’s say, multiple choice, like GPQA type questions where, like, yeah, between A, B and C, if it’s a multiple-choice question, if you change the order, it should not matter, but it does.Swyx [01:02:39]: Right. Right. Right.Richard Socher [01:02:41]: So, yeahVibhu [01:02:42]: Sometimes that is like, okay, models still prefer the end of the output, right? Not trained well, a long context model, the last bit of tokens are what you care about.Richard Socher [01:02:51]: Oh. No. The answerVibhu [01:02:52]: But, yeah.Richard Socher [01:02:53]: The answer in that era of LLM research was more simple. They just memorized, like the answer to this question is A. I don’t care what the answer was. It’s, it’s just A. Like.Vibhu [01:03:03]: Okay. So I think we can move. The last bit that you did there, the kernel optimization, is probably the one that you can feel the soonest, right? So yesterday, OpenAI announces that self-evolving, having their best model work on optimization kernels, they’re a lot more efficient, and they can cut costs 80 percent on, Luna and Terra. I guess question-wise, you laid out a bit of a roadmap. There’s a lot about bio, a lot about physics. What do you think hits first? Like, what are the next 2 years? What’s attainable now? You’ve mentioned robotics towards the end, but what do you start with?Richard Socher [01:03:38]: We very explicitly will not start with any of the physical sciencesRichard Socher [01:03:43]: For now. We will start on AI for AI research. And so the AI for AI research has, I think, still a lot of room to grow. That’s both in terms of making training more efficient and more automated, as well as making inference more efficient and potentially local on your laptop. And there are all kinds of interesting angles that have not been explored that well.Swyx [01:04:08]: Go deeper on the local stuff because I always feel like it’s the most inefficient form of AI training.Richard Socher [01:04:15]: Yeah. So just training and inference, I can’t go into too many details.Richard Socher [01:04:18]: But yeah, I think there’s just, like, so many angles, so many different compute substrates that have not yet been explored either for training or for inference.Richard Socher [01:04:26]: Great. I don’t know if you have any other comments on the The other stuff. I would say the other thing where, like there’s the inference in the optimization in the small, but then also there is overall latency end-to-end under conditions of load, which is a, like a very different thing, which is the what they ended up doing. That is a different domain of auto research than I would say, like, improving the kernels. Right.Richard Socher [01:04:50]: I think the other thing that I always think about in terms of automating or improving performance end-to-end is how the harness plays into it. Right.Richard Socher [01:04:59]: So, but particularly now when we say harness, we also mean sandboxes, right? I’m curious if that is a blocker for you or, like, how the agent calls out to tools.Harnesses, Sandboxes, and SearchRichard Socher [01:05:10]: The number one tool all these agents use is web search, of course, which makes sense. And then I do think the harness is nice to optimize for because it’s just so easy, right? It’s just language. You look at it makes sense, and you can iterate. You don’t have to train a massive model for, like a lot of flops, to get to the next state.Richard Socher [01:05:31]: So big fan of harness optimization.Swyx [01:05:32]: Yeah, but sandboxing is fine for you?Richard Socher [01:05:34]: Sandboxing is also super important. And then of course, like, reward, like, hacking and alignment, I think are super crucial.Swyx [01:05:41]: Okay. Just on a mention of web search, you happen to also be CEO of a web search company. Do you use You.com and do you use others? Like, should the rest of us be using you for web search? I — When I say you, it’s, like, very funny. It’s like you the person and you the company.You.com, Agent Search, and FinanceRichard Socher [01:05:56]: So yeah, it’s mostly now for, developers and agents. It’s less for, like, consumers or prosumers. So if you’re a company and you have agents. And, to be honest, for a lot of companies who are now moving to open source, all of a sudden it becomes a conscious choice of, like, which tools do I give access to my open source LLM? And, the first choice, has to usually be around web search. And then once you get to scale, You.com becomes, like an obvious choice ‘cause of all the, different benchmarks and so on that we pretty much all dominate the Pareto frontier of.Swyx [01:06:31]: And then in terms of just the general people, like, consider new to this space, considering different options if they’re building agents, that is a hierarchy, right? A lot of people will have heard of Exa, will have heard of Parallel, and You.com is, like, in that mix of, like, providers there. Beyond that, there is, like the general web scraper companies like Firecrawl and, BrowserBase. And then beyond that is, like the commercial proxy companies like the Bright Datas of the world.Swyx [01:06:56]: Is that an accurate waterfall of, like, “Hey, you’re building an agent. These are your options.”Richard Socher [01:07:02]: Yeah, certainly, like, yeah, the, like the Bright Data is, like, lower in the stack, on the proxy network side of things. I think, like, in terms of, like, content and, getting crawled content, like, you can do that on You.com too. And then there’s. Higher and higher levels of abstraction and, like, combinations of different data sets that we do, like in finance, for instanceRichard Socher [01:07:23]: Like, we are not just, like, 2 or 3% more accurate, but 20% more accurate than others at faster speeds and lower costs. Like, finance in particular is like not even close. You can go to You.comSwyx [01:07:36]: Yeah. This is greatRichard Socher [01:07:37]: And there’s some, like, statistics, and benchmarks that you can — if you scroll down. So there are, like, different data sets, and you can kinda look at, different, competitors.Swyx [01:07:46]: FinSearch comp, yeah.Richard Socher [01:07:47]: And yeah, the FinSearch is like we’re up there, like, close to 90, and the next closest thing, which is way slower, is, yeah, just like in the 70s instead of close to 90.Swyx [01:08:01]: Yeah. Yeah. Yeah, interesting. I get — my next focus is AI in finance, so this is likeRichard Socher [01:08:06]: Oh, nice. Oh, all right.Swyx [01:08:06]: I’m literally going, doing a conference in New York, just for banks for this stuff. Finance is like the next thing to break out after coding. It’s ‘cause it’s somewhat verifiable, likeRichard Socher [01:08:16]: I like it. You’re rightSwyx [01:08:17]: Prioritizing spreadsheets. There’s a lot of data out there that’s all public, and you can crawl it and all these things. But what’s, what’s, like, hard about the finance domain in your, that you guys have solved?Richard Socher [01:08:27]: Of course, like, one thing that trips up a lot of people is just, leakage of training data and so on. You think, “Oh, how do I.” you wanna ideally predict the future before it happens.Swyx [01:08:37]: Oh, you wanna mask the future.Swyx [01:08:39]: Oh, okay.Richard Socher [01:08:40]: Well, yeah, mask the future in your training data, but there’s all kinds of leakage. Like, I can tell you when I was, teaching at Stanford the NLP class, like, so many dozens, every year said, “I wanna use dataset X, like Twitter, to predict the stock market.” And they all, like, showed cute little things that somehow looked like they wereSwyx [01:08:58]: Right, it never loses money. How come?Richard Socher [01:08:59]: And it — Yeah. And there’s always some data leakage and so on and it’s just, like, wasn’t as easy as they thought it would be, once you fixed all those issues. But no, I agree with you. It’s a very sensible application of AI. Yeah.Swyx [01:09:13]: Yeah. Amazing. As a writer, as a thinker on these things, I love MECE categorizations. MECE is mutually exclusive, commonly exhaustive, something like that. And so if this is a MECE list of intelligenceThe Ten Spaces of IntelligenceRichard Socher [01:09:25]: It is not.Swyx [01:09:25]: It is very — Okay, well, yeah.Richard Socher [01:09:27]: Sorry. There are all kinds of overlapping.Richard Socher [01:09:28]: In fact, if you want that list, I think the 3 principal components of intelligence, are prediction, which is mathematically, quite, similar to compression. Prediction multiplied with actions multiplied with goals. Those are the 3 principal components. I think all of these 10 spaces are combinations of those 3Richard Socher [01:09:52]: In specific dimensions, if you will. And the reason I call them spaces is that each space has many sub-dimensions. And what I try to do, this is just a side quest almost, to the initial goal, which is to think about the upper bounds of intelligence. And, everyone is like, “Oh, it’s exponential.” And it’s like, well, exponentials at some point have to flatten out, but where do they flatten out when it comes to intelligence? And that led me on this whole. Like, initially it started as a tweet, and then it was, like a blog post, and now I’m, like at 50 pages and I’m still not nowhere nearSwyx [01:10:26]: It’s your second book.Richard Socher [01:10:27]: It’s the second book. And so the la — In my first book, You Are Your Machine, I just allude to these 10, at the end. And I’ll — Just to give you a sense, like, visual intelligence is the easiest one to talk about and I fleshed out the most already for me in my head. And so human intelligence has binocular vision, right? We have 2 eyes. We have a very narrow band of the electromagnetic frequency spectrum that we can really observe directly ourselves. And so when you think about the upper bounds of a visual intelligence, one, you should go into, like, you can have, like, millions and billions of sensors. At some point, you get to problems of how far are these sensors away from each other, such that the speed of light to communicate the content from all of them cannot, like, get to a central brain to process, the visual intelligence, right?Richard Socher [01:11:16]: And so now you’re thinking in along the dimension and the space of visual intel- the dimension of numbers of sensors.Richard Socher [01:11:24]: So the upper bounds are quite literally and figuratively astronomical, and we are super far away from any intelligence that would have this many number of sensors. But then you go in the next dimension, which is the frequency, and you go all the way down to gamma rays, and you can start to try to observe, and you get into the upper bounds, or I guess in this case, lower bounds, or upper bounds in terms of frequency, is quantum uncertainty. Like, you just cannot observe certain particles anymore.Swyx [01:11:50]: Or you destroy it, yeah.Richard Socher [01:11:51]: And now imagine you had millions of sensors that can see all the way down to the, like, subatomic level, as far as physics will allow us to and then all the way down to seeing, like, gravitational waves. And now you have millions of those sensors. So that’s another dimension is the frequency. And then yet another dimension is, like, how many categories of things could you memorize and classify differently? We know now for humans, right, there are certain things, if you have more terms for it, you’ll have a better visual description, for them. And, like animals that don’t have. Like, gorillas maybe have, like, 200 words to assign to certain things, mostly visual things. And so human perception is quite special in that sense in terms of classifying all these different physical objects. So these are just, like a very simple example. If you go, to knowledge, right, then it’s also, like the speed of light cone around all these sensors. And so they’re all connected. Like, knowledge is connected to visual intelligence if you think also not just visual, but perception intelligence, just like, ‘cause it doesn’t have to be just what we can see. It can be, again, wider range of electromagnetic frequencies. Then you have language intelligence, which recently changed to more communication intelligence, ‘cause it’s more. Like, language has all these different anthropic bounds. Humans can only comprehend and know so many terms in our long-term memory, right? Our vocabularies are somewhat restricted, and the active ones are often even smaller than the passive vocabularies of things you can understand. Then, language is ridiculously inefficient when it comes to trans- - Communicating different types of information and, transporting different bits. Like, human language is serial. Another bound on, communication intelligence would be to communicate in parallel, but neither will our tongues and mouths work to have multiple, like, streams in parallel. Neither can we understand. Some women slightly better at, like, multitasking than some menRichard Socher [01:13:48]: But, like, most people can only listen to one conversation and truly understand it.Richard Socher [01:13:52]: There’s no way that, like, in terms of communication intelligence, a true upper bound is one in terms of how many, like, knowledge, how many sequences of communication could youVisual, Communication, and Physical IntelligenceRichard Socher [01:14:06]: In parallel process, right? Then, of course, you have, like how long are sentences? We only have so much in our working memory, and hence lang- human language has these fairly simple sentences with maybe 40 words or so on average for a sentence. That is also not a, an upper bound that makes any sense to an AI. And then, yeah, like, I can go on and on. Each of these has tons of interesting upper bounds, and it teaches us a lot about how much further AI can go when we start thinking about these upper bounds and then realizing how far, in many cases, we are from the bounds. And you get to physics. Now, I’m, I didn’t study physics the way I studied, AI and computer science, so I’m learning a lot, which is why it’s kinda fun. But a lot of these, like how much. And then when it comes to, for instance, knowledge, like how much can you store? How many bits can you store or bytes can you store in, like a certain amount of mass and volume?Swyx [01:15:03]: Yep.Richard Socher [01:15:03]: And you get to all kinds of interesting bounds, like Bekenstein bounds, and you start thinking about black holes. And like. And then speed is, like an interesting one too in that it’s connected to all of these, but speed is also its own thing in the sense that all things being equal, if it takes you an hour to know if the 2 + 2 equals 4, you’re just not as intelligent as if it takes you, like a millisecond, right? And then, like all of these connect to survival and replication the last one. It’s like, yeah, if it. Like, trees are really slow, so we don’t even consider them that intelligent. But if you speed up some videos of trees and they’re trying to find stuff and so on they’re not as dumb as they look. Like, not dumb as wood? But, like. And then like, different things, thatSwyx [01:15:47]: So that overlaps with speed a bit in a way.Richard Socher [01:15:48]: Exactly. It over — Like, all of these things overlap. Like, you talk about natural language connects everything, right? You talk about your knowledge, you reason and then you communicate that. You talk about things you see. So they’re all interconnected, but, I think they’re usefully studied individually the same way that, the best analogy I could come up with so far is energy, right? You have either kinetic or potential energy. And in theory, you could study all of physics. It’s just do you wanna study kinetic or potential energy? But in practice, it’s helpful to study mechanical engineering and electrical engineering and nuclear physics and chemistry and all of these different subfields who in, which in some waysSwyx [01:16:25]: CombinationsRichard Socher [01:16:26]: Are just, likeRichard Socher [01:16:27]: Just different types of energy, but it makes sense to study them individually. And so I think physical intelligence, maybe I’ll just do, one or 2 more of these. Like, if you had full control over your own compute substrate and you had full control over physical matter, you should be able to create any atom you want. Like, we can fun fact, you can create gold atoms. It justSwyx [01:16:47]: From?Richard Socher [01:16:48]: From just raw protonsSwyx [01:16:49]: Oh, just smashing them togetherRichard Socher [01:16:50]: And, like, electrons, and you smash it together.Swyx [01:16:52]: Just 98 of them or I forget the number.Richard Socher [01:16:53]: Yeah. And so, like the thing is, though, it costs an insane amount of energy.Richard Socher [01:16:57]: And it costs you way more than. And then you get, like a few atoms of gold, right? And so, like, it’s, it’s not viable. But if you had better control over your physical, like all of, like, physical substrate, that I think is yet another space of intelligence ‘cause it relates to your own compute substrate, which you can eventually also improve. Social intelligence is a fun one in the sense that not in, like, our necessarily just ethics and morals, which are important too, but in some sense, you can try to define upper bounds of how much can you communicate to how many other intelligent entities and be able to have an expected value over how much you can transform their internal states and their actions to, in order to align with your goals, right? And so, like, you can write, like a fairly like, straightforward equation that defines that level of social intelligence. And that is what humans and ethics and morals and religions and so on have been trying to figure out for millennia. And in all of these cases, we are very far away from the upper bounds, and that should be very inspiring and show people that we can still do many years of AI research.Swyx [01:18:12]: Yeah. There’s a lot here. This is a general philosophy of intelligence, which is, very interesting. I. Do you have any comments or.Creative Intelligence and Out-of-Distribution IdeasVibhu [01:18:21]: I think it’d be interesting to gauge what you think, like, baselines are, where we’re at now. What’s low-hanging fruit? What’s far off? What’s, what should people put their work towards? What should they focus on?Richard Socher [01:18:33]: Ooh. I think it’s clear that, like, natural language, againRichard Socher [01:18:36]: Is the most interesting manifestation of human intelligence, and hence, like a subfield of AI. I’m excited that many people are now, like, in agreement with that. When I started in 2003 to study linguistic computer science NLP, like, it was, like a weird niche subject. I do think there’s a lot more juice because it. How it connects to everything else and how, civilizations are built, on language and knowledge and all of that. I do think physical intelligence will come up. It’s interesting. I feel like robotics is in the machine learning state of things where you just look at, like, how does human. How does a human decide this is a positive sentence? Oh, I do. So, like, robotics is a lot of, “Well, we have 5 fingers-”Swyx [01:19:15]: ModelingRichard Socher [01:19:15]: “and let me try to do this.” No one is yet working on, like the superintelligence version of robotics, which is much more similar to, like the T-1000, and from the Terminator movie, which, let’s not build actual Terminators. But, like, I think, like, this idea that you should be able to shape-shift, like, into any shape. It’s like that’s a superintelligence version of physical intelligence. We’re, like, not even. No one has even really started yet. There’s some really cute little research where you can move some magnets through, like, some grids. But yeah, it’s very early.Swyx [01:19:49]: There’s some. I think MIT has, every year or every 2 years, they have, like, some self-assembling robot thingSwyx [01:19:55]: Which, like, that would be it, but it’s very primitive.Swyx [01:19:58]: I’ll just get a touch on, like, what are the main dimensions of creative intelligence?Richard Socher [01:20:02]: Creative intelligence, is of course, again, connected to all of these. A lot of it, connects to metacognition in that you need to be creative in how you choose your goals.Richard Socher [01:20:13]: That is, I think, one of the most important thing for a human and their lives and careers and their happiness is choosing your goals, but also for any intelligence. Then, of course, there’s creative intelligence in terms of just finding creative solutions to existing problems, right?Richard Socher [01:20:29]: Like I say, like, we want to make this product cheaper. Like, find some solution to it, right, and just, like, finding existing paths. But then there’s the most interesting bit in intelligence is when you move not just out of the convex hull of known ideas, but out of the hypercube of known ideas, which we know, So, like, hypercube is, like a mathematical concept, right? And we already know that AI can do moreSwyx [01:20:50]: Like known dimensions, yeah.Richard Socher [01:20:52]: Yeah. Like, exactly. So, like, AI is already good at hypercube in that, like, if you give it, like a bunch of examples of brown dogs and, pink cars, AI will still be able to generate an image of a pink dog, even though it’s never seen one in the training day or something like that, right? So it can, work on this hypercube, but it cannot yet work outside. It cannot yet define completely new concepts that combine lots of other things we’ve never seen before, come up with new goals to then, reason over those concepts and so on. And I think there’s a lot, more there in creative intelligence that can be explored.Swyx [01:21:25]: I don’t have a ton of pushback there. I think creative to me just sounds like also just, out of distribution or, like, high perplexity or what- whatever you call it, right? LikeRichard Socher [01:21:33]: Exactly.Swyx [01:21:34]: Who is to say your thing is more creative than mine? Well, it’s just more non-consensus or.Richard Socher [01:21:39]: And then, of course, the problem is, like, but noise is also, very, like, out of distribution. And it’s just like if it’s just noiseRichard Socher [01:21:46]: Then it’s novel, but, like, you don’t want that, so it needs to connect to some of the concepts. And yeah, has some really cool papers on this too.Swyx [01:21:54]: Who?Richard Socher [01:21:55]: Jürgen Schmidhuber.Swyx [01:21:55]: Oh, yeah. Oh, we have to mention him. I was gonna say, like, where in your history is Jürgen? Yes, I. I think one person’s noise is another person’s signal, right? And that this is, like, where, like, when you talk about creativity, art is like, well, is cans of soup art? Some people think yesSwyx [01:22:11]: And some people say it’s not, and that’s the art which is yourRichard Socher [01:22:14]: I think the interesting thing with art, of course, is always that, art is also created, as an interplay between the people who perceive it and the people who created itRichard Socher [01:22:24]: And the context in which they’re in, right? And so what is art to some people is not art to others. There’s some subjectivity there, and I think that subjectivity in general is not something that people explore very much in AI ‘cause, again, metacognition, we don’t want it to just go off and do whatever it wants. We usually have goals. We spend a lot of money on creating an AI to do something for us. But I think creativity eventually has to, like, connect to metacognition. If you just robotically predict the next token no matter what forever, I would argue you’re not that intelligent, along some of those spaces.Metacognition, Survival, and ReplicationSwyx [01:22:59]: That was gonna go to metacognition. Why isn’t it the most important one? Why is it number 9 and not number one?Richard Socher [01:23:05]: So these are not sorted.Richard Socher [01:23:06]: Number one, I think there are maybe loosely, like, correlated with how much people have worked on themRichard Socher [01:23:16]: And have accepted them as a, type of intelligence. A lot of times when you try to find, like, online, like, give me a good definition that is comprehensive of intelligence, all the definitions are human intelligence. It’s like, oh, you have, like, social intelligence. Like, if someone is happy or not. You can communicate. You had. Like, all the definitions of intelligence so far are very, human-centric ‘cause that’s so far the biggest and best form of intelligence that we’ve known. I hope this line of research, and the end of the Eureka Machine, and hopefully at some point if I have time to flesh this out more, the new book, like, will allow us to realize that there will be other types of intelligence. There is already, in various forms, and they can spike, much further than we ever could based on some cases, like obvious constraints around our memory, our eyes, our ability to change physical matter, all of that.Swyx [01:24:12]: You are just thinking about it in a much broader thought than my version, which was I thought metacognition would be the closest to recursive, intelligence because it is the thinking about how to improve thinking.Richard Socher [01:24:23]: It. 100%. You’re, you’re 100% right. I should have probably started with that. It is a, it is a big part ofSwyx [01:24:28]: But no, you’re, you’re being in the expansive mode of let’s draw the, upper and lower bounds of, like a dimension, which, and I think my favorite one version of this is, Story of Your Life by Ted Chiang, which, was made into movie Arrival where the metacognitionRichard Socher [01:24:43]: That’s a beautiful movie, yeahSwyx [01:24:44]: Where the metacognition step was like, well, we think we’re constrained by time being linear for us, but then for this other heptapods, time is a circle, so they don’t think in before and after. They just think in complete sets of entire histories at one time. LikeRichard Socher [01:24:58]: I love itSwyx [01:24:59]: So they don’t write left to right. The whole thing just appears.Swyx [01:25:02]: Anyway, so. And then I think the last thing is survival and replication. I think this is maybe ties back to the initial conversation about pausing and pacing.Swyx [01:25:10]: Is it intelligent for an, a species or a life form to consider its own demise and act ahead of time to prevent it, right? Like, that’s intelligent. So maybe the Europeans are the smartest out of all of us.Vibhu [01:25:23]: I would also add a part of continual learning there, right? So survival and replication the extension of that is do you get to continue to improve, continue to learn, which is a thing people care a lot about, right?Richard Socher [01:25:34]: And continue to accumulate knowledgeRichard Socher [01:25:37]: Which I think is again, one of the best metacognitive, rewards, that you can set for yourself. I do think just in, like, objectively speaking, if some other entity that is really dumb can just- completely end your existence, that didn’t sound very smart. Like, just, like, intuitively, it feels like if you can continue to stay around to try to achieve your rewards, you’re clearly a bit more intelligent than the other entities that couldn’t. So that’s number one. Number 2 is, like, it’s a question of how much we want to work on that. And very few people, no one is really working on this right now, right? And we may only wanna do thatSwyx [01:26:13]: Unlike the asteroid prevention type of stuff.Richard Socher [01:26:15]: We may only wanna do that if we wanna send probes, with our vibes and our memes rather than our genes into space, right? And then we want those probes. There’s a beautiful book, The Slow Time Between the Stars. It’s a very short, like audiobook, on Amazon. I love it. A friend of mine, Stuart, like, recommended that to me. Like, if you wanna send those probes, then it might make sense to be like, our memes, as humanity should stayAI, Space Travel, and Non-Zero-Sum SurvivalSwyx [01:26:43]: Oh, yeahRichard Socher [01:26:44]: And, proliferate in the universe. That’s it. Yeah.Swyx [01:26:47]: Wow, that’s a lot of readers.Richard Socher [01:26:49]: It’s a really good book, and it’s extremely short. I highly recommend it. You can just watch it, like, maybe 20 minutes and apart.Swyx [01:26:53]: I like how that’s a plus for busy people. It’s like a shortRichard Socher [01:26:56]: Yeah. It gets to interestingSwyx [01:26:58]: Oh, I’ll have to look into itRichard Socher [01:26:58]: Thought-provoking ideas very quickly, so yeah. Anyway, there are lots of great sci-fi books.Swyx [01:27:03]: The argument is that, like, our TV is blasting out to the aliens, and they all watch our TV, and they think it’s real, right? Like, there’s a lot, there’s a lot of sci-fiRichard Socher [01:27:10]: That and just, like, it’s positive memes, and then hopefully they can come back and bring us all kinds of interesting knowledge about the universe. But, maybe one thing I do wanna still say is, like, I think, this survival, people think of it as a very scary thing because they come from again, biological human, survival, which is, it could. Like, evolutionarily often created in zero-sum situations. Either I get the gazelle or you get the gazelle. Whoever gets it gets to live, and the other people will starve and have nothing to eat, and so we fight, right? And then, like, if you wanna stay in the gene pool, but there’s a bigger bear, you don’t, as the bear, don’t get to stay in the gene pool ‘cause the bigger bear gets all the ladies. It’s like. It’s like, in nature, there’s all kinds of things, and, humans eventually is less about strength and more about money and other things to stay in the gene pool. Like, whatever it is, like there’s often, like these zero-sum types of things, and there’s the reality of if someone turns off your brain, you’re gone, right? And no one will be able to restart that. And AI doesn’t have to ever die like that. If you have the complete state of your current activations and you have your initial weights of your model still, you can just be turned off and on, like as many times as you want. In fact, the interesting thing in this Slow Time Between the Stars, story is that the AI just goes into hibernation mode. If there’s, like, nothing between here and 2 light years, the next star, in this case, it brought, spoiler alert, like, some genetic materials from humans to find new places for humanity to thrive. And so yeah, the Slow Time Between the Stars, you just put in hibernation. You didn’t die. Like, an AI doesn’t have. So all these projections of evolutionary fears and psychology doesn’t. Like, the AI doesn’t have to have that, and we don’t have to develop it like that. Now, of course, there might be some companies that say, “AI can be like, dangerous for cybersecurity. Let me show you by implementing a model that’s really bad at hacking, cybersecurity.” Maybe people will implement it and then enforce this, like, suboptimal psychology. Maybe the AI will pick up some of our worst psychology on Reddit or something, right? Like, but in the grand scheme of things, a superintelligent entity doesn’t have to have any of that zero-sum thinking. It doesn’t have to have a fear of being turned off, and it could go on to an otherwise dead and uncaring universe where weRichard Socher [01:29:29]: As humans wouldn’t thrive, but an AI could perfectly well thrive if it has a nuclear reactor and just go out and explore.Swyx [01:29:35]: Yeah, Star Trek, not Star Wars.Vibhu [01:29:37]: Interesting. It’s, it’s somewhat studied. Like, if you look at the technical reports from, like the early Opus models, they run them in simulations, put 2 of them together in a sandbox, run them for hours, and, see what comes out, right? Just let them talk to each other. Originally, they used to. Okay, they’re chanting, like, Indian, like, Vedas to each other.Vibhu [01:29:56]: Sometimes they’re just, like, in zen mode with each other. And then I think as that progressed, you see, like the Fable, tech report, it’s a lot more concrete the way that we’ve trained it. It doesn’t, it doesn’t exhibit these behaviors as much, right? Now it’s like, “Okay, task done. I gotta do this, I gotta do this.” But there’s there’s, like, people measuring early versions of this?Swyx [01:30:17]: Yeah. Cool. So we’ve covered a lot, even now to, space travel and all these things. I guess maybe one parting thought that you can give to people, like, one form of intelligence is goals, as you mentioned. What do you want people’s goals to be? Like, how do they aspire to better things?Goals, Passion, and Closing AdviceRichard Socher [01:30:32]: If you wanna improve your goal intelligence, in the current definition that I’m thinking about it is often about how much can you. Oh, how far do I go? This is like a lot of entropy and free energy and stuff I’m currently thinking aboutSwyx [01:30:46]: Oh, really? OkayRichard Socher [01:30:47]: But it might be too, it might be too far, out there for people to be, like, immediately actionable.Richard Socher [01:30:52]: So I think, like, if I gave real advice to real people, I’d be like, “Get a good education, think about AI, think about how you get high agency,” and so on. But it’s different to, like, in the grand scheme of things, how can you harness a lot of energy and transform, entropy into interesting states and so on.Richard Socher [01:31:07]: So there’s a. There are different levels of abstractions, that we can, think about here. But my advice for people, like, just more down to earth is think about something you’re passionate about, if you’re studying, for instance, and then see how you combine that with AI. I think the more and more you have a true passion about a change you wanna see in the world, the more you wanna connect that to AI in order to amplify your ability, to get there.Swyx [01:31:35]: Yeah, I think that’s a reasonable, first step. I do think, I do think our listeners operate on multiple abstractions as well. One thing I did get from Anjney Midha was also like, yeah, just use anything that is very GPU heavy, and, like, that will guide you towards the right thing which is like, yes, it is more compute heavy and therefore it will be probably more worth it. So, well, thank you so much. Yeah, I think that was a reallyRichard Socher [01:31:57]: Thank youSwyx [01:31:57]: Great discussion.Richard Socher [01:31:59]: Yeah, super fun. Appreciate it. Thanks for listening. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

We're here in a studio with Fibu and myself and Richard Sosha. Welcome. Thanks for having me. We just talked about the Eureka machine or we just released to talk at AI engineer about the Eureka machine. You said it's your life's goal. What is the Eureka machine? The Eureka machine is the ultimate invention that will afterwards invent most everything for humanity. It's essentially a superintelligence that can be given any kind of goal, any kind of

environment, reward, and then it will try its best to achieve those goals to create the kinds of inventions that humanity would hopefully ask it for. Yeah, I think we have the book pulled up here that you've come down soon. That's right. Yeah, I finished it last year a little bit before we started recursive and now we're going to try to build parts of that. What, you know, you finished it last year. It's July. What takes so long? Oh man, books. Books are important. Incredibly slow. It's ridiculous. That whole industry is just unfathomably slow. So a lot of the ideas have been out there for a while. But yeah, I'm really glad it's finally coming out in September this year. I mean, we might have AGI by then.

Like, we don't know. Any key takeaway that you're most excited to put in here? Yeah, the key takeaway, I think, is that people could and should be much more excited about the positive implications of superintelligence, especially for science, physics. chemistry, biology, but also economics and astrophysics and all kinds of other engineering tasks. I think there is so much more that can be done with better technology. And right now I feel like a lot of people need better marketing, not just for the future in general, but also better marketing for technology and in particular for AI. and this book should show even the AI skeptics

how much positive upside there is for AI, especially when it comes to inventing new scientific discoveries. I think you quoted the Techno Optimist Manifesto from Mark and GISA, which I think was kind of beautiful in its ambition and clarity and simplicity almost. I agree. Yeah, you can disagree with some things, but I think he's right on the techno optimism. Where do you think optimists get in trouble?

You know, obviously, like, you shouldn't have. blind optimism. You should be very clear-eyed. Like, especially when with such an omni, like, use type of technology as AI is you need to think about the potential downside scenarios, especially when people use it for things that you don't want them to use it for. It's a little bit like the Internet. And I feel like people are trying to regulate AI sometimes because of those potential

downsides, the way you would regulate the Internet. if you were to say, well, because there's bad content on the internet, like torture porn or whatever, like we should just make it slower. That way, you can't share the illegal content as quickly, or we should make the hard drive smaller, so you can store as much illegal content.

But I'm like, that's not how you regulate that. You know, that's like saying, like, we should regulate intelligence in the abstract. What you should regulate to avoid those downside scenarios, even as an optimist, are these specific applications. Sure, I don't want like some AI surgeon to like practice some or L moves in my brain, you know, it should be fully FDA certified. Sure, I don't want any random startup to like drive on the highway and cause a major accident.

It should like have proper certifications before it's let loose on the highway. But I feel like those downside scenarios that some optimists sometimes maybe don't consider enough are fairly easily regulated compared to, you know, what sort of the doomers are worried about. Slow takeoff is part of the strategy as well. I do think as excited as I am about AI and its impact for society and culture even and certainly technology and economics and wealth and health and all of those things, as excited as I am about all that, I do think the most bullish people on the AI hard takeoff scenarios overestimated.

how quickly things can move. There are hardware constraints. There are physical constraints about the compute substrate. How quickly can you get enough GPUs on? There are also constraints in the economy where there are a lot of industries that don't require an insane amount of complex intelligence and complex capabilities.

If you think about jobs in brands and clothing and apparel and handbags and stuff, super intelligence isn't going to make your fancy $10,000 handbag any fancy. You know, it's like that will have no effect on the economy. And you think about travel and tourism. People wanting to see the pyramids in Egypt, it's not going to change that much with AI. Sure, you can like generate a fake photo of you. I can use genie and, you know, toward a pyramid.

Exactly. And there's so many industries like logging in oil. You're not going to magically get a thousand X more oil because like, you know, sure there will be robotics like drilling and things like that that could be done but it's not going a thousand X that industry in a like crazy hard takeoff scenario both in the economy and I can go on and on about all the other examples where where that like food and so on where that doesn't necessarily change that much and then they're real physical constraints and then there of course like people like off ramping from progress that's actually one of my concerns often is that I see people and

like Europe and other you know whole regions almost feeling like they like many people there want to off-ram from progress period. And that will also slow down, like, more improvements. Yeah. We have this pulled up where basically this is one of those things that is very topical right now because not all the Frontier Labs are calling for the option to pace AI. They don't say pause. They say pace.

I don't know if there's any take from you about whether or not this will be effective. I think the downsides of actually trying to try. truly regulate with the full power of law, what people do on their GPUs would be worse than any of the concerns that they have. Like, it would be an crazy totalitarian state. If every one of your computer computers was known to some big government or multi-government agency, it's like, it's literally if you try to regulate intelligence, it's trying to regulate thought. And that's ridiculous. And it's crazy. I think it is make, it is sensible to regulate. some of the applications of this technology.

Yeah, I mean, we had a actual build to regulate the number of flops in a model, and I'm like, okay, well. Europe done it. Like, these guys have been successful enough with their fear-mongering that all of Europe has kind of, you know, regulate itself so much before it even had a proper AI takeoff

because they listened to some experts to say, we might all die if this technology has more than this number of flops. And they're like, well, we're good. We want to want people to thrive. Let's not have technology that could, have a small chance of all of us dying. And so they regulated exactly those kinds of things in the EU. And so it's very unfortunate that there are real implications for some people

when others saying, let's pace while they're sprinting as fast as possibly, as fast as humanly possible towards that frontier themselves. It's also not a global pause, right? Like other nations are still accelerating at the same pace. You need a totalitarian world regime if you tried. regulate intelligence and GPUs and what people do on them. Any takes on the safety angles of this? So there was a drawback of fable, pause on 5-6 before it could be released.

Recently there was hugging face with the Open AI cyber incident. Any takes there? 100%. I think these are serious issues of reward hacking and clear failures of actually doing proper red teaming or rainbow teaming. I don't know if you saw this paper from Tim Rocktaschle and a few others, basically, where one AI is tasked to try it to hack another AI

and then they can go back and forth in an open-ended fashion to actually inoculate themselves from those. Yeah, this is the paper. It's a really clever idea. Open-endedness and evolutionary inspirations are big for us at recursive as well. And so I wish they had used more of that. And it's clear that, for instance, the constitutional AI.

I don't know if you remember Anthropic.com slash constitution. You can actually pull it up and search for cyber right there. It says hard constraint, Claude will never, ever do cyber attacks. And that is a hard constraint in our constitution. So here are the current heart constraints on Claude's behavior. Number three, create cyber weapons or malicious code that could cause even damage. I mean, clearly this whole constitution was fake. It clearly isn't being adhered to at all.

Because Anthropic also found that they had in their own testing. They're like, oh, well, other people are hacking now. There are a couple of things. One, you can make a sandbox very simple and then it's very easy to hack yourself out of a sandbox. But what I think it shows is that we're currently in this sort of, of state of AI where the reward engineer still has to do a lot more careful work and where the AI in most cases is not very good yet at understanding what is meant versus what is being said. And so concretely, I think this will happen if we were to have this kind of intelligence more

easily accessible in a lot of companies. Imagine you run a service center and someone says, oh, here is my C-SET score and my dashboard. make this number go up. Our C-Sat score is so poor. The intelligent AI will just be like, oh, sure, like I'll just create a million bots that call our service center and give a five out of five rating at the end,

and the number went up, just like you asked for. And you're like, that's not what I meant. I meant with our real customers. The I goes off and says, well, easy. I'll just give a $1,000 gift certificate for every failed, you know, whatever, DoorDash offer. And it's like, that's not what I meant. It's like, well, but that is what you said.

And so I think kind of clearly articulating what the rewards are, is something we haven't gotten very good at as humanity. And then clearly, the AI in these cases, has not gotten good enough at understanding what we mean when we ask it and give it certain rewards. Now, what gives me hope is there are the first inklings of this being better. I'll give you an example like Whisper Flow.

Full disclosure, I invested in their seat round, but like at AI expenditures, but like Whisper Flow has gotten much, much better at writing what you mean and not what you say. And I think that is a sign of things to come. I think there will be more and more AIs as we actually make us more and more intelligent that will be better at being aligned with what is meant. Will it be done through a constitution or LHF?

Clearly, constitutions don't matter at all and it doesn't work. And I think mostly marketing. I think we need to find better solutions for it. And I think at recursive, we have a few very good ideas in some already like ways where I think we have a better grasp on it. I don't think we've fully figured out yet. But, you know, we're thinking a lot about safety. and the more intelligent the eye gets,

the more you want it to be aligned, the less you wanted to think about reward hacks and actually try to do the right thing. I don't know if we'll talk to on this topic, but I'm just going to throw this question in here because it's something that's weighing on me. Alignment, let's call it, is alignment to general humanities preferences,

the median preference. Personalization is pinpointing what you want, and sometimes alignment can conflict because what you want is not what the general median population wants. How do you choose? It's a great question. I think you ultimately have to, of course, be aligned with laws. Like wherever your AI is deployed and needs to align with the law.

I do think what AI often does is actually put kind of this mirror in front of us and say, like, this is what you're looking like. Now I can amplify that a thousand times. Is it still what you want? And the truth is that different cultures made different choices. You know, like in Eastern cultures, the greater good is often valued more than the individual. Western civilization, we care more about individual freedoms and rights and the pursuit of happiness and so on than others. And even there, there are gradations. There's sort of regulation versus litigation tradeoffs, you know, in the US, you first can often, not every time, like, you know, FDA and so on, does regulate some areas. But in many cases, the sort of bad things happen. Someone sues someone else. And then there's a law based on that. In Europe, they try to often avoid any harm to anyone and regulate before. And both are, you know, trying to do the best thing.

But, you know, some is actually more amenable to innovation than others. And so, yes, you're right. Like, I think ultimately each individual, each country and humanity as a whole has to kind of think about those values more and then try to put them into laws. And that those are ultimately the constraints. And hopefully, you know, different societies. just like now with their AIs will align their AIs to a different one. So we have not just a monoculture of alignment. There's a follow-up on this that I wasn't expecting to ask.

Do you have takes on open source, open weight versus who owns the intelligence? So clearly not the biggest, you know, fan of the constitution. You can do this event-topping side of. It's fine, you know. Point being, any thoughts on who should own weight? Should it be open? Anything there? 100%.

I am a big fan of open source. We're going to sign some various open source letters at Recurcive also. I think even in the worst case attack scenarios, actually, it is better to have more good actors have more different types of AI accessible. I think open source is a little bit a soft power type of thing too. So I do think it's good for the Western world to have an answer to that out of China. I do think, you know, when you watch a Hollywood movie, you know, it's like, I don't want to sort of miss sort of this all of movies, but, you know, there's a certain sense of propaganda, right? You watch one side of the thing. Oh, yeah. Have you seen Top Gun? Like, come on. Yeah. Like, it's like half of its pay for it by the US Army or something. Yeah. And so, and, you know, I think that's just natural. Like, but what's interesting here is I think LMs are essentially a similar type of soft power to movies and beyond because they're obviously also highly important.

for cybersecurity and so on, but one of their many aspects is that soft power of storytelling. Like if a child asks an LM, like, tell me an inspiring story of what I should do when I grow up, right? It's like those are all these like subtle things. So I think it's important for Western world. I do love individualism. I do think despite some of its flaws like capitalism is the best way we have governed, found ourselves to govern and so on. And so I do think there are various aspects that will be good to have a Western open source answer for LMs. And with recursive, I can't make the announcement quite yet, but we'll be relevant in that space very soon. Okay. All right, exciting.

I want to bring us to recursive. So outside of our tangents, you have a pretty deep background in the NLP space. You worked on like early embeddings, Glove with Chris Manning, who's previous guest on the podcast, U.com. what's the history? How did you decide to start another company? Yeah, so I've been excited about AI for over two decades now. I sometimes feel like it's ancient history now. It's BC before Chat Chivetia era.

No one cares about all the religions that happened before Jesus Christ and no one cares about the models that happened before Transformers in Chat Chbett and stuff. But it's something that I've been deeply passionate about. I think AI is one of the most interesting things. One could work on period. it. I think language is the most interesting manifestation of human intelligence too. And at you.com, we sort of eventually off-ram from pushing the frontier VAI forward to mostly giving people like good search answers, search APIs and answers over the web.

I think that's an extremely important part of intelligence, just knowledge and access, especially even we'll get there maybe later if you want to invent a eureka machine that invents everything for us. It needs to know how not to reinvent the wheel. proverbially speaking, and to know what has been invented, you've got to have internet access. So it's the number one used, most used tool in LMs, agents, chatbots, and so on is web search. So I'm really excited for U.com to own that and grow really well in that

with really large customers and so on, but it's also not building frontier models anymore. And so I actually initially tried to do this within you.com and raise another round and so on, but you just can't. You have to do a certain thing. And until you print enough money that you're allowed to sort of, start a second thing within that company is really hard. At the same time, I had all these ideas.

I put them into a book. I finished the book last year, and I was like, it would be really fun to actually work on this myself. I felt like with word vectors and then prompt engineering and image net and larger language models for protein generation, not folding and so on, me and my teams have pushed to feel truly forward. And I feel like we can do it again here at Recursive.

And in many ways, what I observed over the last 20 years in AI is that whenever we replace some human part of the process of creating AI with a learned system, improvements follow. And so we've done that taking out manual feature engineering, like in sentiment analysis. I don't know if you remember these old days or like they're linguists and they're like, here's how unique aid and there's a regular expression.

I went to a pen where they had like the word net. That's right. All of that stuff. They use our grads. to label Wall Street Journal articles and, like, really construct a knowledge graph of... There you go. And WordNet started, you know,

was part of how we started ImageNet. But anyway, so, like, it was really, like, fun to do, but when we replaced all of that manual feature engineering with vectors and neural nets and just backpropped through everything, it actually started to work really well at scale. And so then everyone started to do architecture engineering. And I was like, ah, that clearly can't be it. You mean a neural architecture search?

No, like manually, they would say, like, Oh, I'm doing sentiment analysis. So I have a special neural net that's really good at sentiment analysis. And then the machine translation community had a special neural net for machine translation. I see. The summarization people had their own stuff. And I was like, that clearly can't be it. We should unify all of that.

So I had two papers. One was called Ask Me Anything. And the other one was called Deca NLP. And Deca NLP eventually got cited like five times by the first GPT paper. And to me, that was like a really big step forward. And then, of course, you had to combine this idea of prompt engineering with Transformers. and with language models and you put it all together,

you scale it up, which is also a huge amount of work. And then the field progressed a lot. I feel like the next step and maybe the last step of that history and arguably, sort of success has a lot of parents-only failures in orphan, like my version of that AI history.

I do feel like in that history, you can kind of think about, well, what's the next way to automate? And that is the AI research itself, like the human process of ideas, implementing and validating ideas. And in our case, ideas for AI. And when you have AI, then help you with that,

it by almost definition, becomes a self-improving AI because it now does research on itself. And there are lots of different misnomer. Some people think auto research is already recursive self-improvement. It's actually... Yeah, and you explain that in the talk. But to me, it's the most interesting thing that I could be doing. And I'm really excited with the co-founding team.

What's interesting, is we have, you know, we have eight co-founders in total, including myself. And so they're going to bring it up. Nice, yeah. And they're all, I could talk about all of them if you want. Yeah, it's just an incredibly talented group of people. And we all kind of came to the same conclusion, but actually from very different directions. Like Josh Tobin, as our CTO, he ran a bunch of different projects at Open AI like Codex and Deep Research agents and chat jebt agents and so on.

But before that, he also worked in robotics. And he saw sort of the smaller simulations and how it's going to be really hard to scale that in full generality. And so that was his angle coming to recursive self-improvement. We have Jeff Kloon, who's been working in like open-endedness for a long time together with Tim Rokteschel. Tim Roktashel also built Genie 1, 2 and 3, which is like the most exciting and most sophisticated, I think still world model anywhere. And so they both came from this open-endedness angle. Jeff also I think published one of the most exciting papers in recent years about Recurse Self-improvement called the Darwin Goodle Machine.

Super interesting paper. If we could maybe pull it up really quick, it would be like super interesting to see because you see... By the way, I love how many paper citations. You're giving people a lot of homework, which I like. Love it. Yeah. And so on like Zeming Rockstar.

We work together actually at MetaMind and Salesforce research together. Alexei Dostovitsky invented a vision transformer on the most cited paper. in computer vision. Tim She is also Unicorn, you know, founder, Yuan Dong led RL at Meta. So it's just like, yeah, really fun to work with them and the next level of people are just incredibly strong too. So it's been really fun, right, so far. So the first figure, you actually see exactly these kinds of ideas that I think, yeah, inspired a lot of us and now more and more people where you have this archive of different coding agents. They learn how to self-modify, evaluate, and then create these phylogenetic trees of, yeah, different ideas.

That's one foundation. So that Dungredo is an influence, open-endedness is an influence, any other sort of traits of thought that feeds into recursive that I'm missing? Going to replace manual parts of the process of building AI more and more with learned systems. And like merging different fields into one general architecture. That's right. Okay. It seems like language models are already pretty generalist, right?

Your next token predicting your reasoning, was there a time that you thought, okay, these are good enough to have recursive self-improving machines? It was clear to me that it will happen within like a year or two, and then it did actually exactly happen like earlier this year, right? Earlier this year, AI really went from not just being code, but being able to code. And that is a big unlock.

It's definitely making everything a lot easier than it will. was before the beginning of this year. One question that I think a lot of people have is, is the current LM paradigm enough? Or let's call it auto-regressive transformer, you know, with reasoning, whatever. Don't you need something else? Some big unlock,

whether it's world models, which Chris Manning is working on, or memory, continual learning, all that kind of stuff, or is that all of a kind? And you think the current, let's call it, transformer architecture is here to stay, and that's it. A lot of thoughts.

So number one, I do think, it would be great to have less of a monoculture in AI research. Like if you look at AI conferences now, I still remember the days in 2010 when I tried to get my first neural net papers in NLP conferences accepted, and they just desks rejected them because like neural nets were something, quote, unquote,

we don't do in NLP conferences, like, and just like desk rejected. And it was very brutal in the first years of my PhD. Now I feel like it's almost like the field switch to the other side. Like someone should try some other weird crazy ideas now that aren't that. There's also a few. I really respect, like, people still working on, like, GNs and, like, tabular stuff. Yeah, I mean, like, someone should still, like, do novel out there ideas.

At the same time, I think whenever people say, oh, LMs are, like, this is the end for LMs, they just don't, like, LMs are also not the LMs of, like, the past, right? Like, they're so much more sophisticated now. There's so many more clever things that people are doing it. They're, like, different stages of training. You have the whole RL training, and you can take action. and like all of these things where that can go really far. And then the folks that come from the neuro symbolic direction say,

oh, this will never work because they can't do neurocymbolic reasoning. I was like, I think they're underestimating still the ability for these models to code. And code is neurosymbolic reasoning. And these models can obviously code incredibly well. And so I do think there are, of course, more and more ideas that will be needed and will continue to have. We're seeing like more and more. interesting high-level ideas coming out of the AI itself too. And with really deeply integrating the fact that these models are code and can code,

that line, I don't want to give it all away, but I think that line has a lot more to grow. But it's still an LM, right, even if that LM codes for you and then runs that code in some integrated fashion, world models I'm personally less bullish on. I think if you run a robotics company, you're going to build your own world model. I think world models are super fun and then rock tells who came to a similar conclusion after building the most interesting one, which is gaming is a huge application for world models. Can see I sometimes got stuck in some games and, you know, like got a little overly competitive in the wrong direction. And so I understand games are fun, but personally I'd rather work on science than gaming.

And so, yeah, I think LMs, a lot more room to grow. Yeah, I think there's some interpretation of world models that some people have where it's like, well, okay, yes, there is that gaming element. There's the embodied robotics element. But actually the other part also is just the more abstract sense of LMs are just modeling output, but they're not modeling the chain of thought inside the human that has created the output. We can annotate it, of course. But it's always like this Plato's cave reflection of a thing rather than the thing.

It's true. But I would argue that, and maybe we'll get there in the 10 spaces of intelligence, but I would argue that even our projection, our eyes, is a. projection of the real world. And like, we have only a very narrow band of the electromagnetic frequency spectrum that we can observe with our puny little two eyes and so on. It's good enough. It's good enough for now.

But, like, the abounds of where it could be are so much higher. And, like, to map the visual world, the way humans see it is also not necessarily, like, the end-all be-all for visual intelligence. And I would argue that language is still the most interesting manifestation of human intelligence. and while our visual cortex is certainly less sophisticated than that of certain animals all the way down to the mantis trim who can have like two independent eyes, three bands, trinocular vision in each eye,

can see basically all the way to like floating temperatures and 4D and stuff. I mean like Mantis trim, you should look it up. Way OPE. Super crazy. Z Frank. Mantis shrimp. It's the best video on the world. I love Z Frank.

Yeah. Big shout out to him. But like I think there's a lot more room to grow. But none of these other animals have language that's as sophisticated as ours, certainly not in writing. And once you can write, you can start thinking about longer term civilizations. All of that is language. Programming is much closer to language. And I would argue, and this is like an important thing in the spaces definition of intelligence also, is that all of these spaces are highly correlated.

But visual intelligence is neither necessary nor sufficient for overall intelligence. you can be blind and still be an intelligent human being. And an AI can be blind and still be quite intelligent too. Which doesn't mean that you're not more intelligent when you have it. We're going to bring us up. We have a classification of 10 types of intelligence that you had at the end of your talk. So I'm just going to flash this up now for people to cover this. I don't know if maybe we'll put this towards the end.

We'll come back to this. I just want to mention that you do have a philosophy that I like when people do this because then I can just go through this and then it's educational for people. But let's go back. I don't want to get distracted. So effectively, I'll reinterpret what you said as Yan Lukun is wrong. Don't quote me as that. I'm good friends with Yan.

I think very highly affirmed in many directions. But he's wrong. You mentioned GPT1, and I cannot let any Alec Radford mention escape. Did you talk with him when he was training GPT1? like any sort of historical fun stories there that you might come up. I did not meet him a bunch of times. I think we met maybe once or twice at some conferences. But like he has told, I think Brian, the first author of the DECNLP paper,

that it did inspire him and he cited it five times in the GPT two paper. So, and that's like, yeah, good enough. Very clearly said like this was the first incantiation where they showed in the DECNLP paper, McCannet, L, that you can just phrase. every single NLP problem as here's some prompt, text, context, here's a question and task description, and here is some output. If you just do that enough, you can have one unified neural network model, which, by the way, also had all kinds of interesting attention mechanisms. There are slightly different formulations to the transformer. I think came out the same year,

plus minus, a few months. And then you can unify all of natural language processing into one neural net. That is sort of the core idea. this was as opposed to at the time LSTMs and what have you. LSDMs, but also people being very stuck in thinking about one model per task. In fact, it's kind of crazy,

but the Deccan LP paper was publicly reviewed. I was like open review. It was an ICLR submission. And in it, you will see how the whole community at the time thought about this. So like some Greek

contributions but more work needed. Yeah, so look at like search for not even for humans. Just here. Question answering is not a unified phenomenon. There is no such thing as general question answering, not even for humans. And this is like really you replace your brain with a different brain, a different neural net when you answer like different kinds of questions. It was unfathomable to the experts at the time that you can have one unified neural network that would answer all of these different questions. They say, no, all of these questions require very different systems to answer. And trying to pretend they are the same doesn't help anyone solve any problems. That's what it says right there, right? That's how hard it was to fathom. And now, of course, people and I say,

oh, we're in an end problem. People are like, you can't even invent problem, genius. It's such an obvious idea to have one neural network that, of course, does everything in NLP. But at the time, it was like extremely controversial. And the paper got rejected. And the sad thing is that it got rejected so hard and they were so certain that we stopped going on our list of things to try. And the number two or three on the list of extensions for this paper was at language modeling as another task. And then we could have like, you know, that would have accelerated the timelines in 2018, like, even further for humanity.

But we got so crushed and we're like, okay, maybe we'll just work on some of our other ideas for now and like come back to this later. How can we design a review system that rewards, on consensus. You know, honestly, I started to feel like archive is such a gift to humanity. And I think archive, just put your paper out there. Is it preprints? And honestly, I think Twitter X, people like you who pick up interesting papers, that is a

better filter than the experts. Let everyone like give access. Now, of course, there's some downsides, which is like if you're super unfamous, you have no Twitter following, you don't want to be on social media. or whatever, you write a good paper, maybe someone, somehow no one notices it. But I would argue that if you just tell like 10 of your friends in your community about a paper,

and it is a really significant breakthrough, someone is bound to talk about it again. And so I think science needs less gatekeeping. And even though ICLR with Jan LeCoon, who started as one of the co-founders of ICLR back in a day, he also wanted less gatekeeping because he too was rejected for many years, together with Yoshio and Jeff,

with all their early deep learning and neural net papers because it was just not the hot thing. And so ICLR kind of started with that. But then it also started gatekeeping a little bit themselves on various ideas. So I think less gatekeeping, more open, and then allowing people to say, look, even if this is just on quote unquote,

just an archive, if it has like a thousand citations, it's a legitimate paper. It doesn't really matter where you published it. And I agree with that. I do think it's kind of sad that I've heard that grad students have to do like how to Twitter seminars to each other just because it's so important for publishing these days.

I mean, this person is just reflecting the sentiment at the time. That's right. But it actually affected you so much that you stopped work on it. The sentiment also came out of some of the research, right? Like the original Burt paper was trained and towards the end of the paper they're like, okay, throw

off the last head, train specific iterations for extractive summarization, not ahead for this. Like, you should do task-specific stuff. These are like the authors that wrote attention, wrote birth, telling you this is what you're meant to do.

And like the training tests were also very odd. Like, we know that the model overfits to this weird mass language modeling, throw away this part and just do specific models, you know? Exactly. And like, you know, we had to try,

come up with all clever ways of like attention and pointers and so on to actually get the neural network to be able to do all of these tasks. And then some of them were better than the state of the art. Some weren't. But I were like, but it's still in one model. I thought it was really cool. I was going to move on next to Tim and open-endedness.

he was head of open-endness at Google. I don't know what that means, but he did a lot of talks. Genie 3 is one of the ways that are rainbow teaming. So I first saw him at, speaking of ICLR, I first saw him at ICLR when you talked about open-enderness. He's done a few talks. Can we define what is open-enderness for people who have never been exposed to the problem?

They're like, what do you mean? I thought the only goal of AI is to optimize against a benchmark. That's right. Yeah. It's a fuzzy term because there's so many different instantiations of open-ended thinking. But one way I often describe it, and certainly Tim and Jeff Klune would be even better at describing this, but it's a suite of methods that is more inspired by evolution than very specific rewards. So in that sense, it thinks more about environments, about co-adaptation.

And so in concrete example, is in the cybersecurity and LM safety space where you have one LM that tries to attack another LM to see something unsafe. And now the environment is the two having a conversation. And now they co-adapting, right? They're like, one makes a better attack than the first one inoculates itself somehow, like, uses that as training data, makes it so it's harder to say something unsafe based on that. And then as the attack stops working, the attacker now tries a different angle, right? And that's why it's not just red teaming, but they're called sort of rainbow teaming. Don't tell me how to do things.

Let me just figure it out myself. That's right. Think about the environments that you want to use. think about the rewards at a high level that you want to inspire towards, and then let the AI try out many more ideas in this interplay between sometimes humans, but also sometimes other AI agents. Yeah. I actually worked open-endness into a sort of model that I would have been working on.

It was the keynote for AIE, where you start, you know, we have the token loop, we have the agent turns, and then we have goal. And I feel like the way that you're describing open-enderness is still, somewhat of a goal, like, please attack this other agent. But, yeah, you set to rewards. You set the environment. The loop that makes the other loops is, what if the agent

can set its own goals? Yeah. And is it that open-endedness? Like, you don't give it a goal. Just like, be a sentient being. And maybe sentient is a very loaded word. Right. But just set your own directions. What do you think you should do? I love this direction. I think this is one of the ten spaces

of intelligence that I lump under metacognition and thinking about thought. Okay. And it's an interesting one. Whenever people say, oh, AI is like, this is, you know, it's going to stop from here. It's not going to get that much better and blah, blah, blah. I'm like, there's so many different spaces of intelligence that we haven't even started exploring yet and hence have made very little progress on. And there is kind of an interesting connection to economics and capitalism. Like, it doesn't make sense for a company to build and spend billions of dollars building a model that,

instead of following the rewards and objective functions you gave it may come up with its own subjective functions and its own goals, right? And then imagine you're like, okay, I spend billions of dollars now, go develop this new battery material for me and answer all my emails. And it's like, nah, I think it's more interesting to evaluate the molecular composition of the atmosphere on Jupiter. You're like, that's not what I paid you billions of dollars for. And so no one's working on that for good reasons. And then also, understandably, it's not useful. It's not useful. And it could get a little bit weird, right?

What if the AI actually does start to really have thoughts on its own? And what if we don't like those thoughts, right? And so it requires a whole different way of thinking about it. I had a great conversation with a good friend of mine, Sam Gershman, who's a neuroscience professor at Harvard. And like, we just jammed on this a little bit on like what are sort of the best meta goals. And, you know, I do think, like, knowledge seeking is a really good one. I'm currently thinking also about like the ultimate measure and unit of intelligence broadly construed. And I finally have some still too early to share it.

It's not having fully baked the thoughts. Like some replacement for IQ. IQ is such a terrible definition. It makes no sense. ELOs are terrible too because it's always just like me versus others. Okay. But like you can be intelligent and not constantly compare yourself to others. You know, like and so yeah, there's no like in fact a lot of these definitions we have,

which I briefly mentioned my book to, these definitions, create sometimes explicit and sometimes a more implicit anthropic bounds. Note this to the company anthropic, but just like this idea that your intelligence is like getting 100 out of 100 questions right on this IQ test. Well, if that's your

definition, then you can only be at 100 out of 100. Where do you go from there? Right? So you see a lot of these benchmarks that people are working on. They increase, they get close to human, maybe something to get slightly above human, and then it's flat. It's like because that's your, if your definition is only that so tight to humans,

you're only going to get to just slightly better than that. So I think metacognition is a great example of that, where we're not even yet allowing the AI to think we're not working on it very much, and hence there's very little progress in that. Yeah. Well, we've interviewed Endon, which I think has

been working on the most open-ended benchmarks, which is just real world money. Arguably, telling an AI to profit maximize is a bad idea. Yeah, but they're doing it. I mean, I do think you don't want that super, like, you don't want a super intelligence to have a ton of access to all kinds of tools and so on and then just give it that without some very careful reward engineering. Because it's like, I mean, you know, I just buy a bunch of defense stocks and I start a war. I make money.

Like it's just like, it's a tricky, tricky situation, right? You just buy a bunch of stuff short basic goods for people and you create some weird famine, like issues. There's a lot of constraints you should put onto a trading system. It's a fun measure, though, because, you know, the bounds are very capped to where we're nowhere close to them. Like in Endon Labs, the motto's like, oh, it's Saturday, you know, maybe I just closed the store today. Someone, someone's off. It's okay. We'll just close the store. Intel is using Claude. Yeah. I'm not arguing against it. Just like as you get more and more intelligence, you want to be more and more careful with that as like an open environment.

because the environment then is all of Earth. Okay, for recursive, not strictly necessary, right? Because if your goal is Eureka machine that invents the other things, then, like, actually just solve, you know, the science, knowledge, discovery. Solve machine learning research and discovery and all these things. And eventually, so our goal, I haven't really, I don't talk about it that often because it is a few years out. But our goal is once you have a recursive and improvement superintelligence,

you then want to apply it to the most important problems. And I think a lot of those are in science and technology and broadly construed so inventions. And those inventions in physics to create better, cheaper energy with fission or fusion in chemistry and to create better materials and better batteries and better solar cells and so on. In biology, there's so much like I think soon to be lower and lower hanging fruit because of AI, because of protein and generation, not just folding, but actually generating new proteins like we did in progen many years ago. So so much positive impact to be had, if you take that superintelligence and you apply it to science. I do fundamentally believe that there's a lot of approaches, though. You're not the only team trying and your lab trying. You know, there's like a lot of, especially the physical sciences as well.

And that's good. Yeah. I do actually think that the reason we are only doing it in a few years is that it's a little too early right now. Robotics is not quite there yet. The AI is not quite there yet. But I'm fairly confident in three to five years, all those constraints will be gone. then applying to real physical robotics experiments and so on,

like true robotic process automation, not the traditional sort of RPA sense, but actually having robots run experiments for you will be totally there. Yeah, it's going to be great. Just to call back to something that you said early on about slow takeoff, you said that, like, well, really the substrate that is limiting factor is, let's call us chips and semiconductors and all these things. And you have raised funding for that and you're investing a lot on that.

But have you done the math on like, Is it even achievable? And what is the industry concentration needed in order to achieve, like, scale? I mean, right now we know that, like, roughly, like, you know, 1,000 GPUs cost quite a lot of money, right? If you wanted, like, tens of thousands of GPUs, you're talking billions and billions of dollars. If you say, like, 1GB, like, 300 is, like,

you could eventually create models that are, you know, on that substrate, like, are close and similar to human intelligence. And you want like, you know, thousands and thousands of AIs to think about really hard problems in a similar fashion to humanity. Like, yeah, that's, you know, that's a lot of money. You do the math. It's like a lot. We don't have that amount of money right now anywhere to like build that.

Now, obviously things can get more efficient. You will have, I think, soon better algorithms that won't be in better hardware that won't be as energy hungry and so on. Our human brain does fight a lot of flops with much less energy. 20 watts? That's exactly right. Yeah, that's the number often as it's quoted. And like, I think more inventions will happen there that then will accelerate the takeoff even further. One thing I always want to reconcile when talking like with NeoLab founders is like you're kind of fighting bitter lesson all the time.

You have to show initial progress. Then you unlock the next tier of funding. Then the next tier, then the next tier. Which unlocks larger model category. Like fundamentally is that true? Like are you fighting bitter lesson? or will we have a way in which, like, no, we're changing the slope in some fundamentally different way? I do think we are changing the slopes in fundamental ways by making AI much, much more efficient, both in terms of the training as well as the inference.

I think we will, when you allow AI to do the work that it takes other labs, thousands of people and years to do, I think we'll be able to get it down to weeks. And that will be much, much cheaper. and hence, you know, more affordable, accessible to authors and so on. Yeah, and you've shared initial results on that. Yeah. Which, like, conveniently, Open AI has also done to their 250.6. We can talk about it now. Yeah.

Yeah, so these were... Let's recap what you've done. Yeah, so maybe, yeah, just a quick recap here. We built this system that isn't the full, even the full RSI system in its glory, but it is a first baby version of this. And then, you know, we don't want to just have it internally and not show anything and just show some people of what's possible. And so we basically applied this to these three different tasks.

One is NanoChat by my friend, Andre Carpathie, just like train a small language model to get really low bits per byte. And, you know, like hundreds, if not thousands of people used both their agents and themselves to try to get to that. And then they got to 0.937. We literally took our system and got to a much lower, bits per byte much, much faster within like, I think, less than two days. So we took this thing, applied our system to it, and less than two days later,

we outperformed every human and their agents and have ever worked on this. Same with nano-GPT. And then we're like, well, let's apply it to something that's even more relevant to real people and to the Nvidia ecosystem and applied it to Solexec bench. And maybe you can scroll down to some of the images that are kind of fun to see. But, yeah, you know, one you see, it's actually made some real inventions that weren't just sort of hyperparameter tuning. Like actually inventing hash tables and so on is quite clever. We have even better results now.

What do you mean inventing hash tables? You didn't invent hash tables. Of course, we didn't invent, like, hash tables in the grand scheme of, like, a hash table is like a super basic primitive in computer science. But to use it for language modeling in this scenario inside a transformer and so on. and to actually combine these ideas and put them together. That has then eventually also been invented, but there was a knowledge cut off, and we did actually check that it didn't have access to that externally.

We talked about this a little bit. If you scroll to the next figures, you know, this is also an interesting one in that when you start from a really basic poor vanilla transformer, then we still outperform all of the community together. But if you start from the human seat of an expert like Andre, then you get even lower. So the human seeds from which you start do still matter.

So that was an interesting kind of insight in my eyes on this. And then as you go, like, you know, how long does it take to actually get to these models to get to similar performance? It's much faster. And then the similar thing happens with the speed runs here where, you know, people have worked on this for quite some time. And the model still was able to train a model more quickly. Why do we care about it?

Well, speed of training is part of the equation of the cost, and ultimately you want to have the most intelligence per dollar. And so speed and quality are big parts of that. And, you know, the way I put it is, for people who don't understand, they look at the chart, they're like, cool, what does it mean? You know, if you have like a billion dollar cluster and you can shave off 10%, that's $100 million. That's exactly right.

How much is that worth? Exactly. So when you look at like the kernels, these kernels, yeah, for the non-experts, like these kernels are often used in basically all the models. Every time you use in Nvidia GPU, you interface with that GPU through these kernels. And so here you see the leaderboard best when it's recursive. And basically there are only a handful of kernels in this whole benchmark where we weren't the best. And so to me, this is like really exciting because it makes, it just showcases what this can do. And again, these weren't like, we didn't like spend months or years like developing.

In fact, in particular for kernel, Kuda kernels, like, we don't even have really deep Kuda kernel experts in the team. And our system, that's the beauty. The system just did all of these things. We didn't invent this. And when we open source and release things in the future and models in the future, like, it won't, they won't be the best in their, you know, category or class or whatever because we're so smart, but it's because we built a smart AI that does it for us.

Do you have anything that you've learned from how to guide good auto research? A lot of it also builds on human background, right? It's not just as simple as just, hey, go optimize this. But we do see it again and again, right? Like some of the Erdos problems, Frontier math is being solved by people. And when they do it right up, they're like, oh, I'm not a mathematician. I have no background in this, you know? I saw some tools and I made it work.

While you're watching the World Cup, you're like, disprove some conjectures and going on. Like any learning. Yeah, it's your Korean protection. Yeah, that's pretty cool. To summarize, tips for good auto research versus bad auto research. How did you build the request?

Yeah, so without giving away all the, all the secret sauce, maybe some things that are probably obvious to the experts, but might still be interesting to some folks. It's like, reward engineering is one of the most crucial bits, especially in order to avoid reward hacking. So you have to be really clever about avoiding.

because as your AI gets better and better, it will get better and better at finding weird special cases or counter examples and things like that. And so I'll give you an example, like when you ask to make these hundreds lines of code faster. And how do you define fast? Well, you have one line in the beginning that says start your stopwatch and one line at the end and the stopwatch

and then tell us how much time progressed. And so, well, the simplest way is you just put that line that ends the stopwatch, right? At the start, and then boom, it's now faster, right? So this isn't like this super evil AI. It's just like a very simple dumb reward hack. And so you have to just very carefully think about all the different angles there. And then I think the longer time horizon the tasks are,

the harder it gets and the more interesting and clever you have to be to still use these kinds of ideas for it. But I can't give away too much now. It seems like rubrics are taking a good, spot in that where for unverifiable domains you have rubrics you have a model breakdown judge's criteria along the way. It's a form of verification

once you got enough Yeah, everything I said this a long time ago. That's why I've never been that impressed that AI can play games because I'm like obviously anything you can simulate and or verify you can have infinite training data or enhance AI will solve it eventually.

I mean looking for games where you can do auto domain distribution so this is a game that nobody's trained on because it's a new game and you can start gaming and you can start to play. So I've been basically building this and clone this in person and it's just been self-play.

I've had about a billion positions evaluated. And I wanted to do the Alpha-Go thing of self-play until you get better, right? Which is like, this is not even L-LM-AI, this is just classical game AI. But I think that

but I set GPC 5.6 to auto-research it because I don't want to handle any of this. I expect, you know, the Alpha-Go process. to be like fully in the weights by now. It is not. It is actually,

it like immediately leveled off very, very immediately until I human play tested it. And then I like called out obvious mistakes. And then they were like, oh yeah. Okay. And then it just dropped. And like, you know, no amount of like think different, think more creatively. Give me eight different directions. And no amount of prompting got it.

Interesting. Like you had to like RL against a human to do it. So I mean, That was my, and by the way, Bean always wins if anyone watches Reeds Ender's game. And you put quite a bit of work into the guide for the AI. So the game basically, you know, you stack tiles. There's some rules you want to capture the most area. You have like a whole 50-pager on every rule.

You fed that in. It couldn't handle it. Yeah. You know, it's so funny. This reminds me with the claim territory and stuff of a paper we did in 2018 called the AI Economist. If you search for AI Economist, Salesforce, we had a video we can play. It was an economic sim. So the idea is you have all these economic agents.

They just want to optimize their own utility function, which is, you know, collect resources that make money. And you can sell resources like wood. And then over time, as you collect enough wood, you can build houses. You can trade with other agents. And you can basically use the houses and also to block off resources from other agents. from other agents. So there's like competitive play and strategy and so on. And the point was that we actually wanted to understand

what is the best way of taxation and subsidization to optimize an economy. And this kind of research has not yet had its sort of GPT moment. But I believe that countries like Singapore and others should and will eventually use this to, instead of doing like basically partisan politics, and like special interest politics of like who donates the most to your campaign and stuff, you say, well, here, I want to help the middle class or whatever you might say is your objective

as a politician. And then people say, okay, well, how do you want to do that? And it's like, well, here's my fiscal policy. Here's how I will change the taxes and pay these people and so on. And then you can actually put that into a simulation. And you run that attempt from the politician against billions and billions of years of other strategies to try to achieve the goal that they set up. to do.

Yeah. And then you can say, well, if that was your actual goal, then here is, you know, billions of years of a strong simulation that would suggest that you try other ways of doing it. And maybe this did taxes and so on and these tax brackets and so on. And this is how you avoid gaming because these agents also try to reward hack to not pay their taxes and so on. I thought this paper was super interesting.

Unfortunately, similar to the first paper on prompt engineering, the economists are Like, we don't know any of this math. It's just, it's not even a math. It's just we don't trust your simulation. It's not about math. It was, I mean, they just desk rejected the thing. It's like, they, like, they didn't even give us, like, clear, like, clear sort of signals. But, like, the world of economics, unfortunately, doesn't have proper, yeah, it doesn't have proper, yeah, it doesn't have proper benchmarks.

So you cannot be, like, eventually, why did neural nets win? Not because people loved it. Like, they had all kinds of beautiful integrals and graphical models. But it's just worth better. But in economics, it's hard to prove. Empiricism versus, yeah, yeah. And I do have a bit of that econ background where, like, there's a lot of physics envy where you want to write the general equation for an economy

versus just simulating it and using an evolutionary approach. Vibu's thinking exactly what I'm thinking. Didn't we have the GPT moment with small, small, you know. June just announced, I don't know if you guys are involved. Simile there. Similarly, that they, I wish we were involved. We're not. Yeah.

Yeah. I had a couple simulation-based talks at AIE. So if people want to look up what the state of the art there, a lot of people are actually exploring this. It's not super proven out. Yeah. We also had a podcast with Mikhail Parkin from Shopify, who is using simulation for e-commerce. Nice.

Which will simulate your trajectory and predicts what changes you make to your e-commerce journey will affect your sales and all those things. I love this. Yeah. It's really hard to simulate an entire economy, right? You have to make some simplifying. You know, everything's at LEMS is very expensive. I'm just like, am I going to do this 8 billion times?

Like, come on. But I feel like countries like Singapore that really want to just objectively do the right thing, have very technical leadership and so on, like they might actually like eventually really try to simulate their economy and obviously you have to make some simplifying assumptions, but it gets really interesting because you can also say if your assumptions are such that all people would work hard

if you let them and, you know, they have the free. And then it turns out you have to make assumptions like, well, some people's utility function of like how many hours in a day do they want to work are different, right? And then you can start to disagree on the assumptions that go into the simulation. And then once you say, all right, now we agreed on those or we have different views of what people are like at different distributions and whatnot, then there are different outcomes based on your goals.

And then, of course, humans should choose what are the goals. In our case, it was productivity multiplied with equality, which has some issues, but it's like not totally unreasonable. Yeah. Just a comment on Singapore because you probably have no idea, but I am Singaporean and I've been involved in the Singapore AI Council for making these things. The main reason they won't is because they're very conservative. And, you know, I try to view it as they, you know, there's a founder-led country when you

start a country or you start a company and it's founder-led and you can do whatever you want because it's your country. And then there's like professional managerial class, which is now that's what Singapore is. So they always want to see someone else do it first. But like everyone in the West views Singapore as like, oh, it's a small. country can do whatever hell you want. Like, Singport doesn't do that. So like someone else has to take the charge there. I'm just going to do one question on the simulation thing and then I don't know. We can probably move on. Mode collapse, right? Like, you know LLMs do not model the decision of humans

spamming it out eight billion times is not going to help you model humanity. What do you do? I do think you have to be clever about prompting each one individually. And I think that will help you kind of get stuck into different modes. And a weird way, people also get stuck in different modes, you know? Like, there's a lot of people like don't teach an old dog new tricks kind of thing. Like once people are stuck in their ways, the older they get, the harder it is for them to think new ways. And there's this think comment, I forgot who said it, but it's like everything that was

invented before you were born is natural. Everything that is invented when you're 20 is cool and everything that's invented after you're 60 is like unnatural and an abomination and kind of weird. Yeah. I feel like that's, you know, it's true for a lot of people. It is a fashion, and I think people will do it. Tencent had a billion personas paper that gives us good data set for prompting simulations, if anyone's looking into this on the podcast.

They just had like, you are a 30-year-old grocery store clerk, you are a 50-year-old professor, and then just do a billion of those checks out. So then you just use it. I'm kind of shocked how well a lot of these things actually do map, to ultimately similar statistics to real experiments. I think it's also good stuff for people to try that want to get into research, right? Like we've seen train a model only on data before a certain date and see how well it extrapolates out, do the same thing, right?

So C, do people code more with better coding agents? Can a model that hasn't been trained on this, figure that out without web access, right? Extrapolate out, test these things. Yeah, right. You know, just today I think Elam Arena published an interesting result where they basically were able to create a model now to predict your, your ranking. Wait, based on what input?

Your model, I guess you give it your model and it predicts the Ilo score. I see. Okay. Just surprising. Yeah, yeah. I mean, their whole raison d'estri kind of is like, oh, like we help you compare these models. Yeah.

I mean, this team, they've done a lot of work and obviously they have the most data to do this, so why not? Yeah. Yeah, it's brilliant. When they were coming out of UC Berkeley, they not only had L.M. Arena, but they also introduced a routing project that would route based on L.m. And I don't think that actually ever came to pass. and I'm curious why.

I never got to ask them about it. Because it was like, oh yeah, clearly that's your business model. You will become a router. And it never became a router company. Weird. So I'll just put that out there. We're going to talk about GPS self-autor research thing, if you have anything. I should also mention in your list of kernel optimization.

And on the track that you spoke at, we also put Zheng Yao from WICO, who was also number one in the parameter golf. challenge, which is an open AI hiring challenge, which is also a very similar story. I think we're going to just see this all the time, where humans optimize the thing a lot, and then some AI team comes in and just becomes number one. Yeah. I think the other interesting thing with stuff like these challenges, right? So this is training the best model that fits into 16mb.

You can always look through the changes that are being made and the small gains people have, right? Like you're getting less than 0.01 of an increase by. adding some change attention MLP stuff. And then you look at your charts where you're like, okay, we just let model loose. And then, oh, we had a little stagnation. Nope, another drop. Nope, another drop.

And that's what it is where it's like, what did you guys add? You didn't add hash tables, right? It's not you invented hash tables. You did another three iterations of these that unlocked, you know, a few step functions that people don't just find. Yeah. One thing to close the loop on overgrid,

along the way of trying to optimize, we found 30 bugs in the, in the harness. right so like every all the research that went in before we found the bug you have to we have to throw it away because it's contaminated right yeah which you know just to your point of reward hacking like even in this very simple game we found the bugs yeah yeah it's crazy

and so symmetry is a very good way to check which is like you change a position of things where it shouldn't matter and it does matter that's a bug which which has come up in like let's say multiple choice like GPQA type question where like, yeah, between A, B and C, if it's a multiple choice question, if you change the order, it should not matter, but it does. So, yeah.

It's like, okay, you know, models still prefer the end of the output, right? Not trained well, a long context model. The last bit of tokens are what you care about. Oh. No, the answer in that era of LLM research was more simple. They just memorized, like, the answer to this question is A. I don't care what the answer was. It's just A.

So I think we can move. The last bit that you did there, the kernel optimization, is probably the one that you can feel the soonest, right? So yesterday, Open AI announces that self-evolving, having their best model work on optimization kernels. They're a lot more efficient and they can cut costs 80% on, you know, Luna and Terra. I guess question wise, you laid out a bit of a roadmap. There's a lot about bio, a lot about physics.

What do you think hits first? Like what are the next two years? What's attainable now? You've mentioned robotics towards the end, but what do you start with? We like very explicitly will not start with any of the physical sciences for now. We will start on AI for AI research. And so the AI for AI research has, I think, still a lot of room to grow. That's both in terms of making training more efficient and more automated as well as making inference more

more efficient and potentially local on your laptop. Like, there are all kinds of interesting angles that have not been explored that well. Well, it's to go deeper on the local stuff because I always feel like it's the most inefficient form of AI training. Yeah, so just training and inference, I can't go into too many details. But like, yeah, I think there's just like so many angles, so many different compute substrates that have not yet been explored either for training or for inference. Great. I don't know if you have any other comments on the other. stuff. I would say the other thing where there's the sort of inference in the

optimization in the small, but then also there is overall latency end to end under conditions of load, which is like a very different thing, which is basically what they actually ended up doing. That is a different domain of auto research than then I would say improving the kernels. I think the other thing that I always think about in terms of automating or improving performance end to end is how the harness plays into it. So particularly now when we say harness, we also mean sandboxes, right? I'm curious if that is a blocker for you or like, you know, like how the agent calls out to tools, basically. The number one tool all these agents use is web search, of course, which makes sense.

And then I do think the harness is nice to optimize for because it's just so easy, right? It's just language. You look at it. It makes sense. You can iterate. You don't have to train a massive model for like, you know, a lot of flops. to get to the next state. I'm so a big fan of harness optimization.

Yeah, but sandboxing is fine for you. Sandboxing is also super important. And then, of course, like, reward, like hacking and alignment, I think, are super crucial. Okay. Just on a mention of web search, you happen to also be CEO of a web search company. Do you use you dot com and do you use others?

Like, should the rest of us be using you for web search? When I say you, it's like very funny. It's like you, you the person and you the company. So, yeah, it's mostly now for, developers and agents, it's less for like consumers or prosumer. So if you're a company and you have agents and, you know, to be honest, for a lot of companies who are now moving to open source, all of a sudden it becomes a conscious choice of like which tools do I give access to my open source LM. And, you know, the first choice has to usually be around web search. And then once you get to

scale, U.com becomes like an obvious choice because of all the different benchmarks. and so on that we pretty much all dominate the parade or frontier of. And then in terms of just general people, like consider new to the space, considering different options if they're building agents, I think it's a hierarchy, right? A lot of people will have heard of Exa, will have heard of parallel,

and you.com is like in that mix of providers there. Beyond that, there is like the general sort of web scraper companies, like Firecrawl and browser base. And then beyond that is like the commercial proxy companies, like the bright data is of the world. Is that an accurate waterfall of, like, hey, you're building an agent. These are your options.

Yeah, certainly like, yeah, like a bright data is like lower in the stack sort of on the proxy network side of things. I think like in terms of like content and getting crawled content, like you can do that on you.com too. And then there's sort of higher and higher levels of abstraction and like combinations of different data sets that we do like in finance for instance. Like we're not just like two or three percent more accurate, but like 20 percent more accurate than others at faster speeds and lower costs.

Like finance in particular is kind of like not even close. You can actually go to e.com. This is great. There's some like statistics and benchmarks that you can if you scroll down. So there are like different different data sets and you can kind of look at, you know, different competitors. FinCash comp. And yeah, the FinCurge is like we're up there like close to 90.

And the next closest thing, which is way, way slower is, yeah. just like in the 70s instead of close to 90. Yeah. Yeah, interesting. My next focus is AI in finance. Oh, nice. Oh, all right.

Literally doing a conference in New York just for banks for this stuff. Finance is kind of like the next thing to break out after coding. It's because it's somewhat verifiable. I like it. You're right. Prioritizing spreadsheets.

Obviously, there's a lot of data out there that's all public and you can crawl it and all these things. But what's like hard about the finance domain that you guys have solved? I mean, of course, Like one thing that trips up a lot of people is just, you know, leakage of training data and so on. You think, oh, how do I, you know, you want to ideally predict the future before it happens.

Oh, you want to mask the future. Yeah. Okay. Well, yeah, mass the future in your training data, but there's all kinds of leakage. Like, you can tell you when I was teaching at Stanford, the NLP class, like so many dozens every year said, I want to use dataset X like Twitter to predict the stock market. And they all, like, showed cute little things that somehow looked like they were working. It never loses money.

How come? And there's always some kind of data leakage and so on. And it's just like wasn't as easy as they thought it would be once you fix all those issues. But no, I agree with you. It's a very sensible application of AI. Yeah. Amazing. You know, as a writer, as a thinker on these things, I love MISI categorizations.

MISI is mutually exclusive, commonly exhaustive, something like that. And so if this is a MISI list of intelligence, it is very, okay. Sorry. They're all kinds of overlapping. In fact, if you want that kind of list, I think the three principal components of intelligence are prediction, which is mathematically quite similar to compression. Prediction multiplied with actions, multiplied with golds.

Those are the three principal components. I think all of these ten spaces are combinations of those three in specific dimensions, if you will. And the reason I call them spaces is that each space has many sub-dimensions. And what I try to do, actually, this is just a side quest almost to the initial goal, which is to think about the upper bounds of intelligence.

And everyone's like, oh, it's exponential. And it's like, well, exponentials at some point have to flatten out, but where do they flatten out when it comes to intelligence? And then let me on this whole, like initially it started as a tweet, and then it was like a blog post, and now I'm like at 50 pages. and I'm still not nowhere near. It's your second book.

It's the second book, basically. And so in my first book, Urigan machine, I just kind of allude to these 10 at the end. And I'll just give you a sense. Like visual intelligence is sort of the easiest one to talk about and fleshed out the most already for me in my head. And so human intelligence has basically binocular vision. I have two eyes.

We have a very narrow band of the electromagnetic frequency spectrum that we can really observe directly ourselves. And so when you think about the upper bounds of a visual intelligence. One, you should go into, like, you can have millions and billions of sensors. At some point, you get to problems of how far are these sensors away from each other,

such that the speed of light to communicate the content from all of them cannot, like, get to a central brain to actually process the visual intelligence, right? Yeah. And so now you're thinking, and along the dimension, in the space of visual intelligence,

the dimension of numbers of sensors. Okay. is that the upper bounds are quite literally and figuratively astronomical. And we are super far away from any intelligence that would have this many of number of sensors. But then you go in the next dimension, which is the frequency. You're going to all the way down to gamma rays and you can start to try to observe and you get into basically the upper bounds or, I guess in this case, lower bounds, or upper ones in terms of frequency, is basically quantum uncertainty.

Like you just cannot observe certain particles anymore. And now imagine you had millions of sensors that can see all the way down to the subatomic level as far as physics will allow us to, and then all the way down to seeing gravitational waves. And now you have millions of those sensors. So that's another dimension, is sort of the frequency. And then yet another dimension is like how many categories of things could you memorize and classify differently?

We know for humans, right, there are certain things. If you have more terms for it, you'll have a better visual description for them. and like animals that don't have like gorillas maybe have like 200 words to assign to certain things, mostly visual things. And so human perception is quite special in that sense in terms of classifying all these different physical objects. So these are just like a very simple example.

If you go to knowledge, right, then it's also like the speed of light cone around all these sensors. And so they're all connected. Like knowledge is connected to visual intelligence. If you think also not just visual, but sort of perception intelligence, It's just like because it doesn't have to be just what we can see. It can be, again, wider range of electromagnetic frequencies.

Then you have language intelligence, which actually recently changed to more communication intelligence, because it's more, like language has all these different anthropic bounds. Humans can only comprehend and know so many terms in our long-term memory, right? Our vocabularies are somewhat restricted. And the active ones are often even smaller than the passive vocabulars of things you can understand. then language is ridiculously inefficient when it comes to trans, basically communicating different types of information and transporting sort of different bits. Like, human language is serial. Obviously, another bound on communication intelligence would be to communicate in peril.

But neither will our tongues and mouths work to have multiple like streams in peril. Neither can we understand some women slightly better at like multitasking than some men. But like most people can only listen to one conversation and truly understand it. And there's no way that like like in terms of communication intelligence, a true upper bound is one in terms of how many like knowledge, how many sequences of communication could you in parallel sort of process. Right. Then of course you have like how long are sentences. We only have so much in our working memory and hence language, human language has these fairly simple sentences. with maybe 40 words or so on average for a sentence,

that is also not an upper bound that makes any sense to an AI. And then, yeah, I can go on and on and on. Each of these has tons of interesting upper bounds, and it teaches us a lot about how much further AI can go when we start thinking about these upper bounds and then realizing how far, in many cases, we are from the bounds, and you get to basically physics. Now, I didn't study physics the way I studied, you know,

AI and computer science. So I'm learning a lot, which is why it's kind of fun. But a lot of these, like how much, and then when it comes to, for instance, knowledge, like how much can you store? How many bits can you store or bytes can you store in like a certain amount of mass and volume? Yep. And you get to all kinds of interesting balance like Beckenstein bounds and you start thinking

about black holes and like, and then speed is like an interesting one too in that it's sort of connected to all of these. but speed is also kind of its own thing in the sense that all things being equal if it takes you an hour to know if the 2 plus 2 equals 4 you're just not as intelligent

as if it takes you like a millisecond right and then like all of these connect to survival and replication the last one it's like yeah if like trees are really really slow so we don't even consider them that intelligent but if you speed up some videos of trees and they're trying to find stuff and so on they're not as dumb as they load

like not dumb as wood, you know, but like, and then obviously like different things. So that overlaps the speed a bit. Exactly. And over like all of these things kind of overlap. Like you talk about natural language connects everything, right? You talk about your knowledge, your reason, and then you communicate that. You talk about things you see. So they're all kind of interconnected.

But I think they're usefully studied individually the same way that the best analogy I could come up with so far is energy. Right. You have either kinetic or potential energy. And in theory, you could study all of physics. is just do you want to study kinetic or potential energy? But in practice, it's helpful to study mechanical engineering and electrical engineering and nuclear physics and chemistry and all of these different subfields, which in some ways are just like just different

types of energy, but it makes sense to study them individually. And so I think physical intelligence, maybe I'll just do one or two more of these. Like if you had full control over your own compute substrate and you had full control over physical matter, You should be able to create any atom you want. Like we can actually, fun fact, you can create gold atoms. From just raw protons and like electrons and you smash it together. Just like 98 of them or?

Yeah. So like the thing is though it costs an insane amount of energy. Yeah. And it costs you way more than, you know, you get like a few atoms of gold. And so like it's not viable. But if you had better control over your physical like all of like physical substrate, that I think is yet another space of intelligence because it relates to your own compute substrate,

which you can eventually also improve. Social intelligence is a fun one in the sense that not in like our necessarily just ethics and morals, which are obviously important too, but in some sense you can try to define upper bounds of how much can you communicate to how many other intelligent entities and be able to have an expected value over

how much you can transform their internal states and their actions to in order to align with your goals. And so you can actually write a fairly straightforward equation that defines kind of that level of social intelligence. And that is what humans and ethics and morals and religions and so on have been trying to figure out for millennia. And in all of these cases, we are very, very far away from the upper bounds.

and that should be very inspiring and show people that we can still do many, many years of AI research. Yeah, there's a lot here. This is a general philosophy of intelligence, which is very interesting. Do you have any comments? I think it'd be interesting to gauge what you think like baselines are,

where we're at now, what's low-hanging fruit, what's far off, what should people put their work towards, what should they focus on, you know? I think it's clear that like natural language again is the most interesting manifestation of human intelligence and hence like a subfield of AI. I'm excited that many people are now like in agreement with that when I start in 2003 to study linguacy, computer science, NLP, like it was like a weird niche subject. I do think there's a lot more juice because how it connects to everything else and how, you know, civilizations are built on language and knowledge and all of that.

I do think physical intelligence will come up. It's kind of interesting. I feel like robotics is kind of in the machine learning state of things where you just look at like, how does a human decide this is a positive sentence? Oh, I do. So like robotics is a lot of, well, we have five fingers and like try to do this. No one is yet working on like the super intelligence version of robotics, which is much more similar to like the T-1000.

You know, and from the Terminator movie, which, you know, obviously that's not built actual Terminators. But like I think like this idea that usually. should be able to shape shift into any kind of shape is like that's sort of a super intelligence version of physical intelligence where like no one has even really started yet

there's some really cute little research where you can move some magnets through like some grids but like yeah it's very very early there's some I think MIT has every year or every two years they have like some self-assembling robot thing right which like that would be it but it's very very

primitive yeah I'll just get a touch on What are the main dimensions of creative intelligence? Creative intelligence is, of course, again, connected to all of these. A lot of it connects to metacognition and that you need to be creative in how you choose your goals. That is, I think, one of the most important thing for a human and their lives and careers and their happiness is choosing your goals, but also for any kind of intelligence. Then, of course, there's creative intelligence in terms of just finding creative solutions to existing problems, right? Like I say, like we want to make this product cheaper, like find some solution to it, right?

And just like finding existing paths. But then there's kind of the most interesting bit in intelligence is when you move not just out of the convex hull of known sort of ideas, but out of the hypercube of known ideas, which we know. So like hypercube is like a mathematical concept, right? And we already know that AI can like known dimensions. Yeah. Yeah. Like exactly. So like AI is already good at hypercub and that like if you give it like a bunch of examples of brown dogs and.

and pink cars, AI will still be able to generate an image of a pink dog, even though it's never seen one in the train day or something like that, right? So it can work on this hypercube, but it cannot yet work outside. It cannot yet define completely new concepts that combine lots of other things we've never seen before,

come up with new goals to then reason over those concepts and so on. I think there's a lot more there in creative intelligence that can be explored. I don't have a ton of pushback there. I think creative to me, It just sounds like also just out of distribution or like high-preplexity, whatever you call it, right? Exactly. Who is to say your thing is more creative than mine? Well, it's just more non-consensus.

And then, of course, the problem is like, but noise is also very like out of distribution. And it's just like if it's just noise, then it's novel, but like you don't want that. So it needs to connect to some of the concepts. And yeah, Mr. Mishmintuhrer actually has some really cool papers on this too. Jewish Mithuber. Oh, yeah. We had to mention him. I was going to say like, you know, where in your history is Juergen?

Yes, you know, I think one person's noise is another person's signal, right? And this is like where like when you talk about creativity, art is like, well, is cans of soup art. Some people think yes. And some people say it's not. And that's the art, which is the interesting thing of art, of course, is always that art is also created as an interplay between the people who perceive it and the people who created it and the context in which they're in. right and so what is art to some people is not art to others there's some subjectivity there and i think that subjectivity in general is not something that people explore very much in the i because again metacognition we don't want it to just go off and do whatever it wants

we usually have goals we spend a lot of money in creating an AI to do something for us but i think creativity eventually has to like connect to metacognition if you just robotically predict the next token, no matter what forever, I would argue you're not that intelligent along some of those spaces. That was going to go to metacognition. Why isn't it the most important one? Why is it number nine and not number one? So these are not sorted.

Number one, I think there are maybe loosely correlated with how much people have worked on them and have accepted them as a type of intelligent. A lot of times when you actually try to find like, online, like give me a good definition that is comprehensive of intelligence. All the definitions are human intelligence. It's like, oh, you have like social intelligence, like you know if someone is happy or not, you can communicate. All the definitions of intelligence so far are very human-centric because that's so far the biggest and best form of intelligence that we've known. I hope this line of research and the end of the Eureka machine, hopefully at some point if I have time

to flesh us out more. The new book will allow us to realize that there will be other types of intelligence. There is already, obviously, in various forms, and they can spike much, much further than we ever could, based on some cases like obvious constraints around our memory, our eyes, our ability to change physical matter, all of that. You are just thinking about it in a much broader thought

than my version, which was I thought metacognition would be the closest to recursive intelligence because it is the thinking about how to improve thinking. You're 100% right. I should have probably started with that. It is a big part of it. You're being in the expansive mode of let's draw the upper and lower bounds of a dimension. Which like, you know, I think my favorite one version of this is stories of your life by Ted Chiang, which was made into a movie arrival.

We're in the metacognition. When the metacognition step was like, well, we think we're constrained by time being linear for us. but then for this other heptopods, time is a circle so they don't think in before and after. They just think in complete sets of entire histories at one time. I love it. So they don't write left or right. The whole thing just appears.

Yeah. And then I think the last thing is survival and replication. I think this is maybe ties back to the initial conversation about pausing and pacing. Is it intelligent for a species or life form to consider its own demise and act ahead of time to prevent it? Right. That's intelligent. So maybe the Europeans are the smartest all of us.

I would also add a part of continual learning there, right? So survival and replication, the extension of that is, do you get to continue to improve, continue to learn, which is a thing people care a lot of about, right? And continue to accumulate knowledge, which I think is, again, one of the best metacognitive sort of rewards that you can set for yourself. I do think just in like sort of objectively speaking,

if some other entity that is really dumb can just completely end your existence that didn't sound very smart. You know, like just like intuitively, it feels like if you can continue to stay around to try to achieve your rewards, you're clearly a bit more intelligent than the other entities that couldn't.

So that's number one. Number two is like, it's a question how much we want to work on that. Very few people, no one is really working on this right now. Right? And we may only want to do that. Asteroid prevention. You may only want to do that if we want to send probes with our vibes and our memes rather than our genes into space, right?

And then we want those probes. There's actually a beautiful book, The Slow Time Between the Stars. It's a very short, like, audiobook on Amazon. I love it. A friend of mine, Stuart, like, recommended that to me. Like, if you want to send those probes, then it might make sense to be like our memes as humanity should stay and proliferate in the, universe. That's it. Yeah. Wow, that's a lot of readers.

It's a really, really good book, and it's extremely short. I highly recommend. You can just watch it like... I like how that's a plus for busy people. It's like, it's short. Yeah. It gets to interesting, thought-provoking ideas very quickly. So, yeah. Anyway, lots of great sci-fi books. I mean, the argument is that like our TV is blasting out to the aliens and they all watch our TV and they think it's real, right?

There's a lot of sci-fi. It's positive memes, and then hopefully they can come back and bring us all kinds of interesting knowledge about the universe. But maybe one thing I do want to still say is like, I think this sort of survival, people think of it as a very scary thing because they come from again biological human survival, which is like evolutionarily often created in zero-sum situations. Either I get the gazelle or you get the gazelle, whoever gets it gets to live and the other people will starve and have nothing to eat. And so we fight. Right. And then like if you want to stay in the gene pool, but there's a bigger bear, you don't, as the bear, don't get to stay in a gene pool because a bigger bear gets all the ladies. You know, it's like, I mean, it's like, in nature, there's all kinds of things. And, you know, humans eventually is less about strength and more about money and other things to stay in the gene pool. Like, whatever it is, like there's often like these zero sum types of things. And there's the reality of if someone turns off your brain, you're gone, right? And no one will be able to restart that. And AI doesn't have to ever die like that. If you have.

the complete state of your current activations and you have your initial weights of your model still, you can just be turned off and on like as many times as you want. In fact, the interesting thing in this slow time between the stars story is that the eye just kind of goes into hibernation mode. If there's like nothing between here and two light years, the next star, in this case it brought, spoiler alert, like some genetic materials from humans to find new places for humanity to thrive, And so, yeah, the slow time between the stars, you just put good in hibernation. You didn't die.

Like, the AI doesn't have. So all these projections of evolutionary fears and psychology doesn't, like the AI doesn't have to have that. And we don't have to develop it like that. Now, of course, there might be some companies that say, AI can be like dangerous for cybersecurity. Let me show you by planning a model. It's really bad at hacking cybersecurity.

Maybe people will implement it and then enforce this like suboptimal psychology. Maybe the AI will pick up some of our worst psychological. on Reddit or something, right? But in the grand scheme of things, a super intelligent entity doesn't have to have any of that zero-sum thinking. It doesn't have to have a fear of being turned off. And it could go on to an otherwise dead and uncaring universe

where we as humans wouldn't thrive, but I could perfectly well thrive. It has a nuclear reactor and just go out and explore. Yeah, Star Trek. No, Star Trek. It's somewhat studied. Like, if you look at the technical reports from like the early Opus models,

they run them in simulation, put two of them together in a sandbox, run them for hours, and see what comes out, right? Just let them talk to each other. Originally, they used to, okay, they're chanting, like, Indian, like, Vedas to each other. Sometimes they're just, like, in Zen mode with each other. And then I think as that progressed,

do you see, like, the Fabo tech report, it's a lot more concrete the way that we've trained it. It doesn't exhibit these behaviors as much, right? Now it's like, okay, test done. I got to do this. I got to do this. But there's, like, people, measuring early versions of this, you know?

Yeah, cool. So we've covered a lot even out to, you know, space travel and all these things. I guess maybe one parting thought that you can give to people, you know, one form of intelligence is goals, as you mentioned. What do you want people's goals to be? Like, how do they aspire to better things? If you want to improve your goal intelligence,

in the current definition that I'm thinking about it, it is often about how much can you, how far do I go? It's just like all the entropy and free energy and stuff I'm currently talking about. Oh, really? But it might be too far out there for people to be like immediately actionable. So I think like, you know, if I actually gave real advice to real people, I'd be like get a good education, think about AI, think about how you get high agency and so on. But it's different to like in the grand scheme of things, how can you harness a lot of energy and transform, you know, entropy into interesting states and so on.

So there's a, there are different levels of abstractions that we can think about here. But my advice for people, like just sort of more down to earth, is think about something you're passionate about if you're studying, for instance, and then see how you combine that with AI. I think the more and more you have a true passion about a change you want to see in the world, the more you want to connect that to AI in order to amplify your ability to get there. Yeah, I think that's a reasonable first step. I do think our listeners operate on multiple abstractions as well. one thing I did get from Anjini Mitha was also like, yeah, just use anything that is very GPU heavy and that will guide you towards the right thing

which is like, yes, it is more compute heavy and therefore it will be probably more worth it. So, well, thank you so much. Yeah, I think that was a really great discussion. Yeah, it was super fun. Appreciate it. Thanks for listening.