Periodic Labs: A chatbot walks into a laboratory

Liam Fedus https://www.linkedin.com/in/liam-fedus-26547811 and Dogus Cubuk https://www.linkedin.com/in/ekin-dogus-cubuk-9148b8114 are the co-founders and co-CEOs of Periodic Labs https://periodic.com/, a startup building “synthesis superintelligence.” In practice, that means an AI system that closes the loop between hypothesis, experiment, and learning. Liam previously led post-training at OpenAI, where he was one of the creators of ChatGPT. Dogus spent years at Google DeepMind leading a large team of chemists and materials scientists. Together, they’ve brought those two trajectories to bear on a problem that has stymied science for decades: not just discovering new materials, but making them reliably and at scale. In this conversation, we explore the limits of language models, the commercialization bottleneck in materials science, why Periodic chose high-temperature superconductors as its first target, and what Liam learned from watching ChatGPT go from an internal product to one of the most successful products of all time. — Transcript: https://www.generalist.com/i/215855083/transcript — Every timestamp (00:00) Intro (02:05) Periodic’s mission (03:14) What it takes to build a synthetic superintelligence (04:48) Why “thinkism” falls short (05:46) Why Periodic is pursuing high-temperature superconductors (07:21) How Heike Kamerlingh Onnes discovered superconductivity (09:36) Why synthesizing new materials is so difficult (16:44) Dogus’s path to AI (20:18) Liam’s path to AI (23:28) Lessons from launching ChatGPT (25:24) Why Dogus and Liam chose each other (27:01) How Periodic attracts top talent (29:50) Inside an experimental loop at Periodic (35:10) What AI learns across experiments (40:10) What Periodic builds versus buys (44:37) Balancing diversity and depth across experiments (49:23) The difference in LLM performance on math vs. science (51:26) Why Periodic doesn’t depend on better models (53:20) Periodic’s short-term goals (54:09) The competitive landscape (55:39) Final meditations — Liam and Dogus’s complete reading list The Beginning of Infinity: Explanations That Transform the World (recommended by Liam): https://www.amazon.com/Beginning-Infinity-Explanations-Transform-World/dp/0143121359Subtle Is the Lord: The Science and the Life of Albert Einstein (recommended by Dogus): https://www.amazon.com/Subtle-Lord-Science-Albert-Einstein/dp/0192806726— Other resources Periodic Labs: https://periodic.com https://periodic.com/Bell Labs: https://www.nokia.com/bell-labsQBasic: https://en.wikipedia.org/wiki/QBasicGoogle DeepMind: https://deepmind.google https://deepmind.google/Google Brain: https://en.wikipedia.org/wiki/Google_BrainOpenAI: https://openai.com https://openai.com/— Where to find Liam Fedus LinkedIn: https://www.linkedin.com/in/liam-fedus-26547811 X: https://x.com/LiamFedus — Where to find Dogus Cubuk LinkedIn: https://www.linkedin.com/in/ekin-dogus-cubuk-9148b8114 X: https://x.com/ekindogus — Production and marketing by penname.co https://penname.co/. For inquiries about sponsoring the podcast, email [email protected].

• 0:00Speaker 00

  L L M's, if you talk to them, they're incredibly intelligent. Their performance is insane, but then you haven't seen that kind of impact on the physical world. Why is that?

• 0:07Speaker 01

  Our goal is to build a system that can intentionally synthesize matter. This isn't going to emerge just from reading textbooks. It's really necessary to use simulations, model the physical world, but then actually carry out the experiment.

• 0:19Speaker 00

  If we are successful with periodic, it will actually increase the number of scientists in the world.

• 0:23Speaker 02

  Maybe let's get really granular to understand how this all comes together.

• 0:27Speaker 01

  Anyone could go into a lab and mix a bunch of powders and say they ran a lot of experiment. However, if you're not seeing what you're doing, there's no error correction loop, there's no intentionality to it. And so those are some of the first A I systems we had to build. What we're targeting now are areas where the frontier is just not sufficient, or we can do it vastly cheaper. A system that can literally engineer matter is going to be of key importance to everybody.

• 0:50Speaker 02

  If you had unlimited resources and no operational constraints, what is an experiment you'd really like to run?

• 0:56Speaker 01

  I would love to see.

• 1:04Speaker 02

  A I has gotten very good at writing software. That's no accident. After all, code has an answer key a computer can check in a fraction of a second. The program either runs or doesn't. Physics has an answer key too, the physical world. But consulting it means running actual experiments, something an A I model can't do on its own. Liam Fedes and Doj Chubuk are building a company to change that. They're the founders of Periodic Labs, a startup building infrastructure to create what they call synthetic superintelligence. Liam previously led post-training at OpenAI and was one of the creators of ChatGPT, while Doj was a research scientist at Google DeepMind.

• 1:46Speaker 02

  They've brought their experience together to build what might be described as an A I scientist, capable of conducting thousands of experiments a day and learning from each one of them. In this episode, we discuss the limits of L L M's, why science is so hard, and what Liam learned from launching ChatGPT. I'm Mario and this is The Generalist. I'm really excited to have you both here. I've been really looking forward to this conversation. I think you're building such a fascinating company, so maybe we just start there. Tell me a little bit about periodic. What are you trying to do?

• 2:17Speaker 01

  Our goal is to build a system that can intentionally synthesize matter. So we're calling this a synthesis superintelligence. Our core belief is in order to do this, you need to make connection with the real world. So a system like this isn't going to emerge just from reading textbooks, reading papers. It's really necessary to use simulations, model the physical world, but then actually carry out the experiments.

• 2:44Speaker 01

  We kind of talk about like thinkism is not enough. It's really necessary to get the new information from the world to conjecture how it might be, get the experiments back and do this end-to-end loop. By doing this end-to-end training against the physical world, new types of intelligence and A I systems will emerge. So rather than a system that has just been trained on the body of final scientific literature, it's been trained on the scientific process itself. So rather than knowing the state of science and knows how to do science.

• 3:12Speaker 02

  I've heard one way that you've maybe described the mission in the past is almost to build an A I scientist, right? And decomposing the different pieces of that puzzle. Through that sort of lens, how do you think about, you know, fundamentally, what does a scientist do? And what do you need to replicate?

• 3:29Speaker 00

  The traditional way to look at science is to say you come up with a hypothesis. And then based on the hypothesis, you do some experiments and the experiments either validate that hypothesis or invalidate it. And then you update your hypothesis. And I think what's really critical here for us is this observation that science always tries to change the boundary of our understanding, which by definition means it cannot just rely on what we already know. Because science is trying to figure out what we don't yet know. And this is almost at odds with the traditional description of machine learning, where machine learning trains on a training set and does well on a test set that has the same distribution as the training set. So there's like a clear tension between what we call machine learning and what we call science.

• 4:13Speaker 00

  And that's where we're basically trying to make progress. And how have humans gotten around this problem, right? Because at the end of the day, humans are also machine learning algorithms. And they often have figured out things that they didn't know before. And they had to iterate, like they had to run experiments or they had to write down a theory, get yelled at by reviewers and then updated. You know, Einstein had to do this, Bardin had to do this. So we feel like L L M's will have to do it too. So we've been trying to spend a lot of time and focus on building labs that can have an L L M and simulation iterate with it. That's kind of our goal, yeah.

• 4:49Speaker 02

  Liam, you mentioned sort of this idea of thinkism. For folks that maybe don't, you know, know what that really means, maybe you can tell us a little bit about it and, you know, parse the differentiation a bit there.

• 4:59Speaker 01

  Yeah, I mean, I think it's this concept that an A I that just thinks for a very long time. Maybe if you extrapolate out the inference tokens well beyond anything we've seen, it can just do anything. One of our deep scientific goals is to discover new high temp superconductors.

• 5:16Speaker 02

  Yes.

• 5:16Speaker 01

  And we don't think that a system trained in the way that they're trained today will think for a billion or a trillion tokens and suddenly into it, here's a new high temp superconductor. We think that's just not in line with how we've pushed knowledge forward and science forward traditionally. And so I think that's kind of core to our belief. It's like you need to connect it to the physical world and iterate. The data simply just doesn't exist and that's how you're going to actually build new knowledge, not just rehash the existing knowledge.

• 5:47Speaker 02

  High temperature superconductors, how did you land on that as sort of the right first step for this company?

• 5:52Speaker 00

  There are a couple of reasons. Let me talk through a bunch of them. So as Liam said, our goal is synthesis superintelligence. And if we can synthesize anything in the lab, we might as well synthesize the most impactful thing, which is a superconductor. So that's one reason. Another reason is imagine all the different sci-fi futures you want to think of. They often involve superconductivity, like quantum computing, fusion, lossless transmission of energy. These all have superconductors somewhere in there.

• 6:22Speaker 00

  And the other thing is it's like, there's a sense in which superconductors are the most macro scale quantum mechanical phenomenon. There's something very like fascinating about it. And, you know, if you think about the history of condensed matter physics, it goes through these like seven year or so cycles of what's been very popular and then falls out. Like there was topological insulators, heavy fermion physics. Like it kind of roughly matches the length of a PhD that like these topics come up and down. Superconductivity is one topic that never went out of rogue. You know, like since its discovery by ONES back in nineteen ten, people have been fascinated by it since and we're still fascinated by it.

• 6:59Speaker 01

  And it's also in the pursuit of this, you have to build up such sophisticated simulation infrastructure, A I infrastructure. And this has a huge amount of generality.

• 7:09Speaker 02

  So it sort of gives you the right stair steps and also is this massively impactful.

• 7:14Speaker 01

  Exactly. In the pursuit of this super ambitious goal, you build up a huge amount of incredibly impactful technology.

• 7:21Speaker 02

  Doj, I've heard you, you mentioned ONES before. For folks who don't know, you know, that history, what they achieved, can you tell us a little bit about why they're an inspiration?

• 7:30Speaker 00

  Yeah, absolutely. Because ONES is an inspiration to us in many ways. So this is a good thing to talk about. So in the eighteen hundreds, there was a big rush to liquefy as many of the gases as possible. So I think around eighteen fifty, they liquefied oxygen, for example, and ended up being playing a big role in industry. So they were trying to kind of go down in the periodic table and liquefy as many of the gases as possible. As you can imagine, you have to lower the temperature more and more. Helium was the hardest one because, you know, it's a noble gas. So it doesn't really interact with other atoms very strongly, which means this liquefaction temperature will be low. And helium especially among the noble gases is the smallest one. So that was the last one to be tackled. And ONES built a lab to go after this.

• 8:13Speaker 00

  And one of his very interesting ideas was that science should be done at an industrial scale by professionals. So he would hire professional technicians, professional engineers to work in a science lab, which was, I think, a bit like funny at the time to people. And he succeeded. So, you know, after a lot of hard work and he had a large team doing it, he was able to liquefy helium. And interestingly, you know, he got a Nobel Prize for liquefying helium. He didn't get it for discovering superconductivity. So he just like basically created a new capability, which is lowering the temperature to below four Kelvin. As you know, zero Kelvin is the absolute zero, like we don't expect to go any below that. And four Kelvin is really cold. But he reached that, and once he reached four Kelvin, he could like go down a bit more to one point eight.

• 8:56Speaker 00

  And then after he was able to do that, he knew that some people were wondering what would happen to metals if it got really cold. And some really smart people like Lord Calvin thought that if you could go such low temperatures, electrons would basically freeze. And it's kind of intuitive, right? It's so cold, electrons don't want to move anymore.

• 9:13Speaker 02

  Yes.

• 9:14Speaker 00

  So he wanted to try it. And I think one of his scientists there tried putting different metals into liquefied helium or like very cold temperature. And one of the things they tried was mercury. And turns out mercury is a superconductor at such low temperatures. And then once you observed that it was a big deal and that kind of opened up this whole area around superconductivity.

• 9:35Speaker 02

  Thank you for that history. That's very interesting. You, you've both mentioned this phrase, synthesis superintelligence. I'm sure folks have an understanding of what synthesis is, but you know, why that word? Like, why is that the right wrapper or the right description? There's something sort of, I imagine, quite deep there and why you've chosen to frame it that way.

• 9:52Speaker 00

  Yeah, I mean, there are historical reasons and scientific reasons. So one historical reason is U S used to be really good at synthesis. And especially during Bell Labs time. But after Bell Labs stopped, you know, doing research as much, U S kind of lost its power and strength in synthesis. And it's kind of moved on to some other countries.

• 10:11Speaker 02

  Like synthesis, when you say Bell Labs was great at synthesis, like what is an example of that?

• 10:15Speaker 00

  Good question. So like, let's say there's a new compound, new material that is interesting. And I come to you and say, Mario, can you make this?

• 10:21Speaker 02

  Yes.

• 10:22Speaker 00

  It's can be very hard. So if it's been been made before, then you can just follow the previous recipe. It's kind of like baking, but it's never been done before. It's actually hard to figure out because you're basically like at a high level, trying to get the atoms in the right shape. And it's a pretty difficult thing, right? Because you're not actually seeing the atoms, you're not able to move the atoms individually. Today, U S academia has focused a lot more on functional properties, like finding a new battery, finding a new solar cell. But there's this fundamental capability that's still very important, which is being able to synthesize things. Okay, so this one reason. Another reason is in materials discovery, you might have heard of this like twenty year bottleneck, where you might discover a material in the lab, but it usually takes ten to twenty years before it impacts the technology. And one of the main reasons for that is you can make it in the lab,

• 11:05Speaker 00

  but when you need to figure out how to scale up the manufacturing, so you can make it at industrial scale, so you can sell it, there's usually a huge issue. And a lot of the things you see in the lab don't transfer to large scale. So the second, and the third is, to be honest, you know, sometimes people think of materials discovery, the hardest part being like discovering the material. But if you go in the lab, usually the hardest part is making the thing you discovered, because it's kind of relatively easy to figure out what could be an interesting material. What's usually hard is making it. Like, let me give you a quick example. So when Alex Mueller wanted to study transition metal oxides for new superconductors, and you know, he succeeded at the end towards nineteen eighty-five, and he discovered cooperates, which were the first unconventional superconductors,

• 11:49Speaker 00

  and got a Nobel Prize for it. These were the high-temperature superconductors. They were actually originally trying nickelates instead of cooperates. So if you look at the periodic table, nickel is right next to copper. So instead of copper oxygen, it's nickel oxygen. So he tried that, it didn't work, so gave up and went to cooperate. But then, you know, forty years later, our neighbors at Stanford, Harold Huang, discovered that nickelates are also superconductors, once he could figure out how to make it. I guess what we're trying to say is, even for materials discovery as the main goal, being able to synthesize materials efficiently is the biggest bottleneck. So we're spending a lot of our time on figuring out how to make things that we want to make.

• 12:27Speaker 02

  For someone who, you know, last conducted a scientific experiment, probably in tenth grade, and, you know, nothing serious certainly. When you say it's hard to make things, like, what are the ways that that is difficult precisely? Like, is it a question of precision? Is it a question of consistency? Like, you know, for, you know, pick any example, I suppose. What are the sort of most common ways that process fails?

• 12:52Speaker 01

  I mean, like, even if you know how you want to configure those atoms, you're like, okay, if the atoms were arranged in this particular way, it would have the, it'd be stable, it would have the properties we care about. There's all of the processes you need to do to lead up to it. So what elemental precursors do you use? And what proportions? What temperature profile? What pressures? What are all the processing conditions? There's a huge amount of challenges associated with all these things, and it's necessary to actually produce a new material. It's insufficient to just bring some of the precursors together at ambient pressure, ambient temperature. You have to sort of bring it into new regimes. And this is extremely difficult. But once you have a system that can do this, it can intentionally and efficiently start designing matter,

• 13:37Speaker 01

  which you think is extremely foundational.

• 13:39Speaker 02

  Another sort of naive question. But once you have, you know, the recipe, so to speak, for one of these things, how easily can you make that recipe consistently? You know, it sounds like, you know, once you have the idea, okay, hard to then do it the first time and get it right. But once you've sort of dialed that in, is that quite replicable? Is it very replicable? Is it actually surprisingly hard?

• 14:00Speaker 00

  It can really depend. So some reasons it's hard to replicate. Let us just like tell you some examples, so it kind of like gives some like reality to it. Furnaces, for example, so you know, like, if you think of very simple synthesis of them, like solid-state synthesis, you just mix precursor powders, put it in an oven, for furnace, for some time, at some temperature, and then take it out. Like, this is the simplest thing you can do, probably. Even there, there are a lot of issues. So one issue is the precursors might have different amount of impurities. So, you know, you buy this precursor, it's ninety-nine percent one, ninety-nine point one percent pure, but then there is other impurities, and impurities can lead to other results. Another thing is furnaces change their temperature stability over time. Like, one thing we realized in our lab, that there was definitely a pain, is

• 14:45Speaker 00

  the furnace was working pretty well for the first three runs, but then afterwards, this temperature stability was going haywire. And, you know, it's stuff like this that can make replicability difficult. But humans have definitely gotten better at it as they got experience. Like, another example from cooperates is, I think when cooperates were first discovered to be superconductors, actually, some people wondered if it was even real, because they had trouble replicating in the lab, and it was a difficult synthesis procedure. Today, I think people have dialed it in so much, it's in textbooks. Like, if you open up a solid-state chemistry textbook, some of the synthesis examples will include cooperates, like YBCO. So it's one of those things, again, that where at the beginning, it might seem not very replicable, hard to figure out, but as we gain experience from it, we kind of get better at it.

• 15:29Speaker 01

  Yeah, and one key piece to what we do at Periodic is we think about the hardware engineering associated with all these things, too. So something that used to be a hidden variable, where it's like, oh, the oven is actually degrading slightly over time, or there's asymmetries within the furnace. If you can get better telemetry on these things, then that allows you to pipe that information back to the system, and it can have a better sense of this.

• 15:53Speaker 02

  I imagine, you know, one of the things I think about when we think about the sort of state of modern science, modern knowledge, is the replication crisis. You know, it's hard to reproduce the results that we think we know. We don't necessarily know some things that it might not be true, or, you know, we don't know them with full confidence. It sounds like because of that telemetry and, you know, the way that you can sort of engineer this system with a different level of fidelity, you can hopefully start to really figure out why these things fail and maybe have better replication, better reproducibility.

• 16:20Speaker 01

  Yeah, I mean, we also run replicates as well, right? So we have to understand what is the variance of different systems and really teach a system to do a better job of like decision-making under uncertainty. So, for example, if there's some aberrance, some noise in one pattern, don't have it like over-indexed on that and say, actually, no, this should just be rerun. This looks like noise. Having that type of scientific intuition is the thing we want to build into these systems.

• 16:45Speaker 02

  I want to go so much deeper into what you've built, but maybe to, you know, take a step back in time. I think you guys have said before that you met at the gym flipping a tire. How did you both get to that point? Maybe, maybe Doge, we can start with you. You were, you were telling me, you know, before we started that you grew up in Western Turkey, I believe.

• 17:02Speaker 00

  Yeah, that's right. Yeah, so I, like, growing up as a kid, I always knew I want to be a physicist. But what I didn't know at the time, but looking back, kind of makes more sense, is it seemed like I always enjoyed simulating physics and math. Like, I know when I first, if I say Q basic, will you know what this means?

• 17:17Speaker 02

  I will not. But some of the listeners will.

• 17:20Speaker 00

  So, yeah, Q basic was this language that was available when I was a kid. And I remember as soon as I got my hands on it, once I figured out how to program, I was just doing like simulations of like applied math, physics. But at the time, it didn't make sense for me. At the time, I was just thinking of theoretical physics, like heroes were like Einstein, Feynman. And then in high school, again, I was just like learning more languages, like Java. And again, it was just like all those simulating stuff. And then, when I joined my, when I started my PhD, I kind of realized this issue that physicists were using tools from like seventies, sixties for analyzing data. But the amount of data has increased exponentially. And the reason for it is Moore's law, right? Like, people in the nineties had exponentially more compute power than people in the seventies.

• 18:05Speaker 00

  So physicists were producing so much data, but they were analyzing it by hand or by like simple analytic models. So this was back in two thousand and ten. I felt like we should see how machine learning can do physics. And back then, there was all the pushback, because, you know, machine learning is probabilistic, it's biased, it's not always calibrated. And physicists, especially Platonist physicists, I guess physicists have different philosophical leanings. But the Platonist ones tend to think there's one reality, one theory, the beautiful ones, of course. So there was a lot of pushback at the time. But in two thousand and twelve, when the ImageNet moment happened, and then deep learning got more and more prominence, by the time I finished my PhD, machine learning was everywhere. And actually, physics departments had to put machine learning somewhere in the grant proposal, especially in condensed matter physics, to even get a grant.

• 18:50Speaker 00

  So huge change in like five years. And during that time, Google Brain had this residency program. And I think it was like partially an idea by Ilya, it's a brilliant idea. They want to bring in people who have some machine learning background, some interest, but expertise in other fields, to learn deep learning and contribute to deep learning at Google Brain. So that's how I joined, I was part of the residency program. And Liam also joined in a similar way, he'll explain himself. And yeah, so we were both doing deep learning research. I was at the gym, trying to flip a potato, it was difficult, so I picked them as the strongest person I could find.

• 19:23Speaker 02

  Yeah, you saw the arms, this guy's gonna help me.

• 19:25Speaker 00

  And it succeeded, you know, we managed. And so I was continuing to do like regular deep learning research. So you might remember, this was an exciting time, I would just publish at NeurIPS, ICML, like, deep learning was advancing with these conference publications.

• 19:38Speaker 02

  Yes.

• 19:38Speaker 00

  There was a lot of fun to that. And then I was still doing my kind of passion from before, which is simulating solid-state physics. And then, at some point, Liam and our other friends, Barrett, Luke, have left, and then started ChatGPT, and I ended up staying and like really focusing on physics. So by the time, like, before we left, before I left, I was running this large team of chemistry and material scientists at Google DeepMind. And we were doing a bunch of simulations and machine learning. Liam and I often connected and talked about physics and other things. And we felt like the technology had reached a place where we could really impact the market, the industry. So we left our respective roles to start the company.

• 20:16Speaker 02

  Hmm, amazing, okay. What is your origin story?

• 20:19Speaker 01

  Yeah, origin. So I grew up in Maine, and also had like a really early love for physics. And so in college, I was actually a bit between physics, chemistry, economics, and I was part of a dark matter detection program. And one of the key pieces for that was reconstructing the particle tracks, which became like a physics model. But it's really a physics-informed machine learning model. And so I started doing machine learning in physics in a very early stage. This is still an undergrad. And then again, in graduate school, reconstructing particles, this time at CERN. Through this, we were pushing the amount of computation we're putting into this and, you know, some of the algorithms for machine learning.

• 21:02Speaker 01

  But after a while, it became clear, if I really wanted to push machine learning, I should be doing it in industry.

• 21:08Speaker 02

  Yes.

• 21:09Speaker 01

  And so I was actually at Google, part of the Google Accelerated Sciences team, and then later Google Brain. So that's where Doge and I were meeting. I was working on a mix of some things in the science program, but then switched over to generative models, reinforcement learning, and really was just in this like scaling mentality. Wasn't necessarily applying to physics, but just saw how impactful language models were becoming, and was building some of the architecture necessary to scale those up. So this is using sparsity, so switch transformers, mixture of experts. This became an important architecture for serving today's models.

• 21:46Speaker 02

  Yes.

• 21:47Speaker 01

  And continued to scale that up, did some of the first trillion parameter models. But at some point, it became very evident that there is a huge possible impact for this in. In the broader world.

• 21:59Speaker 02

  Yes.

• 21:59Speaker 01

  In a product. And that's what led me and a few others to OpenAI. We felt it was going to be really a great time to start productionizing some of these things. And so at the time when we arrived, there was GPT three point five out as an API. GPT four was internally available and we had some period to develop it. So we're working on that. And then at some point, we realized, let's just do a quick sprint using GPT three point five.

• 22:27Speaker 02

  Yes.

• 22:28Speaker 01

  And it was a two week sprint and then we turned this into ChatGPT.

• 22:33Speaker 02

  And did that product do well?

• 22:34Speaker 01

  Uh, yeah, it was okay. It exceeded some of our expectations.

• 22:37Speaker 02

  Yeah.

• 22:37Speaker 01

  We had internal polls as to how many people use it and people were guessing numbers like ten thousand, fifty thousand, someone. Yeah, like a high bar was like a million.

• 22:44Speaker 02

  No way.

• 22:45Speaker 01

  It was a low key. It was genuinely a low key research preview.

• 22:48Speaker 02

  Your, your mental model must have then sort of like totally shifted, right? I mean, I guess that happened to literally all, all of us, but you were.

• 22:54Speaker 01

  Well, I mean, I think the context before that too is Meta had just released Galactica and had to withdraw it within a few days of release.

• 23:01Speaker 02

  I didn't remember.

• 23:01Speaker 01

  That was only a couple of weeks before and there had been no successful chatbot up until that point. So the prior that this is going to be a valuable technology was quite low. It was further surprising because we had GPT four internally and it wasn't this hugely viral product within the company. So, you know, putting out a weaker model to the public, it was really just part of this iterative deployment. And yeah, it obviously it worked better than our predictions.

• 23:29Speaker 02

  What was the period after the release like for you? Was that just like?

• 23:33Speaker 01

  Just totally frenetic.

• 23:34Speaker 02

  Yeah, chaos at all times.

• 23:36Speaker 01

  Yeah, I mean, like servers couldn't keep up. Um, and you just trying to like rapidly improve it. You know, the early model couldn't do math, couldn't write very well, uh, had, you know, many deep flaws, um, very different than the things we have today.

• 23:49Speaker 02

  Yes.

• 23:50Speaker 01

  But just a series of getting feedback, improving the task distribution, improving the data quality, just all these kind of boring details is what led to the system as of today.

• 24:00Speaker 02

  You know, very few people ever will have an experience like that, I suspect. But I imagine it serves you in very good stead as a founder today. What do you think it taught you? What did it give you?

• 24:11Speaker 01

  Question everything. I mean, so it's really interesting. So people would come in years after the initial creation and it's very easy to anchor on a random decision made a year, year and a half ago. Not necessarily like an optimal decision, but you sort of take that as sort of the assumptions that you're building upon. But I can recall back to, no, that was just one conversation, like one IC made that feel free to change these things. And as we were sort of building that up and eventually I was running the, the full post-training team.

• 24:44Speaker 02

  Yes.

• 24:45Speaker 01

  That sits between research and product. You know, we made incredible progress in the digital world, but in Doge and I were connecting, we felt that the impact of these systems in the physical world was always going to be constrained unless we actually connected it to the real world. We just felt, again, reading on text, reading textbooks, reading papers, deriving questions from that is a great foundational technology, but insufficient. And I think ultimately like the, the main objective for artificial intelligence that we're really excited about is new scientific discoveries, new technologies, something like really tangible you can hold in your hand. And so that was sort of the motivation leading up to periodic.

• 25:24Speaker 02

  You know, you, you both have worked with some of the very best builders and researchers in this space. Why did you pick each other? Like, well, you know, what was it about Doge that you were like, you know, this is the guy I want to spend the next, you know, chapter of my career with?

• 25:38Speaker 01

  I mean, just absolute brilliant thinker. There's no one better in the world to, to do this problem. The Doge has been like at this intersection for fifteen years, more.

• 25:50Speaker 00

  Yeah, fifteen years, yeah.

• 25:51Speaker 01

  Right. And yeah, I mean, we've known each other for a decade. And so I, I think it was just obvious, there's no one better in the world to do it with.

• 26:00Speaker 02

  What about you, Doge? Why, why Liam?

• 26:02Speaker 00

  I mean, many reasons. So one of them is, you know, I feel like with a co-founder, the most important thing is trust. Because you're trusting your own career, but also the people you bring to the company. And, you know, there are some people in our company who I worked with for many years. Liam has worked with many years and I wouldn't do it with someone who I couldn't trust fully. Another one is Liam, like has been extremely successful on the LM side, but he's always been very passionate about physics. Like whenever we talked about just randomly catching up, he'd actually like has his eyes about quantum physics, quantum computing. So like, I think it's really important for the person to really care about these things.

• 26:38Speaker 02

  Yes.

• 26:38Speaker 00

  Uh, and I could tell that Liam really cares about physics. And when there's so many exciting things he could do with the LM side, like he would do this. And then the, of course, the other part is we needed deep expertise in RL and LLMs. And again, like Liam has done RL research and kind of contributed at the frontier for so long. So it was a very obvious, easy choice, yeah.

• 26:58Speaker 02

  You, you mentioned, you know, bringing people into the company. I think one of the things that, I mean, you can see it from the outside, but also talking to some of your, your investors in preparation is the talent density you guys have created is extremely impressive, especially for a company this young and, you know, still in the very early innings of, of what you're up to. How have you convinced some of the, the very, very best folks in, in this industry who are being competed over ferociously to, to pick this as their mission?

• 27:28Speaker 01

  I think mission.

• 27:28Speaker 02

  Yeah.

• 27:29Speaker 01

  Yeah. Like quite simply, it's like people see just the ambition of what we're trying to achieve and the, the possible impact of this. Like something that can actually design the physical world, um, something that understands physics, chemistry is even more foundational than a system that understands language. And I think that really appeals to people. I think it really appeals to people too, in terms of how much leverage, how much of a diff on the world can you have in this position?

• 27:57Speaker 02

  Yes.

• 27:58Speaker 01

  And so while you might be a very successful contributor to a coding system, the coding system, you know, there's always this counterfactual you can run, which is like, okay, well, if I'm not the five hundredth contributor on this coding agent, what's the delta? And I think people are looking at that and like, well, okay, I think my skills could be better used elsewhere and I can have a larger impact on the world. So I think that appeals to a lot of people as well as some of the scientists, just seeing how rapidly AI is advancing in the digital era. So math, theoretical physics, et cetera, and wondering what could this mean for physical sciences? And I think that, that's a huge appeal.

• 28:37Speaker 02

  I imagine it doesn't hurt that, you know, when you look at the, the best researchers in AI, so many of them have physics backgrounds. Like it really seems to have gotten a lot of that talent. So I imagine for a lot of these folks that you're talking to, it's sort of a chance for them to bridge their childhood love of what they were doing and, you know, their first passions with, you know, maybe where they've built some of their career. I don't know if that's how you've, you've seen it.

• 29:00Speaker 00

  And definitely, like I think there are two aspects to this. One, as you said, a lot of these researchers really deeply care about science. And this may be like a personal interesting, but on the other hand, they really want to have AI impact society in a positive way. And I think science is like such a good direction for AI because there's no end to it. For example, we expect that if we are successful with periodic, it will actually increase the number of scientists in the world. Uh, whereas for certain roles, you know, AI, if it replaces the employee, it's kind of like a zero sum game between AI. But science is the opposite. Like the more we can show that science can impact the world, especially South State physics, that South State chemistry, the more people will go into that field.

• 29:43Speaker 00

  And there's no end to it, as you know, like there's no end to science, there's no end to engineering. Like we can figure out more science, we can visit more planets.

• 29:50Speaker 02

  Maybe let's get really granular to understand how this all comes together. To, to the extent that you're, you know, able to share it, what does like one experimental loop look like with periodic? Like what are the, the, the different machines as part of that, the different processes as part of that?

• 30:05Speaker 00

  Yeah, so um, let's say we're doing powder synthesis. And let's say we're trying to make a superconductor.

• 30:10Speaker 02

  Yes.

• 30:10Speaker 00

  So the first decision is, you know, what precursors you want to mix? Because the thing you're trying to make probably doesn't exist on your shelf yet. So we have to first get the ingredients and the ingredients can usually be like these binary powders. Like binary means two kinds of atoms are in it. Uh, it can be like the elemental precursor, which is the single element. But yeah, basically you figure out for the thing you're trying to make, what precursors to combine.

• 30:34Speaker 02

  Is that you figuring it out or is it the, the AI's figuring it out?

• 30:37Speaker 00

  Great question. So traditionally it would be the human figuring it out. For us, uh, the AI can actually really speed up this process. And especially, you know, down the line, if you want to run thousands of experiments a day, we can't have a human do it one by one.

• 30:49Speaker 02

  So when it's speeding it up now, it's maybe saying, here are thirty good options for you. Is that sort of like narrowing the scope of what you might look at and introducing new ideas?

• 30:57Speaker 00

  That, that can happen or you can actually say, okay, for this goal you have, here is what you should mix. And we can try that. Okay, so you mix the precursors, now you need to react them. Because at room temperature, these materials won't react with each other, so they'll just stay as the precursor. So there are different options, but a very common standard option is what we talked about, which is you put it in a furnace with some temperature profile. Uh, and Liam has often brought up that the temperature profile looks like a learning rate decay profile, right? Like in machine learning.

• 31:26Speaker 02

  And sorry, I, I am just so interested in each of these steps, putting it into the furnace. Is a human doing that? Is a robot doing that? Is that like already automated?

• 31:34Speaker 00

  That has already been automated by others before we started periodic. So there's already been attempts at getting a robot or a robot arm to take the mixed precursor, put it in a furnace. And we have both, like we can either do that or we can also do, have a human do it.

• 31:47Speaker 02

  Yep.

• 31:47Speaker 00

  But again, like as you want to scale these up to thousands a day.

• 31:50Speaker 02

  Yes.

• 31:51Speaker 00

  Then you definitely need robots to do it.

• 31:52Speaker 02

  Yes. Okay, so it's in the furnace.

• 31:54Speaker 00

  Yeah.

• 31:54Speaker 02

  Then what comes?

• 31:55Speaker 00

  The very traditional path is you wait a certain amount of time you predetermined or the AI has predetermined and then you take it out. And you hope that during that duration, time duration and during that temperature profile, the reaction you hope happens, happened and the reactions you don't hope happens, didn't happen. Uh, so now you have it, you cool it down a bit to not, you know, burn your hand or something. And at some point you put it in a characterization instrument. For inorganic crystals, the method we use the most is an XRD, X-ray diffraction. It's the same X-rays you get at the dentist. It just happens that X-ray wavelength is of a comparable length scale as the distance between atoms. So then if you deflect X-rays through solid, you can see what the atoms are doing and like how they're positioned against each other.

• 32:41Speaker 00

  And then that tells you if your experiment succeeded or not. And that is one part, actually, we find that AI can play a big role because once you get the X-ray diffraction pattern, it's not always easy for a human to tell what you made. But the AI can help you with that and then tell you, okay, this succeeded or it didn't succeed and send it back to a new experiment. And then if it succeeded, maybe you want to see if it's a superconductor. So then you put it in another instrument that can measure the diamagnetic response, which tells you if it's a superconductor at a given temperature.

• 33:09Speaker 01

  Yeah, I think the XRD is a really interesting example, too, because the materials don't come out labeled.

• 33:15Speaker 02

  Yes. Well, here is what you've done.

• 33:17Speaker 01

  Exactly, that's right. So AI is allowing us to basically do the labeling and figure out what we're doing, to see what we're doing at a much faster pace. Anyone could go into a lab and mix a bunch of powders and say they ran a lot of experiments. However, if you're not seeing what you're doing, there's no error correction loop, there's no intentionality to it. And so those are some of the first AI systems we had to build. And then that allows us to scale up the throughput of the lab significantly.

• 33:44Speaker 02

  Are you gathering sort of, I don't know, lots of ancillary data along the way? Is that useful? Is that not useful? I don't even really know what like that would be, but I imagine a mechanized version of this process permits that in a way that, you know, a purely human version does not.

• 34:00Speaker 01

  I think like better telemetry is one aspect, but whether it's robots or humans, I think stitching the data together is one of the most important pieces.

• 34:09Speaker 02

  Interesting.

• 34:09Speaker 01

  So this is what I was alluding to at the beginning, where it's where most AI has been trained on the final artifacts of science. It's like the final paper, the the final, you know, recount in a textbook. That isn't often how the scientific process unfolded, sort of a retelling of the of the story. And what we do is we're very careful in collecting what were sort of the intentions, what was the hypothesis? What were the computational predictions? What was the execution through all of the machinery? And this data that fully stitched together, this full lineage is very, very unique. That's a something that we've been spending a lot of time putting together. And then a system trained against that is learning not just what the final outcome was, but what was the process? What were the judgments made?

• 34:53Speaker 01

  What were the reruns done as new evidence came in? Basically, what was the process of doing science?

• 34:59Speaker 02

  In a way, maybe this is a bad analogy, you'll you'll know better than I do, but it feels like it's sort of almost chain of thought for science, where you're sort of starting to see.

• 35:07Speaker 01

  Yeah.

• 35:07Speaker 02

  Yeah.

• 35:07Speaker 01

  Interacting with the real world.

• 35:08Speaker 02

  Yeah. Yes, exactly. As you've now started to run some of these experiments, what do you notice that having that, you know, trace actually gives you? Is it like, I don't know, better suggestions on the next experiment that you would have ordinarily got? Is it, you know, the next whole set of experiments is somehow leveled up versus what it would be if you started cold?

• 35:31Speaker 00

  Yeah, I mean, one thing is just like being able to pay attention to detail when there's so much detail. Like, we've seen that the LLM can actually detect major mistakes and debug it. Like, one thing that happened in our lab is turns out the sample, you know, samples are put in this like circular sample holder. And turns out just we, you know, made a mistake and they were shifted by two. So now all the samples are mislabeled, you know, we have no idea what happened. And the LLM was able to just like go through it and then realize, oh, you just like shifted counterclockwise by two. So if you do the inverse rotation, now all of them are labeled correctly. Like something like that a human could do, but it just takes so much attention to detail and like so much time.

• 36:15Speaker 00

  Now imagine doing these experiments thousands at a time. The other thing that's really, I think LLMs are really good at, which I'm really excited by, is they can be good at things that usually require different expertise and different PhDs. Like, for example, our LLM can do simulations at the level of a PhD in simulations, but it can also do experimental data analysis at the level of a PhD in it. And that's very unique, right? Because most people either get a PhD in simulation or PhD in experiment. And some things fall through the cracks. So the fact that we have a platform now that is an expert on all of these different modalities and can reason at the same time. So something that happened for us is the XRD result comes up and it's not what we tried to make. And it's also not a material that's in a database, so we can't tell what we made.

• 37:00Speaker 00

  But then the LLM actually ran new simulations and then figured out what it must have been. And then we were able to validate it with the next experiment. Because I often say the people who should run simulations and the people who know how to run simulations are usually disjointed sets of people.

• 37:14Speaker 02

  Yeah.

• 37:15Speaker 00

  And I think the LLMs will bring these together.

• 37:17Speaker 02

  Yes.

• 37:17Speaker 00

  And it's a fractal, right? So I'm just talking about one interface between simulation experts and non-simulation experts. But it's also true for like chemists and physicists, inorganic chemists and organic chemists. Like, the expertise of humankind is kind of like a fractal. And there are all these interfaces in the fractal. And and LLM, they can speak the language of all these different groups. I think it's like gonna be a qualitative change.

• 37:39Speaker 01

  I think another interesting thing to add, too, is as we build up this database of all the things run, all of the decisions, you can start constructing new types of machine learning tasks where you can envision snapshotting the state of the world. And you're like, okay, well, given the state of the world, computationally, experimentally, what did the scientists do next? And then like, what was the outcome of that? This is a very rich source of training and reinforcement learning. And like, one analog that's been talked about before is let's train an LLM up to general relativity. And then let's see if the system can like produce such a thing. And we're kind of doing that, you know, in our scale, in our own experimental data, our own computational data.

• 38:21Speaker 01

  And so this is like a set of data that really confers like a huge advantage where that just doesn't exist anywhere.

• 38:28Speaker 02

  You mentioned this idea of a sort of a snapshot in time, which I think is such an interesting idea, to maybe go a level deeper with that, like what might be a snapshot periodic would take? And what would you get out of looking at that?

• 38:41Speaker 01

  So let's just say hypothetically, you had 100,000 experimental data points. What you can do is you can say, what was the state of experimental evidence at some date? So you can just literally say, these were computational predictions, these were the experiments we'd run up till that point. Here's some notes and thoughts on what we were thinking. And from that point, that becomes your new environment state. And you'd say, now, given this basket of data, what did the scientists do next? And given what they did, like, what was the outcome of that? What were the new empirical measurements? This becomes a incredibly interesting source of reinforcement learning type tasks and data

• 39:23Speaker 01

  that is incredibly hard to construct outside of periodic. And this is really interesting because now you can build this synthesis superintelligence to begin to predict what are the outcomes based on all of this data. And it's easier for us to do this type of work because as we accrue more evidence and more data that doesn't exist in public literature, we know our model isn't contaminated. And there's a risk when your pre-trained model already knows the answer and you do reinforcement learning derived against an answer that was already in the pre-training corpus.

• 39:54Speaker 02

  Yes.

• 39:54Speaker 01

  That the system can basically fake some work, get to the right answer, and you reinforce those strategies.

• 40:00Speaker 02

  Yes.

• 40:00Speaker 01

  Strategies like that aren't going to generalize and aren't going to be particularly interesting. So we're able to circumvent that by having our basket of data that we know does not exist in literature.

• 40:11Speaker 02

  When I think about, you know, what you're doing, there's obviously so many different pieces to to pull together. What does it not make sense for you to do? Like, you know, do you need to develop every piece of hardware yourself? Or, you know, can you use someone else's robots? Will you, you know, create your own models ultimately? Or can you sort of just fine tune and do these different things as part of the process? Like, what, yeah, what do you feel like you need to uniquely own?

• 40:39Speaker 00

  You know, for the hardware things, like if there's an off-the-shelf component we can just buy and use, we're very happy to do that, you know, we don't have to reinvent the wheel. So we usually do that. Turns out for certain scientific instruments, the throughput isn't there or the quality isn't there. So for those, we actually invest in hardware engineering expertise to improve the throughput, et cetera.

• 41:00Speaker 02

  Like, what's an example of something that just is not is not going to hit it at your scale that you have to sort of reinvent in a way?

• 41:06Speaker 00

  I mean, the the one that already kind of matches our needs is XRD. So there's already so much work into high throughput XRD that we have these instruments that we can put, you know, 40 samples at the same time. And we'll do X-ray diffraction, all of them at the same time. But for things like that are more materials property oriented and it's like slower measurement takes longer, you might get like one sample per hour, that could be a big blocker for us. So then we invest into like new IP, new engineering to make it work. For simulations as well, so for simulations, for certain things, there's already a really good code base. So we don't have to improve it.

• 41:44Speaker 02

  Yeah.

• 41:44Speaker 00

  But for certain things, it's just not there yet or it's not scaled enough. And then we invest in the software engineering expertise to make it happen. And on the LLM side as well, which, yeah.

• 41:54Speaker 01

  Yeah, I mean, I think right now we spend very little time trying to improve software engineering, right? So we can use these systems to build up our code bases, help with these types of analyses. And we are absolutely beneficiaries of just how good codecs and other things have become. What we're targeting, though, are areas where the frontier is just not sufficient or we can do it vastly cheaper. This is where we're focusing our efforts. We deeply believe, and I think I've seen very good evidence, that when you actually start training against this data, you can make the reasoning more efficient, you can kind of add in new types of capabilities. And go beyond what inference time thinking can do.

• 42:36Speaker 01

  Compressing this knowledge into the weights, compressing these strategies into the weight is just fundamentally valuable.

• 42:42Speaker 02

  Yes.

• 42:42Speaker 01

  And that's why all the frontier labs continue to do this. And we didn't stop at GPT four and then just do inference time from there. Compression really needs to happen and you get different types of patterns that emerge. So for example, thing about tool use, there's declarative use of tools. Really, okay, this is how you use the tools. These are the specs, these are the inputs, these are the outputs you should expect. But there's a difference of that versus, you know, the procedural, like, you know, the intuitive knowledge of like, oh, I should use this tool in this situation.

• 43:15Speaker 02

  Yes.

• 43:16Speaker 01

  One of my friends had used the example of, you know, you could like read a manual on like how to play a saxophone versus learning how to play the saxophone. So really wanna, you know, build that into the weights.

• 43:25Speaker 02

  Hmm, how interesting.

• 43:26Speaker 01

  But don't duplicate things where the frontier is already sufficiently capable or cheap enough. We've had a few instances where because we're pushing the throughput of the lab so much, there are some analyses that just would have been prohibitively expensive. At one point, one of our engineers was coming to us and said, to do this analysis, our max throughput of this, our second lab, it would cost roughly $30 million a year. And so that was, you know, a large chunk. So we were looking at how do we actually compress this into smaller models, make it vastly more efficient. So you take it from like $30 million down to like, you know, like a sub million dollar cost. And that allows you to reinvest that capital elsewhere into other machines or training or simulations.

• 44:10Speaker 01

  So this is the type of work that we've been doing. We realized that we could do, conduct very high throughput experiments and they're actually fairly low latency. In a lot of areas of science, there can be lower signal to noise. They can take days, weeks, sometimes months. This is a huge motivation for like some of our starting points.

• 44:28Speaker 02

  Wow, interesting.

• 44:29Speaker 01

  So that allows us to build up this huge quantity of data and also maintain the quality, maintain the diversity, and that becomes this really incredible asset.

• 44:37Speaker 02

  Yeah, the diversity piece was something I was interested in, maybe because I, you know, don't have enough understanding of superconductors. But how do you, yeah, make sure you are doing these different modalities, like getting the data that is sufficiently diverse and not sort of overly specializing?

• 44:51Speaker 00

  Yeah, I'm sure there's definitely a trade off, right? There is an obvious trade off to consider, which is like, if you make it infinite, as diverse as possible.

• 44:58Speaker 02

  Yes.

• 44:59Speaker 00

  Now these experiments don't learn from each other. You haven't really come deep enough in any modality.

• 45:03Speaker 02

  Yes.

• 45:04Speaker 00

  Okay, if you make it really narrow, now there's a big problem because if that particular modality doesn't lead to a new superconductor, you're hosed, no matter how well you do, right? So we're trying to kind of walk this trade off correctly. And, you know, one of our guiding principles is actually pretty technical. We wanna make sure the different modalities we consider benefit each other. Because there are certain scientific experiments that are different attempts, like different synthesis methods, but actually getting good at one also has to get good at the other and vice versa. So then we can build our LLM training, simulations, kind of our hardware in such a way that they benefit from both of them. And that's been really good. I think that's how we kind of build our experimental diversity while also really

• 45:48Speaker 00

  synergistically benefiting from each experiment.

• 45:51Speaker 01

  Yeah, to use like ML parlance, we wanna have positive generalization.

• 45:54Speaker 00

  Yeah.

• 45:54Speaker 02

  Hmm. Does that stem from your intuition and sort of knowledge? Or is that also AI assisted? Like, where does the human have to play in there?

• 46:04Speaker 00

  So originally it was mostly human intuition, but we have seen indications of LLM actually contributing to this already.

• 46:10Speaker 02

  Yeah.

• 46:11Speaker 00

  So sometimes they will say, oh, you've done this experiment, this synthesis method. But you know, if you were to add this, now this is a new modality, but it could help. So it's not like a massive hop in the synthesis methodology space, but it's a small difference. And it will usually actually have a pretty good understanding of the literature. So like what kind of synthesis attempts have been done for this chemistry? And then it can also guide us that maybe you should try this next.

• 46:36Speaker 02

  What is the right way to monetize this sort of product? Like, does it become the case that you should develop, you know, these discoveries yourself and start selling them? Does it make sense that you're doing more contract work? What's the right model to start? And what does the right model at maturity look like?

• 46:53Speaker 01

  Maybe an interesting analog would be what we're seeing in software engineering. So in software engineering, when we first produced ChatGPT and these early models, we didn't promise full fledged output of like software. It's like, okay, we're gonna produce our Kubernetes replacement for you and we're gonna monetize that directly.

• 47:11Speaker 02

  Yes.

• 47:11Speaker 01

  It was a co-pilot to accelerate a software engineer and get them to solutions more quickly. Now, fast forward a few years, we're starting to talk about monetization of outcomes. So like sort of a new concept in AI. And I think a similar thing is going to unroll here. So the tools we've been building up in our own discovery loops are hugely valuable to many different folks in industry. And we're basically able to create custom systems for partners, for customers, to help them get to their solutions much more quickly. But in parallel, we also are kind of talking about these outcome based pricing, where then we retain the IP and like sort of the upside of the final output.

• 47:56Speaker 02

  Yes.

• 47:56Speaker 01

  Of our systems.

• 47:57Speaker 02

  Ah.

• 47:58Speaker 01

  There's also another interesting piece, like, so for example, if the software has in its weights, the recipe to a room temp superconductor, you don't wanna monetize it through, say, a software deal, you wanna retain that IP.

• 48:10Speaker 02

  Yeah.

• 48:10Speaker 01

  So those are sort of the trade offs that we have to balance.

• 48:14Speaker 02

  Are there analogies from other industries that you end up referring to? You know, I imagine that some part of the current scientific apparatus is useful to learn from, but I could also imagine different forms of manufacturing have taken this much further, whether that's, you know, creating a computer or an iPhone, or, you know, a car, whatever it might be.

• 48:34Speaker 00

  I mean, so one thing that's something very relevant to us is the drug discovery field. You know, there was a time where if you discovered a drug, you couldn't make a lot of money from its IP. But that changed, like today, you can discover a drug, maybe go through some clinical trials and then get paid billions or dollars just for the IP of that. So the field kind of changed and their capabilities changed for that market to exist. I think materials is kind of at that point now, like you can make a lot of money by selling materials, but materials IP doesn't really capture the value it deserves. But it could go through this transition similar to drug discovery. And we feel like as predictions get better, IP will have more value. So that's why we're really trying to close the loop all the way from like the discovery

• 49:19Speaker 00

  to synthesis and scale up of synthesis. So that's one of the analogies, I think.

• 49:24Speaker 02

  Before we, you know, press the record button here, Doze, you were mentioning the sort of difference between LLM performance on mathematics versus science.

• 49:33Speaker 00

  Right.

• 49:33Speaker 02

  Yeah, maybe you can tell us a little bit about that.

• 49:35Speaker 00

  Yeah, yeah. So this is something we're very passionate about and something that we kind of like predicted and kind of one of the reasons we started the company, which is that, you know, LLMs, if you talk to them, they're incredibly intelligent, right? You're blown away. And on certain things, they, their performance is insane, like coding, math, theoretical computer science, amazing. But then you haven't seen that kind of impact on the physical world. Why is that? And, you know, we feel like one of the reasons is with math, theoretical computer science and coding, the reward is instantly verifiable. Like you write code, you can immediately do a unit test, you have a mathematical theorem, you can validate it. And another reason is all the context is available to the LLM. Like if you're writing code, all the code the LLM can see. If you're doing math, all the axioms and all the corollaries you can see.

• 50:21Speaker 00

  The physics is, it will never be like that, right? Like, for example, when I do an experiment, there are more than an Avogadro's number of atoms in there. There's no computer that can store all the positions of atoms. So we have to come to terms that in real life, LLMs will never have full context. But to be honest, like outside of math, almost everything that's interesting is of this type where the full context will never be available to you. Like, think about politics, finance, literature. Like, what makes life interesting is when you don't have all the context, but you have to make decisions under uncertainty. Physics is very much like that. So if LLMs will actually play a big role in the real world, in the physical space,

• 51:00Speaker 00

  it has to figure out how to deal with missing context, not instantly verifiable rewards, some noise. And we feel like there is a very exciting future to be part of. So that's why we're working on this.

• 51:12Speaker 02

  I love that. That's so interesting.

• 51:13Speaker 01

  And it's like the strategies that are optimal under math, the strategies optimal under software development may not be the ones for the sort of decision making under uncertainty.

• 51:24Speaker 02

  Yes.

• 51:24Speaker 01

  And so that's the vision we're pushing forward here.

• 51:27Speaker 02

  To what extent is periodic's fate sort of tethered to that of the foundation models? Like if we, this is clearly not going to be the case, but if we didn't get another model from this point forward, like how far do you think you could push things?

• 51:40Speaker 01

  I mean, I think from the software engineering side, we could literally stop today and find every existing Nobel Prize, at least from the physicals. I mean, a lot of the physical sciences are relying on simple analyses. It's not competitive programming.

• 51:55Speaker 02

  Hmm.

• 51:55Speaker 01

  So I think from that perspective, it's very good. So I don't, I mean, we are huge beneficiaries of that and I hope they continue to improve rapidly and I expect they will.

• 52:04Speaker 02

  Yes.

• 52:05Speaker 01

  But I view that as like not a blocker for us anymore.

• 52:08Speaker 00

  Yeah, absolutely. I mean, there's so much to do on experimental analysis, simulation, scale up, data analysis, and the current models and our models are already good enough. Let's say like AI is very popular today, but let's say tomorrow it's not for some reason. There are certain things that happen in physics that will never go back. So one of them is force fields. Have you heard of this term before?

• 52:27Speaker 02

  No.

• 52:28Speaker 00

  So, you know, atoms are, like especially the atoms that we deal with are governed under simple quantum mechanics. And you can do quantum mechanical calculation to calculate the forces on each atom, or you can have an empirical model that tells you how much force will be on each atom. And it's called a force field. It's a very old topic. We use it all the time to model materials, molecules, et cetera. Force fields. Changed so much because of GNNs, like graph neural networks, that I don't think it will ever go back because now we can do very general across the periodic table models. In the past, you could only do like three to five elements. Tough days. So I feel like machine learning has made such an impact today that it will never go back.

• 53:09Speaker 00

  And as Liam said, today's capability is already enough to really advance science and engineering. But of course, we suspect that they'll actually keep getting better and then we'll keep benefiting from it.

• 53:20Speaker 02

  You know, clearly, I think we all agree that these models are just going to keep getting better. So planning with that in mind, what changes for you? What do you expect to get unlocked that isn't unlocked for you now?

• 53:31Speaker 01

  Some things that we're really excited about is just really advanced software engineering. We do a lot of very advanced simulations and we're starting to do a lot more work where we construct more advanced simulations that we can control every like bit of code. And these are things that would have taken teams, you know, a year or two.

• 53:52Speaker 02

  Yes.

• 53:52Speaker 01

  Now it's like an engineer's weekend time. And so I think this is something that we're extremely excited about. And especially in an area where the AI system is able to write software, write simulations that are maximally consistent with like experimental evidence.

• 54:09Speaker 02

  It still feels like what you're trying to do feels very differentiated in this market. Like, you know, this is at the moment max mimicry in many cases, right? You know, someone comes up with one good idea and there's going to be sort of five other folks. And that's probably healthy in many respects. How long before you think that starts to happen to you? And who enters the fray? Like, do the big foundation model companies start to say, actually, this is like invaluable data and we need to be spinning this up ourselves? Like, where does that competition come from in your view?

• 54:40Speaker 01

  I would suspect in the next couple of years, a system that can literally engineer matter is going to be of key importance to everybody.

• 54:49Speaker 02

  Yes.

• 54:49Speaker 01

  And I think that's sort of the motivation for moving as quickly as possible, building out a huge network of labs, of infrastructure, building simulations and AI systems that can control and orchestrate this then. That's our belief. I'm curious here.

• 55:04Speaker 00

  I mean, I do think that like we're pretty unique in the team we gathered and the approach we're taking. But taking a step back, there are now actually a lot of people working on this. Like almost every week now I hear like

• 55:15Speaker 02

  Oh, really?

• 55:15Speaker 00

  A company in the air for science space, which is great. Again, like if you really believe that science is endless, we really want just more and more smart people, dedicated people to work on this stuff, which is great.

• 55:25Speaker 02

  Yes.

• 55:25Speaker 00

  But yeah, I mean, like, looking back, I mean, the, maybe one thing is people are seeing how much LLMs can change the digital world. And they kind of wonder how much it could change the physical world. So it's not a surprise to us that a lot of people are getting into this field.

• 55:40Speaker 02

  You two are co-CEOs, right?

• 55:41Speaker 00

  Yeah.

• 55:42Speaker 02

  How do you sort of divide and conquer? What are the swim lanes that you naturally take?

• 55:46Speaker 01

  So, you know, very naturally, based on our expertise, I spend a ton of time on the AI LLM systems. I think that's been like a really core piece of my thinking on a day to day basis, as well as thinking about like, okay, how do we monetize these AI systems in the real world?

• 56:05Speaker 02

  Yeah, that there's sort of a natural division there. Yes.

• 56:08Speaker 00

  Yeah.

• 56:08Speaker 02

  I always like to wrap up with sort of some more abstract questions. Be curious what both of you would say to this, given what you do. If you had unlimited resources and no operational constraints, what is an experiment you'd really like to run?

• 56:20Speaker 00

  I would love to do reaction calorimetry. Like, there's a experiment you can do that tells you the formation entropy of a crystal. So what that means, like, how much energy would it take to take this crystal and remove it? Like, remove the bonds and remove the connections. The reason that's really nice is that's a single number you can get, like formation entropy at some temperature that you can directly compare against simulation. It's usually hard to find like an exact comparable between simulation and experiment. It's usually like convoluted narrative based, but here it's direct. But it's just like a very complicated, expensive experiment. But we had infinite money. I would love to do that, like thousands of them. So there's some data from NIST, you know, the Standards Institute in US.

• 57:05Speaker 00

  And they've done these experiments for hundreds of materials. And it's like invaluable for solid state chemistry field. So if you had infinite resources, I'd love to just really scale that up and get so much formation entropy data that we can compare with simulation and machine learning.

• 57:20Speaker 02

  Amazing. What about you, Liam?

• 57:22Speaker 01

  I would love to see if we had access to a full fab, how quickly you could ramp up, say, like a new generation of chips. So like, how quickly could you overcome the materials engineering problems of like, of logic, of memory, of processing? And you say, okay, here's the new generation. How can we take the yield that would have taken like many, many months to ramp up and accelerate that significantly? So I think those would be incredibly interesting experiments.

• 57:51Speaker 02

  Fascinating. Okay, last one. If you could recommend a book to everyone on Earth, what would you want to recommend to people?

• 57:59Speaker 01

  I would say, probably a popular one in Silicon Valley, but The Beginning of Infinity by David Deutsch.

• 58:06Speaker 02

  That is a popular choice to this question. I still have not read it. I don't know what my mental resistance is to it.

• 58:11Speaker 01

  I really loved it.

• 58:12Speaker 02

  Yeah.

• 58:12Speaker 01

  Yeah, really. Yeah, I think it's a very much like a spirit of optimism. You know, it kind of goes to Doris's point about like the unboundedness of science. So yeah, I really enjoyed that one.

• 58:24Speaker 02

  Yeah, I think I read part of it at some point. Yeah, anyway.

• 58:27Speaker 00

  I really like this book called Subtle is the Lord. It's a biography of Albert Einstein by one of his friends, who's also an exceptional physicist, Abraham Pace. And the reason this is unique is, you know, most biographies are written by people who are not an expert in the topic that they're writing the biography of. And there's usually like a mismatch in terms of like what they understand and what they should say. This book is incredible because it goes through Einstein's life, but with the very precise technical detail about what Einstein did. So it's like an incredible friendship book because they're friends and like this guy wrote this incredible book for his friend. But it's also incredible physics book. So if you're interested in friendship or physics, it's

• 59:06Speaker 01

  That's a great combo. I love that. Amazing. Liam, Doj, thank you so much. This was a real pleasure.

• 59:12Speaker 00

  Thanks for having us.

• 59:13Speaker 02

  Super fun, yeah. That's it. Thank you for listening to this episode of The Generalist Podcast. Please subscribe on Apple Podcasts, Spotify, or your preferred podcast app. Ratings and reviews help others discover these discussions. So if you enjoyed the conversation, I'd be grateful if you could take a moment to leave one. For all past episodes and more, visit us at thegeneralist.substack.com. See you next time as we continue to explore the future.