Inside a $300 million bet on AI for physical R&D

A big problem with using artificial intelligence to discover new materials? It struggles to predict beyond its training data. That means AI might be better at optimizing known materials than discovering entirely new ones — like a room temperature superconductor or carbon-capture sorbents. But since we last covered the topic in September 2024, a few things have changed. OpenAI released its powerful O1 reasoning model. Large language models have also gotten better at math, physics, and coding. And lab automation — robots mixing liquids and powders, running characterization tests — has improved, allowing for a higher volume of experiments. So, can these improvements overcome AI’s training data problem? In this episode, Shayle talks to Ekin Dogus Cubuk, cofounder of Periodic Labs, which raised $300 million seed round in September. Last year, Dogus took a more cautious view on using AI for materials discovery. Now though, he’s convinced there’s a clearer path forward for physical science research and development, especially materials discovery. Shayle and Dogus cover topics like: Creating experimental and synthetic data to overcome AI’s limitations of predicting beyond its training set Why we should focus on breakthrough discoveries over easier, incremental wins The different roles humans and AI play in the discovery process Period’s focus on automated experimental labs using AI-generated hypotheses Resources: Catalyst: Can AI revolutionize materials discovery? Latitude Media: This ‘superintelligence platform’ just raised $200m in seed funding Latitude Media: Can AI get us closer to fusion? The New York Times: Top A. I. Researchers Leave OpenAI, Google and Meta for New Start-Up Credits: Hosted by Shayle Kann. Produced and edited by Daniel Woldorff. Original music and engineering by Sean Marquand. Stephen Lacey is our executive editor. Catalyst is brought to you by EnergyHub. EnergyHub helps utilities build next-generation virtual power plants that unlock reliable flexibility at every level of the grid. See how EnergyHub helps unlock the power of flexibility at scale, and deliver more value through cross-DER dispatch with their leading Edge DERMS platform, by visiting energyhub.com. Catalyst is brought to you by Bloom Energy. AI data centers can’t wait years for grid power—and with Bloom Energy’s fuel cells, they don’t have to. Bloom Energy delivers affordable, always-on, ultra-reliable onsite power, built for chipmakers, hyperscalers, and data center leaders looking to power their operations at AI speed. Learn more by visiting BloomEnergy.com. Catalyst is supported by Third Way. Third Way’s new PACE study surveyed over 200 clean energy professionals to pinpoint the non-cost barriers delaying clean energy deployment today and offers practical solutions to help get projects over the finish line. Read Third Way's full report, and learn more about their PACE initiative, at ⁠www.thirdway.org/pace⁠.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:02Speaker 00

  Latitude Media, covering the new frontiers of the energy transition.

• 0:07Shayle Kann?

  I'm Shayle Khan, and this is Catalyst.

• 0:11Speaker 02

  I have to say, there's a difference between winning gold medals in math Olympiads and scientific discovery. Like, you can practice for math Olympiads by studying previous year's problems. You can't really practice how to discover the next big theory, but they were getting better at reasoning on complex problems.

• 0:33Shayle Kann?

  Coming up, can AI discover a room temperature superconductor? Volume two.

• 0:48Speaker 00

  Demand flexibility is now an essential part of the data center's speed to power playbook. But how do data centers actually curtail load? Grid Beyond is the end-to-end solution for flexible energy management. Grid Beyond's advanced DERS and VPP platform is tailored to what data centers care about most: maximizing reliability, intelligently meeting curtailment obligations, reducing costs, and managing all your distributed resources in one place. Ready to transform your distributed energy management? Get started at gridbeyond.com. Catalyst is brought to you by Antenna Group, the communications and marketing partner behind the most influential companies in energy, climate, and infrastructure. For three decades, Antenna has helped ambitious companies stand out, build market leadership,

• 1:32Speaker 00

  and become category-defining brands. From breakthrough startups to global industry leaders, their team brings the expertise, relationships, and track record to accelerate your momentum. Learn more at antenna group dot com.

• 1:46Speaker 03

  The power plant for the modern grid isn't a building, it's a network. Energy Hub's DERS platform connects batteries, EVs, thermostats, and commercial and industrial sites into virtual power plants that over 230 utilities dispatch as reliable grid capacity day in, day out. VPPs deploy in 60 to 100 days at 40 to 60% less cost than traditional infrastructure, and they perform like conventional generation from day one. Learn about the flexibility platform for the modern grid at energyhub.com.

• 2:21Shayle Kann?

  I'm Shayle Khan. I lead the early-stage venture strategy at Energy Impact Partners. Welcome.

• 2:27Shayle Kann?

  So, a year ago, a little over a year ago, I had Doge Chubuk on this podcast to talk about using AI for materials discovery, which has all sorts of interesting applications in the spaces that we talk about here. At that time, Doge had been leading efforts in that area for Google DeepMind for some time. And I thought of him as being both very knowledgeable in the space, obviously, also pretty sober about it. Fast forward a year, Doge left Google DeepMind earlier this year and along with Liam Fetis, who was one of the co-creators of ChatGPT, started a company called Periodic Labs, which raised, wait for it, a $300 million seed round led by Andreessen Horowitz. Periodic is doing AI for materials discovery.

• 3:11Shayle Kann?

  And not just that, also physics and chemistry, and they're also very much hardware in the loop. The way I like to frame it is that they are building two kinds of frontier lab at once. There's a frontier AI lab and a frontier scientific lab, the type of lab that we used to talk about. And then they're trying to make those two things work together to make breakthrough discoveries. Notably, one thing we talked about last time was how the AI materials discovery companies at the time tended to start by going after, often, discovery of something like metal-organic frameworks or MOFs for carbon capture, which I think of as less of a breakthrough opportunity, really, from a global scale. Whereas the big, perhaps biggest breakthrough to prove would be the discovery of a room-temperature superconductor.

• 3:54Shayle Kann?

  Well, Periodic makes no promises, but they're very publicly working on high-temperature and maybe room-temperature superconductors. Based on that last conversation, to be honest, I wouldn't have predicted this. So it was time to have Doge back on and hear what changed and how. Here's Doge. Doge, welcome back. It's great to be back. It's great to see you. A lot has changed since the last time we talked. So I looked back. So the last conversation that we had was just over a year ago. It was September 2024. And I was having you explain to me the wild world of AI for materials discovery in particular, and the work that you'd been doing at Google DeepMind, but also just like the broader landscape.

• 4:36Shayle Kann?

  And I'll tell you my takeaway from that conversation, which you could tell me if I had the wrong takeaway at the time. But my takeaway was... Promising field, pretty unclear if and when this new wave of L L Ms and the all the reinforcement learning, all the things that are that have shown up in the past few years, pretty unclear if and when that would generate a real meaningful breakthrough discovery in materials. And we we talked through a bunch of the reasons why it's challenging training data, maybe chief amongst them. But but I came away with a pretty, I think, like a sober view of the the path there. Okay, so fast forward a year and you left

• 5:18Shayle Kann?

  Google DeepMind, started a company to do that amongst other things. So I guess the first question that I have for you is, what changed in the last 12 months to make you to give you conviction that like now is the time?

• 5:31Speaker 02

  Great question. So when we talked, I was doing research in the field of computational material science and machine learning. You know, specifically, we were using graph neural networks, we were using density functional theory, and we were trying to discover materials. One thing that changed since our discussion was the L L Ms have improved even further. So at the time, I wasn't using L L Ms much at all. But I think right around when we were talking, the O one came out, right? The reasoning models started showing up. And that was a huge update for me, because you might remember that one of my big concerns

• 6:13Speaker 02

  is, machine learning works best on the training set distribution. But in science and technology, we almost only care about auto domain generalization. So what O one showed is, if you spend test time compute, you can get better results. So that was very exciting to me, because there was one way of investing resources that was beyond the training set.

• 6:39Shayle Kann?

  So, okay, if I can try to translate that into layperson terms, the reasoning model, like OpenAI's O1 model, introduced, unlocked a door, kind of, that maybe allows you to break this challenge of the limited training data set that you have in materials discovery. That was what we spent a lot of time talking about a year ago. It was like, you know, you can compare the corpus of data that an LLM trains on to do language, which is enormous, it trains on the internet, basically, versus the corpus of data that you were dealing with in trying to discover novel materials. And it was thousands of data points, not tens of billions or whatever.

• 7:24Shayle Kann?

  And so that presumably hasn't changed, at least yet. But you're saying that the reasoning models have gotten good enough that they are able to sort of get around that challenge via reasoning or possibly generating their own synthetic data? Like, what is it that allows them to break that?

• 7:41Speaker 02

  So I'm not saying that they're good enough already, but that was one step in the positive direction. And another thing, you know, we've seen is they've gotten really good at math. So, you know, since last time we talked, they started winning gold medals in math Olympias. And they're doing similarly well on the coding, really well on physics Olympias. And I have to say, there's a difference between those things and scientific discovery. Like, you can practice for math Olympias by studying previous year's problems. You can't really practice how to discover the next big theory. But it did show you that they were getting better

• 8:24Speaker 02

  at reasoning on complex problems. So then what else do we need? So I think the biggest thing we need is to have our own lab. Because once you have a very intelligent reasoning LLM, you still can't discover things unless you make trials, right? Like, just like humans, the LLMs will be wrong often when they try to predict things outside of their training set. But you try many things and then at some point you get a really cool discovery. And this is, you know, as we talked about history last time, this is quite common in solid-state chemistry, solid-state physics, where a lot of discoveries happen somewhat by accident, but of course with a lot of background, understanding of the physical

• 9:08Speaker 02

  system and a lot of trial and error.

• 9:11Shayle Kann?

  So, okay, so this is what you're doing at Periodic, right? You're sort of combining the digital domain with the physical domain. You have a lab in both sense, it's a frontier lab in an L L M sense and a frontier lab in a laboratory sense, in the traditional sense of the word, and you're sort of merging the two. I'm curious in practice, like, how you imagine that that feedback loop working? So is it a traditional, you develop a theory, you run an experiment, you generate data from that experiment, but in this case you feed the experiment back into your customized L L M as an additional set of training data? And then that's the way that the loop works? Or is it more complicated than that?

• 9:49Speaker 02

  Yeah, exactly. I mean, it's pretty simple, I think, as you said. So the L L M can propose, for example, synthesis recipes, or it can propose simulations to run. And because the L L M's are pretty good at tool use, it can actually do it itself. And then you get some results back. So the results from experiment could be some characterization data, results from the simulation can be some, you know, trace of some simulation you did. And now the L M can go through it with the context of its previous training, maybe the context of relevant papers, textbooks, but also now the results that it just got that no one else have ever seen. And then now it can kind of tweak the experiment, tweak the simulation for the next step.

• 10:28Shayle Kann?

  Right. You said one thing in there that I guess is worth pointing out, like, you're trying to automate this as much as possible. The L L M might run the experiment.

• 10:36Speaker 02

  Yeah, absolutely. I mean, one of the other advances that's been happening recently that I think made Periodic possible is the high throughput.

• 10:46Speaker 02

  Experiments have been getting better. You know, there are many examples of this now across academia, industry, where these robots, they became quite commoditized, actually, just mixing powders or mixing liquids and then sending it to characterization. I think one thing that isn't as advanced right now, but we feel like we can do pretty soon is automated characterization itself. So you mix powders, you put it in some characterization tool, you get the result out. What is the actual output? I think this is pretty difficult right now for AI tools, but we feel like we can improve it pretty quickly.

• 11:25Shayle Kann?

  I want to talk about one specific application that I know you're going after that we actually did talk about a year ago. But also, I want to talk about it as a way to see whether one of the other things you described as a fundamental challenge is this change, which was, as I understood it, AI being pretty good at the next incremental discovery, but but not necessarily good at the breakthrough discoveries, you said, through history. Usually that's done accidentally, or often it's done accidentally because it's not, you can't like reason your way to to this massive breakthrough discovery. So so let's talk about superconductive materials. So, right, we talked about this last time where all these companies that existed at the time that were doing AI for materials discovery were starting on things like discovering

• 12:08Shayle Kann?

  a novel moth for carbon capture or whatever. And but we said like the thing that would be the real breakthrough, the big thing, would be a room temperature superconductor. And you guys have since launching been very public about like superconductive materials, maybe not room temperature. I'm curious for you to tell me how likely you think that is, but like high temperature superconductors is on the roadmap. So why, first of all? And then and then, second of all, like this question of Do you think you have a path to the truly breakthrough? What would the path be to a truly breakthrough discovery as opposed to finding something that is a material that is superconductive at a ever so slightly higher temperature than the best that we've got today?

• 12:50Speaker 02

  Yeah, so to answer your first question, I think it's still true that it would be difficult to just reason your way into a much better superconductor. I actually would guess that there's a law out there that we haven't discovered yet that says that you can't just look at your training set as different than what you're trying to discover and just predict it. You know, there's been rules that we discovered from 1800s on where like you connect energy to work. So thermodynamics is the first example. There's more recently Landauer's limit, which shows that you have to spend a certain amount of energy to delete information, which can be used to describe Maxwell's demon contradiction.

• 13:34Speaker 02

  I bet there's something similar for how hard it is to discover things that's outside of your training domain. Okay, so I don't think that's been fixed since last time we talked. But because we have a lab internally, we can just try things and try them at high, large scale and often, and hopefully as intelligently as possible. So even though we won't reason our way into a much better superconductor, we'll be able to push our trials in the direction that's most promising or, you know, most promising for us, given our training set at the time. So, yeah, I think that hasn't changed. And I think there's reason to be hopeful because, you know, in the big scheme of things, it's a pretty new field. I mean, if you look at cooperates, they were from 1985.

• 14:19Speaker 02

  There's been a lot of advances, you know, more recently. So, yeah, we're very excited. One reason we chose superconductivity is if you find a good superconductor, that's impactful immediately, right? Like, last time we talked about how long it can take to translate materials improvements into products. One nice thing about superconductors is if somebody discovers a room temperature superconductor today, even before it makes it into a product, it has huge impact, right? Like, first of all, it changes how we think about the universe. Second, it helps us do physics experiments that wasn't possible before. And you know, whenever you think about like a sci fi-ish technology like quantum computing, fusion, uh, superconductors come up because it's kind of what we need.

• 15:02Speaker 02

  It's kind of like the um, one of the most exciting macro scale quantum phenomena. Um, so that's one reason we picked it because it kind of is exciting as soon as we succeed. The other reason is it requires all sorts of improvements to get there. You know, when we think about OpenAI and DeepMind, I remember back in two thousand and sixteen, people used to make fun of these institutions for prioritizing AGI so much because they were saying we're gonna do AGI. But what happened is they developed so many other tools on the way to AGI that were useful in themselves. But today they have these LMS that, you know, you might consider AGI or like something really impressive. Superconductivity is a bit like that.

• 15:44Speaker 02

  To discover a exciting superconductor, we probably have to develop so many capabilities on the way there that's by themselves very useful. For example, automated synthesis, automated characterization, being able to model or predict high temperature superconductivity because we don't have a theory for it yet. So it's kind of like a nice goal that unites people and requires a lot of other useful things to happen on the way. Um, and it's one of those things that, you know, physicists find really exciting. So the physicists in our company are really excited by this mission, but also computer scientists find it very exciting. It's just one of those things that I think both sides can really appreciate. So those are some of the reasons that we picked it.

• 16:30Speaker 00

  For data centers, connecting to the grid quickly means two things: building distributed energy and having a plan for demand flexibility. Grid Beyond helps with both. Grid Beyond's Terms and VPP platform unlocks the potential of distributed assets and dispatches flexibility into the right market at the right time. And this is more than software. Grid Beyond supports developers every step of the way. From modeling your highest ROI energy mix to orchestrating the full set of assets around a data center. The track record speaks for itself. Grid Beyond manages over one thousand four hundred sites and more than three point seven gigawatts of load across Europe, the UK, and the US. Learn more at gridbeyond.com or click the link in the show notes.

• 17:12Speaker 00

  If you work in energy or climate, you already know the company's Antenna Group has helped build. For three decades, they've been the communications and marketing partner behind the industry's breakout companies, category creators, and market leaders. Today, their team is working across the most consequential areas of the energy economy, from data centers and grid modernization, to nuclear, geothermal, storage, critical materials, and next-generation technologies. Antenna helps ambitious companies become the names everyone in the market knows, trusts, and wants to work with. They build reputation, elevate executives, drive demand, and turn momentum into market leadership. If your company is ready to lead its market, Antenna is ready to help you get there. Visit antenna group.com.

• 17:56Speaker 03

  When you harness the power of millions of DERs in a VPP and put artificial intelligence behind the dispatch, the outcomes exceed conventional generation. Energy Hub's DERMs platform has been trained on years of real dispatch data from over 2.6 million DERs, so utilities can plan with confidence. That means consistent load reduction across a four-hour event window, even during the widest, most uncertain peak periods. Utilities target, shape, and verify capacity down to the meter. Unlock flexible capacity that costs 40% to 60% less to deploy than traditional infrastructure in a fraction of the time it takes you to break ground on a peaker plant. Learn more at energyhub.com.

• 18:39Shayle Kann?

  I guess back to this question of, how do you distinguish between the incremental innovation, which, to be clear, if you develop or discover a superconductor at a higher temperature than anything that we've got today, that's meaningful. But it's not, you know, it's probably orders of magnitude less meaningful than if you discover a room-temperature superconductor. And I presume that the scientific challenge is commensurately distinct between those two. And

• 19:09Shayle Kann?

  You know, the way that L L M's work, as I understand it, at least in part, is on these reward functions. And so are you, are you setting your AI system a goal of, you know, find a room temperature superconductor? And then everything flows back from there? Here are the steps and all the things we have to fix to get to room temperature? Or do you say, improve this characteristic such that we can incrementally, you know, build our way there? In other words, are you going to find 10 super? I think of it as sort of a different thing from, but like the alternative version of this is what happens in fusion, nuclear fusion, where everybody is sort of chasing this same goal of Q is

• 19:50Shayle Kann?

  greater than one, right? Like energy, energy break even. And everybody is getting incrementally closer and closer and closer. And eventually NIF breaks it, or somebody breaks it. Is it going to look like that? Or is it going to look like we've discovered nothing until we discover the room temperature superconductor?

• 20:05Speaker 02

  Well, so as you said, I think there are many different ways of improving superconductors without getting a room temperature superconductor. So one of them could be having a significant increase in T C. But another one could be a really high critical magnetic field, which turns out might even be more important for fusion applications than T C itself. Another one can be more mundane, like some mechanical properties, like a superconductor that also is ductile and, you know, we can make it into devices. So we wouldn't rule out, you know, all of these very exciting developments just for like a room temperature T C.

• 20:41Shayle Kann?

  But how do you set the reward function for your model? What are you, what are you optimizing it for then?

• 20:47Speaker 02

  I mean, I think that's an empirical question. I think one thing I should say is, it's quite nice because it's a bit, it's hard to reward hack. You know, one of these issues with R L and training L L M's is you might worry about reward hacking. And in simulations, again, reward hacking can be a problem, even in D F T. But for real life, experimental measurement of T C, it's much harder to reward hack, which we love. So like, if our reward was increasing T C, that just seems like a Nicer, unhackable reward. But in terms of like, what specifically will get us there? You know, we're not sure yet. I mean, it's an empirical question. We can probably try all of them. I'll, I'll, I'll list the things you propose and we'll try all of them.

• 21:29Shayle Kann?

  It, I guess that gets to the other question, which is like, what, what does the human in the loop look like here? Right? And, and again, as you, you said, like, if we haven't solved the sort of AI is good at incremental innovation and not orthogonal breakthrough innovation thing, but humans are historically at least better at it. Is it like folks on your team, uh, developing a theory of something? And that gets fed through the model and you get the results and you feed it back in, you see whether it's a promising category? Like, is the germ of the original idea of what to look for coming from a human? Or is it coming predominantly from the model and then the humans have to interpret and send it off in various directions?

• 22:14Speaker 02

  Yeah, I mean, that's a great question. You know, we, we're not really prioritizing full automation anyway. So if we get better results with humans doing part of it, that's great. Like, this is also actually a question for our lab, right? Like, do we want to automate every single aspect of the lab? At some point, you end up needing humanoids for that. And I think that's not like, Liam, my co-founder and I, we are trying to be very pragmatic about it. Like, our goal is to get the best result possible on the things we care about. And, you know, how much of the automation comes from the ML? How much of it comes from more traditional tools? And how much of it gets done by humans? I think that's kind of, again, an empirical question.

• 22:56Speaker 02

  So, yeah, we're not, like, I think, as you said, it does seem like today, there are things that ML, AI is better than humans. But one of those things is not hypothesis generation. So, uh, I mean, there are two options. We either have to improve these algorithms and hypothesis generation, which is possible. Or the other option is we have humans providing some of the hypotheses and then AI doing the execution.

• 23:21Shayle Kann?

  I guess the other question here is cost. I mean, you guys raised a $300 million seed round. So that implies on the outside that your cost structure will look similar to other companies that basically have to, are going to use just an enormous amount of compute. And so like, a lot of that cost comes from. Compute, in your context, I could imagine maybe that being true, but also maybe that not being true because you, again, you just don't have the same corpus of data. You can't build a 10 billion parameter model right now because the data isn't there to do it. And so instead, that cost is going to go more toward the robotic lab and all that kind of stuff. Like, how should I be thinking about how much compute you'll use and where that costs come from?

• 24:03Speaker 02

  Yeah, so honestly, compute is very expensive and we are going to train L L Ms. We are going to use GPUs to run simulations, so that does end up being a large part of the cost. Yeah, it's funny that before, you know, if you asked me this question 10 years ago, I would have thought that the biggest part of the cost must come from the labs because like physical is real, you know, you're building this lab, you're buying instruments. But turns out the GPUs are so expensive and training L L Ms is so expensive. So when we were thinking about how much to raise, we kind of laid it out in terms of the GPU cost, the lab cost, and this was kind of the minimum number we felt like was viable. And yeah, we'll see, the GPUs have been getting more expensive recently.

• 24:46Speaker 02

  I guess we'll see how the market dynamics continue.

• 24:51Shayle Kann?

  To what extent are you, do you end up building generalized model or models versus models designed to a specific domain, even a specific scientific domain, right? Like, you guys are doing your material discovery, obviously, but physics and chemistry and these things all intertwine, but like, is the same model going to be equally capable across all domains? Is that the intent? Or is that just not how they're supposed to be architected?

• 25:16Speaker 02

  That's right, and that's actually something we're very excited about. You know, one thing I've been kind of noticing is like in the past, say, three, four years, I had a chance to collaborate with very, you know, world-class, best-in-their-field scientists. And even when you work with them, you realize that While their expertise on a few domains is, you know, incredible, maybe best in history, there's just so much more to know in chemistry and physics that they may not know all the other aspects of it. So this is why I brought up superconductivity, because you might actually need to be really good at, you know, cell state chemistry and synthesis of difficult novel materials, just because, you know, you don't know which chemistry the superconduct is gonna come from. Some of these ideas you might have may not be as stable thermodynamically,

• 26:01Speaker 02

  so you need to be intelligent about how to kinetically force it into that phase you want. But at the same time, in addition to cell state chemistry, you need to be incredible at condensed matter physics, right? Because like, there's so many different kinds of superconductivity. We don't understand most of them very well. And there's nobody in the world who knows both of those equally well, or like sufficiently well. And turns out this is true for many different aspects, right? Like if you need to use robots for high tropo synthesis, again, like, there are only so many people who understand robots and how to use them for synthesis. So I think this was different in eighteen hundreds, probably. Like, there was probably a time when a physicist could contribute and be one of the best in the world on many fields of physics.

• 26:47Speaker 02

  But it's definitely not true today. And this is one of the reasons I think we are very excited about L L Ms, because when you talk to them, they seem like, they have a pretty good understanding of cell state chemistry and cell state physics at the same time already. And we're, you know, trying to improve them further in the physical sciences specifically, because that's where we are really interested in. And then we're hoping that they'll be, yeah, good at multiples of these. And then a really exciting prospect with that is, a lot of the exciting discoveries happen to lie in between fields, right? Like, that's why it's sometimes easier to be interdisciplinary. And there's so many of these surface areas between these different fields. It's like, I guess, science is kind of like a fractal in the way it's hierarchically organized.

• 27:32Speaker 02

  And there's so much, there's so much surface area that humans have exploited, of course. But then there's probably so much left to exploit. And we're excited about an L L M that can basically do that at a scale that humans couldn't yet.

• 27:46Shayle Kann?

  How good are the LLMs today or the best in class of what you guys have at generating synthetic data in this domain? And another way to ask this question is, you know, if you fast forward three years, you're fully up and running and you're operating, how much of the valuable insight you will generate, do you think will come from the the physical data coming out of your lab versus the synthetic data that the LLMs create on top of that?

• 28:17Speaker 02

  Yeah, that's a great question. I obviously don't know the answer, but it's great to brainstorm about that. Because on the one hand, the lab data will be kind of our additional data that other LLMs may not have. And you might think then the lab data will only be as valuable as the results in it. But on the other hand, what's interesting about scientific data is it's not just a few bits of numbers, right? For example, there's certain experiments you can run where the result you get from it is just, say, three floating point numbers. But the implications of those could be tremendous, right? It's not just gonna be like a few bytes.

• 29:00Speaker 02

  It will actually be potentially an incredible amount of understanding just from a few experiments. And this has been how it is in human history, right? Like there's certain experiments that told us so much about how we understand about the universe. And the way to do this with synthetic data can of course be, you know, you run simulations that relate to that experiment. And when you get the experimental result, that actually validates or refute so much of the simulations you ran. And then that is a lot of information in itself. So, you know, it's a very interesting question. I think there's some actually differences about how you think about synthetic data when it comes to an LLM that's good as science. And exactly, I mean, this is one of the reasons I really want to work on this because this

• 29:44Speaker 02

  opens up questions for LLMs and LLM training that's maybe different than what the frontier labs are thinking about right now. You know, if they're only thinking about math and logic and kind of what's on the internet, like accounting tasks, that's a bit different than if you're trying to do experimental physics, experimental chemistry. It just seems like a very exciting question to explore.

• 30:05Shayle Kann?

  I want to talk a little bit about how you build a business out of this. I mean, you mentioned the superconductor example and you said like, there's a lot of value in this long before this novel superconductor falls, goes into a product. But ultimately, kind of has to go into a product of some sort for you guys. And I think we talked about this a little bit last time, too. There's this question, okay, so if your job, your core job, the periodic, is to discover new things that are going to be valuable in the world, say you do it. To my mind, there's sort of a binary decision you have to make at that point. Do you try to sell the discovery, license the technology, license the IP to somebody else who's going to go produce it and turn it into a product? Or do you produce it?

• 30:50Shayle Kann?

  Do you sell the product? Do you sell the discovery or do you sell the product? Do you have a prior on which direction you want to go here?

• 30:57Speaker 02

  Great question. And, you know, I think the two options can be correct depending on the context, depending on the timelines. But honestly, it also depends on like where we are in the company. So at the very beginning, you can imagine our LLMs will be very impactful for other companies doing physical R and D. Like, you know, already today, most people...

• 31:23Shayle Kann?

  Like contract R and D, basically.

• 31:25Speaker 02

  Yeah, exactly. I mean, there's a lot of interest in being able to use these LLMs. Sometimes the data restrictions don't allow it because you don't necessarily want to put your data on an LLM, you know, on the web. Sometimes the other issue is you haven't trained the LLM on your data, so it's actually not as good as it could have been. That kind of improvement could be really impactful because we've seen how impactful LLMs can be in other fields where they have access to the data. So there's a lot of, I think, headroom for impact there. But in the longer run, you can also imagine a case where we, as a field, get really good at designing materials intentionally. You know, that hasn't been the case.

• 32:08Speaker 02

  But if you look at drug design, there was a time when designing drugs wasn't very profitable. And I think people would look at it and say, this is not a good business. But what happened with Genentech is... The field got so good at designing drugs that at some point it became very valuable itself. The machine learning field has been making huge improvements in material science that, you know, was hard, kind of hard to predict. So it'll be interesting to see how far that goes and whether materials discovery by itself becomes a very exciting business similar to drug discovery. But for us, you know, we already see this big need and the big potential for impact by providing these L L M's to do physical R and D.

• 32:53Shayle Kann?

  Yeah, almost like a, this is gonna be the wrong analogy, but it's partially right, like almost like an A W S. Like you're gonna have the infrastructure. In this case, the infrastructure is your custom designed L L M that is smart about physics and chemistry and all these domains and also your physical lab and them being interconnected with each other. And so you have all this infrastructure and scale in that infrastructure that you can use to go convince whatever large company that's doing R and D that they should just be outsourcing it to you rather than building, rebuilding the whole same thing in-house, which is not exactly what the cloud providers are, but you know, there's enough of a relationship there. So that, that feels right, but it is,

• 33:36Shayle Kann?

  I suspect, yeah, I guess this is what you're saying. It is, I suspect a smaller ultimate opportunity than the you proactively discover a bunch of novel materials that change the world and then however you monetize them, you prove you're able to do so repeatedly and then you're Genentech and you know, it's a, it's a whole different category.

• 33:58Speaker 02

  Yeah. And, and you know, one other thing you see is people are so excited about this, you know, they wanna see L L M's not just conquer the digital world, but also, you know, really impact the physical world and, you know, impact the atoms, basically. So I feel like this has to be done. And the, the team has been very excited. We're Like it's it's really amazing. We're hosting weekly seminars where the physicists will teach the computer scientists about the physics and the computer scientists will teach the physicists about L L M's. And of course, there are a lot of people in between, right? Like, it's actually, again, like a fractal. So, yeah, I think there's been a lot of excitement about seeing if these technologies can be used not just for the digital world, but also for constructing the atoms around us.

• 34:42Shayle Kann?

  I I guess maybe the last question, you you said at the beginning, the thing that changed between a year ago and now, in part, was advancements in the in the big L L M's, right? The O one model and so on. Is there a next? Like, what could change in what could OpenAI release in a year or two years from now that would be a big leapfrog for you? Are you branching off now from what the big L L M's are going to do? And, you know, everything that all advancements are going to come from periodic? Or is there something else that they could offer that is a step function change in your capability to discover new materials or whatever?

• 35:23Speaker 02

  Yeah, great question. I I think we actually basically rise with the tide, right? Like, as L L M's get better, there's so many advantages of that to other applications. For example, the L L M's are getting very good at coding. And that's not surprising, because programming is a kind of closed environment. You can just simulate in your computer and get valuable feedback and then quickly improve. But as computers get better get better at coding, that's huge for science, because then you can run simulations more efficiently. The simulations themselves can improve. Similarly, with tool use, experiments. So I think as L L M's improve in general, that's going to help a lot with science applications.

• 36:08Speaker 02

  There are maybe longer term things that can happen. One of them could be, you know, things like hypothesis generation or more auto domain generalization. But then a question there is, will that come from status quo? Like how L L Ns are being trained now? Or will it come from actually labs that try to improve scientific reasoning for these L L Ns? Because then maybe hypothesis generation emerges naturally or auto domain generalization emerges naturally? Because that's what you're kind of trying to get at with your reward. So I think there'll be a very exciting question to see maybe next time we chat.

• 36:44Shayle Kann?

  Love it. All right. Thank you so much for taking some time again. Congrats on periodic. Super excited to see what you guys discover and for when your room temperature superconductor is shooting electricity all around the world around me.

• 36:57Speaker 02

  Okay, well, thanks a lot. It was great chat show.

• 37:01Shayle Kann?

  Doge Chubuk is the co-founder of Periodic Labs and a former researcher at Google DeepMind. This show is a production of Latitude Media. You can head over to latitudemedia.com for links to today's topics. Latitude is supported by Prelude Ventures. This episode was produced by Daniel Waldorf. Mixing and theme song by Sean Marquand. Stephen Lacey is our executive editor. I'm Shayle Khan and this is Catalyst.