打造AI物理学家：ChatGPT联合创始人的新事业

归根结底，科学是由现实世界中的实验驱动的。这就是我们在Periodic Labs所做的。我们采用这些先导技术，并认为，如果你关心科学的进步，我们需要将实验纳入循环中。

Ultimately, science is driven against experiment in the real world. And so that's what we're doing with Periodic Labs. We're taking these precursor technologies, and we're saying, okay. If you care about advancing science, we need to have experiment in the loop.

构建一个可以设计现实世界的AI物理学家（暂且这么称呼），其应用范围非常广泛。你可以将其应用于先进制造、材料科学、化学。任何需要与物理世界进行研发的过程，似乎都能从Periodic正在取得的突破中受益。

The applications of building an AI physicist, for lack of a better word, that can design the real world are so broad. You can apply them to advanced manufacturing. You can apply them to material science, to chemistry. Any process where there's R and D with the physical world required, it seems like we'll benefit from breakthroughs that periodic is working on.

例如，如果我们能发现一种200开尔文的超导体，即使在我们用它制造任何产品之前，能够在如此高的温度下观察到这样的量子效应，我认为这将极大地更新人们对宇宙的看法。

For example, if we could find a 200 Kelvin superconductor, even before we make any product with it, to be able to see such quantum effects on at such high temperatures, I think would be such an update to people's view of how they see the universe.

如果AI能从谈论科学转变为实际做科学会怎样？今天的对话邀请了a16z的普通合伙人Anjane Mitha，以及Periodic Labs的联合创始人Liam Vedas和Doge Chubuk。Periodic Labs是一家前沿研究实验室，致力于为物理和化学构建实验循环AI。他们探讨了为什么现实世界的奖励函数很重要，中期训练和高计算强化学习如何结合，以及为什么超导和磁性是实现AI物理学家的首要目标。他们还谈到了嘈杂数据集和负面结果，当ML研究人员与实验科学家并肩工作时会发生什么，以及短期回报。从半导体到太空和制造业的先进行业副驾驶工具。

What if AI could move from talking about science to doing science? Today's conversation features Anjane Mitha, general partner at a sixteen z, with Liam Vedas and Doge Chubuk, cofounders of Periodic Labs, a frontier research lab building experiment in the loop AI for physics and chemistry. They unpack why real world reward functions matter, how mid training and high compute RL fit together, and why superconductivity and magnetism are the first beelines towards an AI physicist. They also get into noisy datasets and negative results, what happens when ML researchers sit shoulder to shoulder with bench scientists, and the near term payoff. Copilot tools for advanced industries from semiconductors to space and manufacturing.

让我们开始吧。

Let's get into it.

那么，Liam，你是ChatGPT的联合创造者。Doge，你曾在DeepMind负责一些物理团队。谈谈你们是如何相遇的，以及是什么时刻让你们意识到必须离开这两个实验室来创办Periodic？

So, Liam, you were the cocreator of ChatGPT. Doge, you were running some of the physics teams at DeepMind. Let's talk about how you guys met, and what was the moment where you realized that you guys had to leave both of those labs to start periodic?

我相信我们八年前在Google Brain相遇，当时在翻转一个大轮胎。是的，在谷歌。

I believe we met eight years ago at Google Brain flipping over a large tire Yep. At the Google You

得给我们多讲讲这个

gotta give us more on that

故事是这样的。Google Rails是谷歌健身房之一，谷歌的设施。我想那就是我和Doge相遇的地方，那是一个巨大的轮胎，基本上一个人自己是翻不动的。所以Doge试图翻动它，他把我叫了过去。他说，哦，我觉得我们两个人应该能行。

story now. Google Rails was one of the gyms at Google the Google facilities. And I I think that's where Doge and I met, and it was just this massive tire that, like, a single person basically can't flip single, like, by themselves. And so Doge was trying to flip it, and he, like, pulled me over. He's like, oh, I think the two of us could do So

那你为什么要翻这个轮胎呢？

And what why were you trying to flip this tire?

你懂的

You know

明白我的意思吗？为什么不翻呢？对吧。

what I'm saying? Why not? Flip it. Right.

但是，是的，我试着自己翻，但翻不动。然后我就想，我能找到的最强壮的人是谁？可能是Barrett或Liam。我记得是Liam，然后我们成功了。

But, yeah, I I tried doing it. I couldn't do it. And then I was like, who's the strongest person I can find? And I was either Barrett or Liam. And I remember Liam, and it worked.

我们就把它翻过来了。

We just flip it.

对。而且

Right. And

那是你们俩意识到彼此都有物理学背景的时刻吗？那是怎么发生的？你们是怎么从翻轮胎转向搞实验的？

was that the moment where you guys both realized you had physics backgrounds? How did that happen? What what how did you go from from flipping tires to flipping experiments?

是的。我的意思是，我不确定Liam是否记得这个，但我们这些年会保持联系，经常最终会讨论量子力学或超导性。这很常见。但我从没想过他们最终会一起研究物理学。当时Liam在研究LLM，而且进展非常顺利。

Yeah. I mean, so I don't if Liam remembers this, but we would catch up over the years and we would often end up talking about quantum mechanics or superconductivity. This was very common. But I never thought they would end up working on physics together. So Liam was working on LLMs and they were going really well.

我当时没有使用LLM，但我注意到LLM在我的工作中变得越来越有影响力。一个影响方式是，当我试图回忆化学、物理方面的一些知识时，我可以直接和聊天机器人对话，实际上学到了很多我忘记的东西。另一种方式当然是编程。比如我们在写模拟程序，LLM在帮我们编写这些模拟程序时非常有用。

And I was not using LLMs but I was noticing that LLMs are becoming more and more impactful in my work. So one way it was becoming impactful is when I was trying to remember some things about chemistry, physics. I could just talk to the chatbot and actually learn a lot of stuff I forgot. Another way was of course coding. Like we were writing simulations and the LLM was so helpful in writing these simulations for us.

于是问题就变成了：我们能否在物理研究中更把LLM当作一等公民来使用？

So then the question was like can we use LLM's kind of more as a first class citizen in the physics research?

是的。而且我认为，在做出离开的决定之前，Doge和我一直在联系，讨论这些不同的技术路线。我们关注模型在推理能力上的改进。我们看到高算力强化学习能做什么。而在材料科学方面，我们看到物理学和化学内部都存在缩放定律，无论是在模拟还是实验方面，这和机器学习中起作用的原理是类似的。

Yeah. And I think kind of leading up to this decision to leave, Doge and I were just, you know, connecting and talking about these different tech trees. We're looking at the improvements models, on reasoning. We're seeing what high compute reinforcement learning can do. And then on the material science side, we're seeing scaling laws within physics, within chemistry, both with respect to simulations, with respect to experiment, and it's like the same kind of principles at play and ML.

我认为对我们俩以及该领域的许多人来说，这项技术的目标是加速科学，加速物理研发。聊天机器人是沿途的一个重要里程碑，但我们真正希望看到技术走向世界。我们觉得这正是开始的好地方。物理学非常可验证。它是一个很好的奖励函数，迭代循环相当快。

And I think to both of us and to a lot of people in the field, the goal of this technology is accelerate science, accelerate physical R and D. Chatbots was a great milestone along the way, but we really want to see technology out in the world. And we felt like this was just the right place to begin. Physics is very verifiable. It's a great reward function, fairly fast iteration loop.

你们有用于大型物理系统的模拟器。我们觉得为了创造这个AI科学家，这就像是这条道路的开端。因此建立了这种信念，并决定创立Periodic。

You have simulators for large classes of physical systems. And we felt like in order to create this AI scientist, this is like the beginning of this path. So built that conviction and decided to found Periodic.

好吧，让我们花点时间谈谈Periodic是什么，它是做什么的。

Well, let's take a second to talk about what Periodic is and what does it do.

Periodic Labs是一个前沿AI研究实验室，致力于利用LLMs推动物理和化学的发展。我们认为将实验与模拟和LLMs紧密耦合在循环中极为重要。因此我们正在建立一个能够生成高通量、高质量数据的实验室。我们将结合实验使用LLMs和模拟来进行迭代。科学本质上是迭代的，我们相信LLMs利用人类可用的所有这些工具，能够在加速物理研发方面发挥巨大作用。

So Periodic Labs is a frontier AI research lab that's trying to use LLMs to advance physics and chemistry. We feel like having experiment in the loop tightly coupled with simulations and LLMs is extremely important. So we're building up a lab that will generate high throughput, high quality data. And we will use LLMs and simulations in conjunction with the experiments to try to iterate. Science, by its nature, is an iterative direction, and we feel like LLMs, using all these tools that are available to humans, can do a great job in accelerating physical R and D.

我想说目标是取代我们今天使用的数学评分器和代码评分器的奖励函数。举个例子，数学评分器，你有一个提示，二加二等于多少？你知道正确答案是四。你可以对这类可通过程序检查的问题施加很大的优化压力。而我们通过拥有实验室所做的是创建一个基于物理的奖励函数。

I'd say the objective is let's replace the reward function from math graders and code graders that we're using today. So, like, math graders, you know, to give an example, you have a prompt, what is two plus two? You know the ground truth is four. You can put a lot of optimization pressure against problems like that that are programmatically checkable. And what we're doing, and by having the lab, is we create a physically grounded reward function.

这成为我们优化的基础。因此，如果模拟器存在一些缺陷或问题，我们总是会进行修正，因为对我们来说，实验才是基本事实。就像，强化学习环境本质，在我们的设置中就是我们的强化学习环境。

That becomes the basis on which we're optimizing against. And so if a simulator has some deficiencies or some issues, we always air correct because for us, the ground truth is the experiment. Like, the RL environment nature, like, is our RL environment in in our setting.

让我们花点时间为可能不熟悉的朋友解释一下你们所说的‘在现实世界中验证强化学习的实验室’是什么意思。你能谈谈实验是如何运作的吗？今天的AI模型是如何训练的，以及这与它们在Periodic将被训练、开发、后训练和部署的方式有何不同？

Let's just take a second for folks who might not be familiar to explain what you guys mean by a lab that will verify RL in the real world. Can you talk a little bit about how experiments work? How do how do how are AI models trained today, and how is it how are those different from how they're gonna be trained and and developed and post trained and deployed at Periodic?

谈谈你们是如何创建ChatGPT的可能会有所帮助。ChatGPT最初的技术在过去几年发展非常迅速。当我们最初创建它时，它是一个非常标准的RHF流程。所以你有一个预训练模型，它有点像这种原始基质。你试图做的是把这个自动补全模型变成有用的东西。

And it might be helpful to talk about how you created ChatGPT. So ChatGPT originally, the technology evolved very rapidly over the last few years. When we were first creating it, it was a very standard RHF pipeline. So you have a pre trained model, and it's sort of like this raw substrate. And what you're trying to do is take this autocompletion model and turn it into something useful.

当时我们的做法是使用监督数据。也就是说，给定一些输入，我们会指定一个期望的输出。如果我们想让它扮演助手角色，就会创建这样的元组。然后运行强化学习，但这次是根据人类偏好训练的奖励函数进行学习。人类会说，给定这个输入，我更喜欢完成方案A而不是B。

The way we did it at that point was we would have supervised data. So given some input, we would say this is a desired output. So if we're trying to get it to act as an assistant, you know, we create some tuples like that. Then you run reinforcement learning, but now you're learning against a reward function that's trained against human preferences. So humans will say, well, given this input, I would prefer completion A to completion B.

不断重复这个过程，就能创建一个可供优化的奖励函数。这基本上就是我们创建ChatGPT的基础原理。但从原始模型到今天的产品还存在巨大差距。我认为部分原因在于推理能力，但另一部分在于奖励函数变得更加优秀和精确。我们最初使用的奖励函数无法判断数学正确性。

And you do that over and over again, and you can create a reward function that can then be optimized against. That is sort of the basis of how we created ChatGPT. But then there's a huge gap between the original model and what we have today. And I think part of that is reasoning, but also part of that is just much better, more precise reward functions. So the reward functions that we were using originally couldn't determine whether you were mathematically correct or not.

所以早期版本的ChatGPT在数学方面并不强，这很大程度上源于奖励函数的设计。你优化的是什么？奖励函数基本上编码的要求是：成为友好助手，帮助人们达成目标，但它没有判断数学正确性或代码有效性的能力。后来我们在奖励函数的正确性方面取得了巨大进步。

So early versions of Chatuchu Bt were mathematically not particularly strong, and it sort of results from the reward function. What did you optimize against? You the reward function basically encoded be a friendly assistant, try to help people get to their thing, but it had no sense of is this mathematically correct or not? Is this code valid or not? And we made huge advances over the correctness of our reward functions.

但这些都是数字层面的。我们基于互联网、教科书和论文创建任务，这很棒，奠定了基础。但最终，科学是由现实世界的实验驱动的。这就是我们在Periodic Labs正在做的事情。

But this is all digital. We're creating tasks based on the Internet, textbooks, papers, and this is great. This lays the foundation. But ultimately, science is driven against experiment in the real world. And so that's what we're doing with Periodic Labs.

我们采用这些前置技术并认为：如果想要推动科学发展，就需要将实验纳入循环，这将成为我们智能体的奖励函数。正如Josh所说，我们的智能体做着与编码或回答查询类似的事情。但现在不仅仅是提供Python或浏览器等工具，我们还拥有量子力学等工具来模拟不同系统。但最终，我们要进入实验室，这将构成系统优化的基础。

We're taking these precursor technologies and we're saying, okay, if you care about advancing science, we need to have experiment in the loop, and that becomes our reward function for our agents. So as Josh was saying, our agents are doing the same type of things you'd use for coding or to help answer a query. But now instead of just giving tools like, here's Python, here's a browser, now we have tools like quantum mechanics, so simulate different systems. But ultimately, we're going to a lab, and then that becomes like the basis of what is the the system optimizing against.

所以

So

所以这就像是这些系统自然发展的终极状态。

So that's sort of just like the natural, like, end state of these systems.

AI领域的人经常提到实验室。他们所说的实验室通常与你们理解的实验室含义大相径庭。Doge，区别在哪里？

People in AI often say lab. Often what they're referring to is quite different from you what you guys mean by lab. Doge, what what's the difference?

没错。正如D提到的，到目前为止，语言模型在逻辑和数学方面已经变得非常出色。这些领域有可验证的回报。在逻辑和数学之后，下一个探索前沿是什么？我认为是物理学。

That's right. So as D mentioned, so far, LMs have gotten really good at logic and math. There's verifiable rewards. What is the next frontier in terms of inquiry after logic and math? I'd say it's physics.

说到物理学，它涉及不同的能量尺度。有天体物理学研究星系，有核聚变、核物理，但还有与我们生活更相关的物理学能量尺度。那就是量子力学，比如薛定谔方程。生物学、化学、材料科学都在这个尺度发生作用。所以我们认为第一个实验室应该主要探索这个量子力学能量尺度。

And then when you say physics, are different energy scales. So there's astrophysics studying galaxies, there's fusion, nuclear physics, but then there's the energy scale of physics that's more relevant to our life. And that's the quantum mechanics like Schrodinger's equation. This is where biology happens, chemistry happens, materials happen. So we felt like our first lab should be basically probing that quantum mechanical energy scale.

对我们而言，这将是固态物理、材料科学和化学层面的物理学。制造我们周围物品的更基本方法之一是粉末合成。你取现有材料的粉末，混合它们，加热到特定温度，就会形成新材料。这是我们的实验室之一。我们将设立粉末合成实验室，事实证明这是机器人可以执行的廉价简单方法。

And for us that would be physics at the level of solid state physics, material science and chemistry. Kind of one of more fundamental ways of making things around us is powder synthesis. So you take powders of existing materials, you mix them and you heat them up to certain temperature and it becomes a new material. So this is one of our labs. We're going to have a powder synthesis lab and turns out this is one of those methods where robots can do it, like very cheap simple methods.

不知道你是否见过旧金山机场的咖啡制作机器人。那种水平的机器人完全可以混合粉末并放入熔炉。这个领域非常丰富，通过这种方法实际上可以发现新的超导体、磁体等各种对我们周围技术至关重要的材料。但其核心就是量子力学。我们认为教会这些大语言模型成为量子力学的基础模型，将是LLM的下一个前沿领域。

I don't know if you saw this coffee making robot in the SF Airport. You know a robot that's basically at that level can mix powders and put it in a furnace. And there's a very rich field, you can actually, using that method, discover new superconductors, magnets, all kinds of materials that are very important for technologies around us. But at the core of it is just quantum mechanics. And we feel like teaching these LLMs to be foundation models, but for quantum mechanics will be the next frontier for LLMs.

为什么当前世界上已部署的模型还做不到这一点？

Why haven't the models that are currently out in the world and deployed able to do this?

好问题。正如你之前提到的，科学本质上是迭代的。最聪明的人类在发现成果前都经历了多次尝试。我认为这可能是LLM令人困惑的一点。LLM可以非常聪明。

Great question. I think as you mentioned earlier, science is by its nature iterative. The smartest humans tried many times before they discovered the things they discovered. And I think maybe this is one of the confusing points by LLMs. LLMs can be very smart.

但如果它们不在科学上进行迭代，就无法发现科学。老实说，人类也不行。你把一个人关在房间里，没有任何机会对事物进行迭代，他们也不会发现任何重要的东西。所以我们认为，教给这些大型语言模型的关键在于科学探究的方法。你要进行模拟，进行理论计算，进行实验，得到结果，这些结果最初可能不正确或不理想，但你要在此基础上迭代改进。

But if they're not iterating on science, they won't discover science. To be honest, humans won't either. You put a human in a room without any chance to iterate on something, they won't discover anything important. So we feel like the important thing to teach these LLMs is the method of scientific inquiry. So you do simulations, you do theoretical calculations, you do experiments, you get results, and the results are probably incorrect or not what you want at first, but you iterate on it.

我们觉得这一点尚未实现。所以这就是我们想做的，但我们认为必须用真实的物理来实践，而不仅仅是模拟。这就是为什么我们拥有自己的实验室，让大型语言模型有机会迭代其对量子力学的理解。

And we feel like that hasn't been done yet. So this is what we want to do, but we feel like you have to do it with the real physics, not just the simulation. So this is why we have our own lab where the LLM will have the opportunity to iterate on its understanding of quantum mechanics.

从根本上说，机器学习模型擅长你训练它们做的事情。没错，这有点像是它的本质。所以如果一个模型表现不佳，你会想，你是否训练过它执行那个任务？顺着Doge的观点延伸，这里存在一种认知上的不确定性，一种可减少的不确定性，除非你真正进行实验，否则你并没有真正构建或消除这种不确定性。

Fundamentally, machine learning models are good at what you train them to do. Right. And that's sort of like the nature of it. And so if a model is acting badly, you're like, well, did you train it to do that task? Kind of building on Doge's point, there's sort of like an epistemic uncertainty, this like reducible uncertainty that you aren't really building or collapsing unless you're actually running an experiment.

例如，我们团队的一位工程师在查阅文献中报道的某种物理性质时，发现其数值跨越了许多数量级。如果我基于这些数据训练一个系统，这些系统并非魔法。它最多只能复制那种分布，但并未更深入地理解宇宙、物理或化学。另一点是，发表负面结果非常罕见。所有结果基本上都是正面的，而一个有效的负面结果非常有价值。

So for instance, one of the engineers on our team was looking at a reported property of some physical property in the literature, and it spanned many orders of magnitude. So if I train a system on that, these systems aren't magic. The best it can do is replicate that distribution, but it's really no closer to a deeper understanding of the universe, physics, chemistry. Then another point is it's very uncommon to publish negative results. All of the results are basically positive and a valid negative result is very valuable.

负面结果可能被丢弃，因为那被认为是草率的科学。但确实存在有效的负面结果，这是一种学习信号，我们的实验室也会产生这样的结果。所以我认为这三点——嘈杂的数据、缺乏负面结果，以及需要行动能力才能真正进行科学这一迭代性努力——是我们需要实验室的核心论点。

A negative result could be discarded because, well, that was sloppy science. But there are valid negative results and that's a learning signal and this is something that our lab will produce as well. So I think these three things, there's just like noisy data, no negative results, and you need the ability to act in order to actually do science, which is an iterative endeavor. Those are like the core theses of why we need a lab.

在你们看来，衡量周期公司（periodic）朝着该目标进展的核心方式可能是什么？

And what might be the core way to measure if periodic's progress against that goal in your guys' minds?

一个简单的例子是，比如高温超导。我们合成的最高温度超导体是什么？目前常压下的最佳记录大约是135开尔文。所以如果我们能超越这个数字，我们就很容易知道我们做得不错。这是相当根本的衡量标准。

One simple one is, let's say high temperature superconductivity. What is the highest temperature superconductor we synthesized? Today the best number for ambient pressure is 135 Kelvin or so. So we'll know very easily if we're doing well, if we can go beyond that number. So that's pretty fundamental.

在更应用的层面上，有材料加工及其对材料性能的影响。所以我们可以直接测量这些性能。比如说延展性、韧性、材料的强度。当我们测量时，LLM会得到非常清晰的信号。这很难被篡改，你知道，不像其他LLM训练技术那样。

On the more applied side, there's processing of materials and its effect on the materials properties. So we can just measure these properties directly. Let's say it's the ductility, it's the toughness, strength of the material. And as we measure it the LLM will get very clear signal. It's hard to hack, you know, unless unlike these other LLM training techniques.

这就像你在现实生活中看到的，就是传递给LLM的信号。

It's like really what you see in real life is the signal that's going to the LLM.

是的，确实如此。那么，你能设计你周围的世界吗？嗯。所以你就像说，我需要具有这种特性的东西。

Yeah. Effectively. So, like, do you can you design the world around you? Mhmm. So you're like, I need something with this property.

系统能发现并生产出来吗？嗯。既从基础科学发现的角度，也从工业的角度。比如，有人在航天、国防或半导体领域工作，然后说，是的，我们遇到了这些问题。我们正试图实现这种材料或这一层的这种特性。

System discover and produce that? Mhmm. Both from, like, a fundamental scientific discovery perspective, but also an industry. So, like, someone's working in space or defense or semiconductors and, like, yeah, we're having these issues. We're trying to achieve this property of this material or this layer.

系统能加速这些技术的发展吗？所以它非常接地气。

Can the system accelerate the development of those technologies? So it's it's very grounded.

这就是我们如何知道它在工作。感觉构建一个AI物理学家（暂时没有更好的词）来解决现实世界设计问题的应用是如此广泛。你可以将其应用于先进制造、材料科学、化学，以及任何需要与物理世界进行研发的过程。似乎我们将从Periodic正在努力取得的突破中受益。

That's how we'll know it's working. It feels like the applications of solving building an AI physicist, for lack of a better word, that can design the real world are so broad. You can apply them to advanced manufacturing. You can apply them to material science, to chemistry, to all anything that inter any process where there's r and d with the physical world required. It seems like we'll benefit from breakthroughs that periodic is working on.

为什么以前没有做到？是什么让现在这个历史时刻成为解决这个问题的合适时机？

Why hasn't it been done before? And what is it about this moment in history that makes it the right time to attack this problem?

也许一个评论很难。是什么让它如此困难？我的意思是，我认为部分原因在于团队。嗯。所以在我们的观点中，这是由过去几年的前沿技术所促成的。

Maybe one comment is difficult. What makes it so difficult? I mean, I think part of it is the team. Mhmm. So in our view, this has been enabled by frontier technology in the last couple of years.

所以Doja和我一直非常专注于组建这个独一无二的团队。就像这群物理学家、化学家、模拟专家以及世界上一些最优秀的机器学习研究人员，从未参与过这样一个协同努力。我们觉得，要真正实现这个目标，你需要所有这些专业知识。你需要这些支柱来完成这件事。所以当你

And so Doja and I have been so focused on basically putting together this n of one team. Like, these group of physicists, chemists, simulation experts, and some of the best machine learning researchers in the world have never been part of one concerted effort. And we feel in order to actually achieve this, you need all these expertise. You need these these pillars to do this. So when you

你们在设计团队时，你知道，在离开OpenAI和DeepMind之后，你们用来指导自己确定想要谁加入团队的主要启发式方法是什么？

guys went about designing the team, you know, after you left OpenAI and DeepMind, what was the primary heuristic that you used to guide yourself in figuring out who he wanted on the team?

所以在专业知识方面，我们想要覆盖LLM专业知识、实验专业知识和模拟。对于每一个领域，我们都希望拥有世界级的人才。当然，每个团队实际上还有很多子团队，就像专业知识是非常分形的。所以在实验方面，我们希望覆盖细胞状态化学、细胞状态物理、自动化，以及更多设施方面，比如实验的更多操作层面。在模拟方面，有更多理论物理部分，也有更多模拟的编码方面。

So in terms of expertise, wanted to have LLM expertise covered, the experimental expertise and simulation. And for each of these we wanted to have world class talent. And of course for each team there's actually a lot of sub teams, like fractal, expertise is very fractal like. So for the experimental side we want to cover cell state chemistry, cell state physics, automation, and kind of the more facilities, like the more operational aspects of experiments. On the simulation side there's the more theoretical physics parts, there's the more coding aspects of simulations.

而在LLM方面，当然，有中期训练、强化学习、基础设施。是的，对于每一个，我们试图基本上找到在这些子支柱中创新过的最优秀的人。

And the LLM side, of course, there's mid training, RL, infra. And, yeah, for each of these, we try to get basically the best people who have innovated in these, like, sub pillars.

是的。所以我认为就像没有一个团队能做到这一点。我们认为必要的技术真的只是在过去几年才出现，而这些数据并不像在Reddit论坛上那样。就像，你需要实际去产生实验数据、模拟数据。它被隔离在所有这些先进行业中。

Yeah. And so I think it's like there's not a team to do it. The technology that we think is necessary to do it has really just emerged in the last couple of years, and this data isn't, like, on a Reddit forum or something. Like, you need to actually go produce experimental data, simulation data. It's siloed across all of these advanced industries.

而其中许多，尽管有愿望，他们可能不了解，你知道，一些最近推动这波AI浪潮的最新技术？

And many of them, while there's a desire, they may not have knowledge of, you know, some of the most recent techniques that's been driving this this recent wave in AI?

曾经有一个时期，像GPT-3论文这样的模型或论文提出了语言模型或小样本学习器的概念，并提出了缩放定律的理念。随后OpenAI有一篇后续论文，我记得叫《生成式建模的缩放定律》，它表明只要以正确的组合不断增加计算量和数据规模，就能非常可预测地提升这些模型的性能。理论上，如果无限持续这样做，就会出现大量涌现能力。

There was a moment in time when models like or or papers like the GPT three paper, for example, that, you know, said language models or few shot learners and proposed the idea of scaling laws. And then there was a follow-up paper, if you guys remember from OpenAI that was called, I think, scaling laws for generative modeling. That just showed that as long as you just kept throwing you scaled up the amount of compute and data in the right combination. You could very predictably improve the performance of of these models. And the theory was that if you just kept doing that, you know, in in at at Infinidum, there would be a bunch of emergent capabilities.

这些模型将能够推理各种领域外、分布外的问题。这是否意味着——你如何调和这种观点：当前前沿实验室的大多数预训练和后训练流程最终难道不会也破解物理学吗？为什么物理验证这个概念如此必要？这种推理思路是错误的吗？

These models would be able to reason about all kinds of problems out of domain, out of distribution. Is it wouldn't that argue how would you how would you square the circle with that school of thought that, you know, that naively the current pre training and post training sort of pipelines at most of the frontier labs won't just eventually crack physics as well? Why why is is this idea of physical verification so necessary? And are is that school of is that school of sort of reasoning wrong?

是的。很好的问题。缩放定律在经验上似乎持续成立。所以这点没有疑问。嗯。

Yep. Excellent question. Scaling laws empirically seem to continue to hold. So that's not in question. Mhmm.

但我认为问题在于这个y轴代表什么？那个测试分布与我们讨论的内容非常不同。假设你在互联网上进行预训练，测试分布可能是互联网的代表性样本，你会看到这些可预测的缩放特性。但这不会捕捉到针对不同分布时完全不同的缩放特性集合。

But I think there's a question of what is this y axis? And that test distribution is very different from like what we're talking about. That test distribution, let's say you're pre shitting on the Internet, might be, you know, a representative sat from the Internet, and you'll have these sort of predictable scaling properties. But that's not going to capture that you have a very different set of scaling properties with respect to different distributions.

嗯。

Mhmm.

我试着说得更具体些。假设我们正在训练一个编程模型，用单元测试提供奖励信号。模型编写一些PR（拉取请求），我们检查单元测试从失败变为通过，就说这是成功的。我们会强化这些行为。

So I'll try to make this a little bit more concrete. Let's say, hypothetically, we're training a coding model, and we have unit tests to provide some reward signal. So the model writes some PR. We check that the unit tests go from failing to passing, and we say, this was successful. We're gonna reinforce these things.

你可能会说，开始优化这个系统后，它编写自开发代码的能力会越来越强。就会出现这种加速，这种起飞场景。代码是最有前景的领域之一，因为网上有海量数据。存在这种反馈循环，系统可以开始自我改进。这是非常有前景的技术，我们都看到了高级编程模型的好处，而且正在快速加速。

You might say, you start optimizing this, and now the system is becoming ever more capable of writing code for its own development. And you have this acceleration, you have this kind of takeoff scenario. Code is one of the most promising areas for this because there's abundant of data online. You have this feedback loop where the system itself can begin to improve itself. And it's it's a very promising technique, and we're all seeing the benefits of, advanced coding models, and it's accelerating quickly.

然而，那个模型并不会因此治愈癌症。相关知识根本不存在。它无法做到这一点，你需要针对你所关心的分布进行优化。所以那个模型，虽然作为软件工程师和分析工具会非常有价值，但它根本没有数据、知识或在该环境中迭代的专业能力。

However, that model is not going to then cure cancer. The knowledge simply doesn't exist. It it doesn't you need to optimize against the distribution you care about. So that model, while it's gonna be a very valuable tool as a software engineer, analysis. It simply doesn't have the data, the knowledge, or the expertise iterating against that environment.

我认为这基本上就是我们的核心理念。

And I think that's just sort of like the fundamental belief we have.

是的。实际上，当Liam和我研究视觉模型的缩放定律时，我们对此进行了一些探讨。OpenAI的CLIP论文中也经常提到这一点。域内泛化和自动域泛化是单调相关的，但不一定是线性的。这意味着你可以持续改进模型，它会按照域内的幂律进行提升。

Yeah. I mean, so actually, Liam and I worked on this a bit when we look at the scaling laws for vision models. And this also came up a lot in the CLIP paper from OpenAI. Like, the in domain generalization and the auto domain generalization are monotonically correlated, but it's not linear necessarily. And so what that means is you can keep improving your model and it will improve as the power law in domain.

而对于自动域任务——正如Liam所说，即那些与你训练集内容略有不同的目标——它们也会遵循幂律改进，但幂律的斜率可能不够理想。因此，你可能需要花费数百年才能达到想要的结果。例如，我们在非论文中看到了这一点。我们发表了一篇论文，发现随着训练集规模的增加，域内性能（IID性能）会按幂律提升，自动域性能也会按幂律提升，但取决于自动域的具体情况（即与训练分布的偏离程度），幂律的斜率可能非常小，以至于基本上毫无用处。

And for auto domain tasks, by which I mean, as Liam said, the things that you're trying to do that's a bit different than what's in your training set, will also improve the power law, but the slope of the power law may not be good enough. So that you might need to spend centuries before you get to the result you want. We saw this in the non paper, for example. We published a paper where we saw that as you increase the size of your training set, the IID performance, the in domain performance improves as a parallel. Auto domain performance also improves as a parallel, but depending on what the auto domain is, like far you are from training distribution, the power law might have such a small slope that it's basically useless.

因此，我们认为取得进展的最佳方式是让目标尽可能接近域内训练。实现这一点的最佳方法基本上是迭代调整训练集，使其更接近你想要的任务。这是一种解决方案。另一种可能更简单：我们想要的实验数据实际上并不存在。

So this is one of the reasons we feel like the best way to make progress is to make your target as close to your in domain training as possible. And the best way of doing this is to basically iterate on changing your training set to be more like what you want to do. So this is one answer. The other one is actually maybe even simpler. The experimental data we want actually doesn't exist.

例如，如果你想利用文献中的实验数据学习合成相关任务，结果发现形成焓标签（即组装原子成所需形状所需的能量）的噪声非常高，以至于在其上训练机器学习模型的预测能力不足以预测下一个结果。原因之一，正如Liam提到的，人们通常不发表阴性结果。而阴性结果通常高度依赖上下文，所以对某人来说是阴性的结果，如果方法不同可能会变成阳性。是的，所以不仅存在域偏移问题（即你的目标与训练集不同导致幂律斜率不够大），

So for example, if you look at, like you wanna say learn on the experimental data in literature for synthesis, turns out the formation enthalpy labels, which is like the energy it takes to basically assemble the atoms in the shape you want, It's so high that if you train a machine learning model on it, it's not predictive enough to predict the next one. And one of the reasons for this is, as Liam mentioned, people don't usually publish negative results. And negative results are usually very context dependent. So what's a negative result for someone might be positive if they do things differently. So, yes, so not only is there this domain shift problem where what you're trying to do might be different than your training set, so the power law won't have the large enough slope you want.

另一个问题是，对于我们想做的某些事情，根本没有数据可用。例如，对于超导性，虽然有很多数据集可供参考，但它们的噪声基底太高，以至于在其上训练

But the other problem is for some of these things we want to do, there's no data for it. For example, for superconductivity, there's a lot of data sets you can look at, but the noise floor on them is so high that training on them

通常没什么帮助。Doge、我以及整个团队都深信扩展性和扩展法则，但关键是直奔你关心的目标。对我们而言，我们关心的是推动科学发展、推进物理研发。这就像是我们的核心理念。

usually doesn't help. Doge, me, the entire team are deep believers in scaling up and scaling laws, but it's just do a beeline for the thing you care about. And in our case, we care about advancing science, advancing physical R and D. That's that's sort of like the thesis.

在超级比特构建与单纯增加算力解决问题之间，以及你们实验室描述的领域特定流程之间是否存在张力？就Periodic而言，我想你们提到最初的重点是超导和磁性领域，对吗？这些领域为何成为首批流程的理想候选？它们是通往跨领域通用AI物理学家的中途站，还是存在风险，可能成为偏离轨道的岔路，无法实现你们所追求的AI科学超级智能这一北极星目标？

Is there a tension between being super bitter less than build and just throwing more compute at the problem and the, I guess, domain specific pipelines that the lab you guys just described will have to focus on? In the case of periodic, I think you mentioned the first beelines you guys are making are towards superconductivity and magnetism. Right? What is it about those domains that make them good candidates for the first for the first few pipelines that periodic is working on? And why are they just are they it stops along the way to an AI physicist that generalizes across all kinds of domains, or is there a danger of them being essentially off ramps that don't result in sort of a a the the AI sort of scientific superintelligence that is the North Star for what you guys are doing?

是的。我觉得，比如高温超导目标实际上包含许多子目标。这有点像DeepMind和OpenAI起步时说要做AGI，但他们指的是在取得酷炫成果前必须完成许多事情。对我们来说，如果想实现高温超导体，可能需要精通自主合成、自主表征，擅长用LLM正确运行模拟来表征材料的不同方面。

Yeah. I feel like, for example, the high temperature superconductivity goal is actually a goal that has so many sub goals in it. It's a bit like when DeepMind and OpenAI started and said we're gonna do AGI. But what they meant was they had to do so many things before they got to these cool results. Like for us, if we want to get a high temperature superconductor, we probably need to get good at autonomous synthesis, autonomous characterisation, we need to get good at characterising different aspects of the material, using the LLM to run the simulations correctly.

所以它是一个北极星，沿途有许多目标会对社区产生巨大影响。这是一个原因。另一个原因是高温超导本身就是一个根本性有趣的问题。例如，如果我们能找到200开尔文的超导体，即使尚未用它制作任何产品，这本身就能揭示许多我们尚未了解的宇宙奥秘。能在如此高温下观察到量子效应，我想这会彻底更新人们对宇宙的认知。

So it's a North Star and there's so many goals on the way that would be very impactful for the community. That's one reason. Another reason is feel like high temperature superconductivity is such a fundamentally interesting question. For example, if we could find a 200 Kelvin superconductor, even before we make any product with it, that in itself says so much about the universe that we didn't know yet. To be able to see such quantum effects on at such high temperatures, I think would be such an update to people's view of how they see the universe.

因此我们认为，即使在产品化之前，它也会对人类产生深远影响。我想这是原因之一。技术原因在于超导是一种相变，因此对我们尚无法模拟的某些细节相对稳健。例如，制作材料时，超导温度通常更取决于其晶体基本属性，而非缺陷或微观结构。

So we feel like it'll be really impactful for humanity even before we make product out of it. I think that's one of the reasons. A technical reason also is superconductivity is a phase transition. So it's pretty robust to some of these details that we cannot simulate yet. So for example, when you make the material, the superconducting temperature usually is more dominated by its kind of crystal fundamental property than the defects or microstructure.

而其他某些材料特性则不同，即使晶体具备所需特性，还有许多无法模拟的因素会阻碍该特性的显现。所以超导既有哲学上的优势，也有技术上的优势，并能团结物理学家群体。有研究物理四十年的学者对超导充满热情，也有从未学过物理但对此非常兴奋的人。很难找到一个能凝聚整个团队的话题。

Whereas there are certain other materials properties where even if the crystal has the property you want, there's so many other factors that you cannot simulate that would prevent you from seeing that property. So superconductivity has this nice philosophical upside to it, has this technical upside to it, and it really rallies both the physicists. There are people who study physics for forty years and really excited about superconductivity. And there are people who've never studied physics but very excited about superconductivity. It's quite rare to find a topic that unites the whole team.

是的。就像Dosh说的，为了实现这个目标，需要解决许多基础问题。我们的策略是，要真正实现AI科学家的目标，必须在某个地方完成闭环。如果只是泛泛而谈，最终只会回到论文库和教科书。所以对我们来说，完成闭环并创建可重复的过程至关重要。

Yeah. I mean, it's like Dosh said, it's like in order to do this, there are so many foundational pieces to solve. And our tactic is, in order to actually get to this goal of AI scientists, you need to make contact do the full loop somewhere. If you say you're doing this in just like very vague terms, you sort of just end up back on archive papers and textbooks. And so it's really important for us to do the loop, but then create this repeatable process.

比如，如何从一个子领域跨越到另一个子领域？这里存在一些非常有趣的问题，关于机器学习系统在这些领域间的泛化能力究竟如何。例如，一个系统在超导数据和磁性数据之间的泛化能力是怎样的？这可能与它在流体力学领域的泛化能力截然不同。我认为这里面有一些根本性的论点可以探讨。

Like, how do you go from subdomain to subdomain? And there's really interesting questions about how well do the ML systems generalize between these things. What is the generalization of a system between superconductivity data to magnetism data, for instance? And maybe that looks very different than its ability to generalize to fluid mechanics. And I think there's like fundamental arguments to make there.

但目标是创建一个可重复的系统，验证它，然后以这种方式遍历不同的领域。

But the goal is create this repeatable system, prove it, and then just go through the different domains that way.

所以我能理解为什么从实验基础上破解室温超导对人类来说具有非凡的价值，但你们正在创办一家初创公司。用一个类比来解释为什么需要有一条清晰的中期或较短期的路径，通往一个既具有商业可行性又对社会有净正面影响的北极星目标。例如，我们看到其他前沿实验室正在自动化白领工作或软件知识工作，他们的北极星目标是打造AI研究员。但在实现这一目标的过程中，有一系列子目标等等。不过，在通往AI研究员的道路上，一个具体的应用——AI编程——打开了巨大的商业价值并为用户带来了诸多好处。

So I I can see the argument for why cracking room temperature superconductivity from an experimental basis is is extraordinary valuable for humanity, but you guys are building a startup. And to use an analogy for why you need to have a clear medium term path or shorter medium term path along the way to a north star that is both commercially viable and net positive to to society. What we've seen, for example, with other frontier labs that are working on automating white collar work or or software knowledge work is that, you know, there's this north star of an AI researcher. But that along the way, there were a bunch of sub goals and so on. But a concrete kind of application that opened up a ton of commercial value and and benefits for users on the way to that AI researcher was the idea of AI programming.

对吧？软件工程可能已经成为第一个主要领域，让人们真正更新了他们对AI模型在消费应用之外实用性的先验认知。就生产力而言，它们在短短几个月内的影响已经非常显著。所以，如果传统前沿实验室的北极星是AI研究员，而实现这一目标的路径是编程，即AI编程。那么对于Periodic来说，这又是什么呢？

Right? Software engineering has become probably the first major domain that's caused people to really update their priors about how useful AI models are beyond kind of consumer applications. And in terms of productivity, their impact has been extraordinary just in a few short months. So if the traditional Frontier Labs as North Star was an AI researcher, and the path along the way to get there was programming, AI programming. What is that for periodic?

基本上，是为先进行业的工程师和研究员提供副驾驶助手。也许因为我们在硅谷，我们更多地思考计算机导向的工作。一切都是数字化的，一切都是比特。但还有很多其他行业。

Basically, copilots for engineers, researchers in advanced industries. So maybe perhaps just being in Silicon Valley, we really think about computer oriented work. Everything is digital. Everything is bits. But there's so many industries.

比如，我们刚才讨论了一些，比如太空、国防、半导体行业，它们处理的是材料物理的迭代，这是他们工作流程的一部分。他们是如何设计这些新技术、新设备的？在缺乏数据和良好系统的情况下，他们并没有特别好的工具。这就是我们的机会，而这些行业拥有庞大的研发预算。所以，是的，虽然高温超导是一个伟大的北极星目标，但我们非常清楚技术与资本是紧密交织的。

Like, we were kinda talking about a few, like, you know, space, defense, semiconductors, where they're dealing with iteration of materials of physics, and that's part of their workflow. How are they designing these new technologies, these new devices? And in the absence of data, in the absence of good systems, they don't really have particularly good tools. That is our opportunity, and these are massive R and D budgets. So, yeah, while high temp superconductivity is a great north star, we very much understand that technology and capital are intertwined.

如果这是一个极其成功的商业实体，我们将能够最大限度地加速科学发展。为此，我们希望加速所有这些不同行业的先进制造，成为所有这些团队的智能层，加速他们的工作流程，开始缩短迭代时间，让他们更快地找到更好的解决方案，加速他们的研究员和工程师的工作。

We're going to be able to maximally accelerate science if this is a wildly successful commercial entity. And to do so, we want to accelerate advanced manufacturing in all these different industries, become like an intelligence layer for all these teams to accelerate their workflow and start reducing their iteration time, get them to better solutions more quickly, accelerate their researchers and their engineers.

让我们更深入地探讨一下实践中的情况，比如一个周期性团队成员的一天，假设团队中大约一半是机器学习背景的ML科学家，另一半是物理或化学背景的物理科学家。你如何开始融合这两种文化？对吧？如何让一个职业生涯主要在实验室做物理和化学实验的人对机器学习产生直觉，反之亦然？

Let let's click a little bit deeper on the that in practice, sort of a day in the life of a periodic team member, where let's say, half the team is this roughly right? About half the team are ML scientists with machine learning backgrounds and the remaining half are physical scientists with physics or chemistry backgrounds. How do you start by uniting the cultures? Right? How do you take somebody whose primary career so far in work has been experiments in the lab, in in wet labs, doing physics and chemistry, and give them an intuition for ML and vice versa?

因为，你们俩都是物理学家，然后职业生涯轨迹让你们有机会在Frontier AI Labs工作，并参与了像Jack GPT、GNOME这样现在被视为里程碑式的机器学习系统的训练。但对于那些可能来自单一领域的人，你如何让团队对另一个领域建立直觉？

Because, you you guys are both physicists who then had the career trajectory where you also had the chance to be at Frontier AI Labs and were part of training systems that are now considered sort of landmark, hallmark machine learning systems like Jack GPT, like GNOME. But for others who might be coming from one domain, how do you get the team to build an intuition for the other?

是的。这是个很好的问题。我们认为确保这些团队紧密合作实际上至关重要。所以我们看到的一点是，物理学家和化学家需要弄清楚如何教LLM推理这些事情。因为我认为前沿AI实验室已经弄清楚了如何在数学、逻辑和化学上训练它们。

Yeah. So this is a great question. I mean, we feel like it's actually crucial for us to make sure these teams work very closely with each other. So one of the things we're seeing is the physics and the chemists need to figure out how to teach the LLM how to reason about these things. Because I think the frontier AI labs have figured out how to train them on math and logic, chemistry.

所以我们认为非常富有成效的一点是，物理学家和化学家正在思考我们应该在中途训练、RL训练中包含哪些步骤，以教LLM如何正确推理量子力学，如何正确推理这些物理系统。当然，另一方面是LLM研究人员也在学习很多关于物理、模拟工具和目标的知识。所以他们合作得非常好。我们有每周的教学会议，LLM研究人员教RL循环如何工作，数据清洗如何操作，而物理学家和化学家则教授科学的不同方面、科学史，这也非常重要。所以我们觉得进展非常顺利。

So one thing we're seeing that's been really, I think, productive is the physicists and chemists are thinking about what are the steps we should include in the mid training, in the RL training that will teach the LLM how to reason correctly about quantum mechanics, how to reason correctly about these physical systems. Another one of course is the LLM researchers are learning quite a bit about the physics, the simulation tools, the goals. So they've been working together really well. We have weekly teaching sessions where the LLM researchers teach how the RL loops work, how the data cleaning works, and then the physicists and chemists are teaching about different aspects of the science, the history of science, that's also very important. So we feel like that's been going really well.

看待这个问题的一种方式是，我们必须教LLM能够发现，比如说，超导体，包括能够很好地阅读文献，比如阅读所有论文、教科书，找到相关部分，然后能够运行模拟、理论计算，并采取行动运行实验。我们认为这与这些公司中的物理研发研究人员非常相似。他们必须阅读文献，可能阅读内部或外部文件。然后运行模拟、理论计算，并实际尝试实验性思考，从中学习。所以我们觉得我们在内部超导或物理目标上取得的所有进展，实际上让我们的LLM在服务客户方面表现得更好，这些客户正在进行非常相似的工作流程。

And one way of looking at this is the things we have to teach the LLM to be able to discover, say, a superconductor, includes being able to read the literature really well, like read all the papers, the textbooks, find the relevant parts, and then being able to run simulations, theoretical calculations, and then take action, run experiments. We feel like this is quite similar to the physical R and D researchers in these companies. They have to read the literature, read maybe internal documents or external documents. And then run simulations, run theoretical calculations, and then actually attempt to think experimentally, learn from that. So we feel like all the progress we're making towards our internal superconductivity or physics goals actually is making our LLMs much better at serving our customers who are doing very similar workflows.

是的。我认为文化上，没有愚蠢的问题。

Yeah. I think culture, no stupid questions.

你可以问最愚蠢的物理问题，最愚蠢的ML问题。我的意思是，我们公司有一些教员，他们实际上是优秀的老师。所以，这些学习会议真的非常棒。我注意到的另一件事是计算机科学家经常从API的角度思考。所以科学家会说些什么，他们总是试图将其映射。

You can ask just like the dumbest, like, physics question, the dumbest ML question. And, I mean, there's a few faculty as part of our company, and they're actually excellent teachers. So, I mean, these, like, learning sessions have been really fantastic. And another thing I noticed is computer scientists often think in terms of, like, APIs. So scientists will say something, and they're always trying to map it.

你可能会想，好吧。那么输入是什么？输出是什么？目标是什么？我该如何将其映射回去？

You're like, okay. Well, what's the input? What's the output? What's the target? How do I map that back?

这始终就像是这种翻译过程。我认为我们团队中也培养了一些人，他们就像是处于不同领域的边缘。比如，如果你有一个单纯形，比如纯粹的机器学习、大语言模型、纯粹实验主义者、纯粹模拟，也有人深入其中。他们一直是连接这些不同人群的优秀桥梁。所以这是一种主动学习，学习其他领域的知识，创建API，然后这些起到桥梁连接作用的人。

What are and it's it's always just like this translation. And I think we also have built up as part of the team, there's there's people, like, on these different edges. So, like, if you have a a simplex of, like, you know, pure ML, LLM, pure experimentalist, pure simulation, there's people who live in this inside as well. And so they've been excellent bridges for translating between these different groups of people. So it's active learning to, like, learn the other spaces, creating APIs, and then these kind of bridge connector peoples.

我认为Doge就是一个很好的例子。

I think Doge being an excellent example of that.

加入Periodic是否必须拥有物理或化学的高级学位？

Is it a requirement for somebody who wants to join periodic to have to have an advanced degree in physics or chemistry?

绝对不是。

Absolutely not.

你知道，

You know,

我们开的一个玩笑是，那位NBA球员是谁说的，我比你们更接近勒布朗·詹姆斯？我们对两位候选人说的是相反的话，因为即使是我们最好的物理学家，他们对物理不了解的部分也远大于他们了解的部分。所以对于这位新候选人，即使他们没有物理背景，他们需要学习我们所做事情的知识量，实际上与我们最好的物理学家需要学习的量相差不大，因为有太多化学知识要学，太多材料科学要学。我认为这是当今科学的一个有趣方面。过去，在十九世纪，有一些物理学家能在前沿领域做很多事情。

one of the jokes we're making is who was the NBA player who was saying that I'm much closer to LeBron James than you are to me? We were saying the opposite of that two candidates because the amount that even our best physicist doesn't know about physics is much bigger than the amount that they know about physics. So for this new candidate, even if they have no background in physics, how much they have to learn about what we're trying to do is actually not that different than how much the best physicist has to learn because there's so much chemistry to learn, so much material science to learn. And I think this is one of the interesting aspects of science today. In the past, in eighteen hundred's, there were these physicists that could do so many different things at the frontier.

如今，我们的知识体系已经庞大到顶尖思想家通常只能在某个非常特定的领域取得进展。也许这实际上阻碍了我们，比如要发现一种惊人的超导体——我们一直用这个例子——你必须精通化学、物理、合成、表征等多个领域。遗憾的是，我认为没有人能全面掌握所有这些知识。因此我们必须协作。所以我认为我们的团队就是一个小型范例，正如利亚姆所说，我们在那个单纯形中拥有许多不同的点。

Today, we've reached a point where our intellectual knowledge is so large that a leading thinker can usually only advance in one very specific field. And maybe this is actually holding us back because, say to discover an amazing superconductor, as we keep going back to this example, you have to know so much about chemistry, physics, synthesis, characterization. And unfortunately, don't think any human knows enough about all of these. So we have to collaborate. So I think our team is kind of like a small example of this where we have, as Liam said, a lot of different points in that simplex.

对任何人来说，他们都有太多需要学习的东西，但这基本上适用于其他所有人。例如，我算是来自物理学背景，但我在团队中向来自化学不同领域、物理不同领域的成员学到了更多物理知识。我认为这对LLM研究人员也是如此。我的意思是，他们可能直到开始与我们团队的其他研究人员合作后，才了解到LLM的某些方面。所以我认为这很棒，就像是我们试图用LLM所做事情的一个小缩影——因为我们正试图将我们作为研究人员学到的所有这些不同知识教给这个LLM。我觉得这是一次非常有趣的经历。

And for any person, they have so much to learn, but that's true for basically every other So for example, I supposedly come from the physics side of it, but I've been learning so much more physics because we now have people from different areas of chemistry in the team, different areas of physics. And I think it's true for LLM researchers as well. I mean, come in, there are aspects of LLM that they probably didn't know until they started working with other researchers in our team. So I think it's a great and it's like a small example of what we're trying to do with the LLM because we're trying to teach this LLM all these different things that we're learning as researchers. It's like a really fun experience, I think.

是的。

Yeah.

你们发现Periodic的优秀研究员与OpenAI、Anthropic或DeepMind的优秀研究员有何不同之处？

And what are you finding makes a great researcher at Periodic that's different from what might make a great researcher at OpenAI or Anthropic or DeepMind?

我会说重合度很高，但最大的决定因素之一可能是：你是否关心这个使命？加速科学对你来说是否是首要目标？看看现在的团队，这是一群极具使命驱动的人，他们认同这就是北极星，让我们为之努力。如果有人真的想改进某些大公司的产品，那可能更适合去那些大公司迭代和优化产品。

I would say there's very high overlap, but probably one of the biggest determinants is, do you care about this mission? Is accelerating science to you, is that like the big goal? And I think looking at the team right now, it's just an incredibly mission driven set of folks who are like, Yep, this is the North Star. Let's do that. If someone really wants to improve some megacorp's products, yeah, you'd probably be better off at that megacorp in iterating and improving their products.

但如果你关心科学发现，我认为Periodic Labs是实现这一目标的最佳场所。

But if you care about scientific discovery, I think Periodic Labs is the best place to do that.

团队现在有多大？

How big is the team today?

我们大约有30人，

We're roughly 30,

我相信。是的。差不多。

I believe. Yeah. Almost.

当你考虑将公司正在进行的大量研究部署到现实世界中时，我们讨论过的那些客户类型——航天、国防、先进制造业——这些都是关键任务型行业，众所周知，它们对所在经济领域至关重要。但通常，它们并不是最快速采纳新技术的。你如何看待在我们讨论过的那些前沿智能体——擅长科学、精通物理——部署到可能在人工智能或机器学习方面远不如你们成熟的公司或组织中？你是否有一个可行的工作理论，以确保进步的步伐不会在部署环节受阻？听起来你在研究方面已经有了相当不错的理论来推动科学进步的进程。

And as you think about taking a lot of the research that's going on at the company and deploying that out in the real world, the kinds of customers that we've talked about, space, defense, advanced manufacturing, these are these are mission critical industries that are known for being, you know, essential to whatever part of the economy they're part of. But often, they're not the most they're not the fastest to adopt new technology. How do you think about deploying the kinds of frontier agents that we've talked about that are great at science, great at physics in companies or organizations that might not be anywhere close to as sophisticated as you are in AI or ML. Is is there do you have a working thesis for how to make sure that the arc of progress is not bottlenecked on deployment? It sounds like you have a fairly good thesis on how to unblock the arc of scientific progress on the research side.

但在部署方面，你们乐观地认为什么样的工作理论可以帮助将Periodic正在构建的系统推广到现实世界中？

But when it comes to deployment, what might be a working theory that you guys are optimistic about that would help get the systems that periodic is building out into the real world?

嗯，

Well,

也许我们在与所有这些公司的对话中注意到的一件事是，他们都在寻找自己的人工智能战略。他们明白技术正在快速变化，他们审视自己的工作方式，却发现变化速度没有达到他们的预期。一些行业还在流失不同领域的关键专业知识，比如失去那些资深工程师、资深研究人员，然后他们就在想，我们该如何保留这些？但一个理论是理解这一点，思考这些API，思考评估标准是什么？

maybe one thing that we've noticed in in our conversations with all these companies is they all are looking for their AI strategy. They understand that, like, the technology is shifting really quickly, and they're looking at how they're doing their work, and it's not changing as quickly as they think it should be. Some industries also are losing, like, kinda key expertise in different fields, and they're, like, losing these, like, senior engineers, senior researchers, and they're like, okay. How do we, like, preserve that? But one thesis is understand it's kinda thinking about these APIs and thinking about what are the evaluations?

这些公司面临的最大瓶颈是什么？审视他们遇到的一些问题，我们可以将其映射到我们的系统中。然后我们说，我们认为我们可以显著加速这一进程。所以这不是一上来就说，嘿，我们第一天就要改造你的生产线，我们要彻底改变你的一切工作方式。

What are the biggest bottlenecks for these companies? Looking at some of the problems they face, and we can map that to our systems. And we say, well, we think we can dramatically accelerate this. And so it's not coming in and saying, hey, we're gonna transform your fab line on day one. We're gonna transform how you're doing everything.

忘掉一切。就像是，不，我们要

Forget everything. It's like, no, we're

解决

gonna solve

一个非常关键的问题，范围明确，评估标准清晰。你基本上是与他们共同起草方案，然后向他们展示这项技术在针对你所关心的事项进行优化时能有多么强大。所以，这里没有什么特别令人惊讶的地方，但正如你所料，这是一种类似'先立足再扩展'的方法。但真正要寻找的是该公司内部最大的推动者是谁，最大的问题是什么，确保你正在为他们解决一个非常实际的问题，并将其与我们技术能力最强的领域相结合。

a really critical problem, well scoped, very clear evaluations. You kinda co draft that with them, and just show them, like, how powerful this technology can be when you optimize against the thing you care about. So, you know, nothing particularly, like, surprising here, but, you know, sort of like a land and expand type method as as you might expect. But really looking for who are the biggest promoters within that company, what are the biggest problems, make sure you're solving a very real thing for them and intersect that with where is our technical capability the highest.

你知道，今天早上你和你销售渠道中的一位客户通了电话。我们不需要指名道姓，但你听到他们最迫切希望Periodic解决的问题有哪些？

You know, you were on a call this morning with one of the customers in your pipeline. We don't we don't need to name who, but what what were some of the things you heard as their as their most urgent problems that they'd like for periodic to solve?

其中之一是模拟。你知道，他们花费大量时间培训人员使用某些模拟，这对他们的发展至关重要。我认为能够自动化这些模拟将会非常有帮助。设计过程，以及一些小事，比如格式匹配，能够将模拟结果输入到设计流程中。所有这些似乎都非常重要，然后能够在同一个地方整合处理数据。

So one of them was simulations. You know, they spend a lot of time training people on some of these simulations they need to use is critical for their development. And being able to automate those simulations I think would be quite enabling. The design process and then some of the small things like matching the formats, being able to feed the simulation results into the design pipeline. All of these seem quite important, and then being able to treat the data together in the same place.

还有别的吗？

What else?

嗯，我认为有一个非常根本的问题。很多这些公司会依赖检索。这有点像是一种超级轻量级的东西。有人带着一个神经网络出现，他们就觉得太好了。我们只需检索你所有的数据，那就是你的解决方案了。

Well, I think there's a really fundamental question. So a lot of these companies will rely on retrieval. So that's sort of like a super lightweight lightweight thing. Someone shows up with a neural net, and they're like, great. We'll just retrieve over all of your data, and then that's your solution.

然而，正如我们在ChachibT等案例中所见，当你对数据进行预训练、将知识实际编码到权重中时，它不仅仅是检索系统，而是对材料有了更丰富、更深刻的理解。我认为这是一个重大的根本性挑战。例如，对于这个客户，他们可以授予员工权限，让系统以用户身份进行检索操作，从而匹配相同的访问权限。但如果你开始对不同部分进行预训练或中期训练，问题就来了——如果你对所有数据都进行预训练，这些数据可能只有公司CEO才能访问。因此，你必须思考如何对这些知识进行分类，并创建不同类型的系统。

However, as we've seen with things like ChachibT and other things, it's when you pre train on the data, when you actually encode the knowledge into the weights, it's not just a retrieval system, you have a richer, deeper understanding of the material. And I think this is a big fundamental challenge. So for instance, for this customer, they can give privileges to their employees and have retrieval as acting on behalf The system acts as the user, and so you can match those same privileges for access. But if you start doing pre training or mid training on different parts, it's like, well, if you pre train on every piece of data, that might only be accessible to say, like the CEO of that company. So then you have to figure out how do you sort of bucket that knowledge and create different types of systems.

但根据与用户的交流，目前他们似乎还没有很好的解决方案来将所有知识提炼到单个模型或一组模型中。也就是说，要超越检索，进行真正的训练。此外，他们正在进行的监督训练其实类似于ChatGPT早期阶段——输入、输出、少量示例，这种转变新思维的方式其实不对。高计算量的强化学习才是真正有效的方法，这才是你应该考虑的策略方向。

But I think right now, after talking with the user, they don't seem to have a great solution for sort of distilling all of the knowledge into, like, a single model or into a set of models. So, like, going, you know, going beyond retrieval to, you know, proper training. And then I think also the supervised training they're doing is really akin to, like, the early days of ChatGPT, where it's like input, output, you have a few examples, and kinda transforming this new way of thinking was like, no. High compute reinforcement learning is really effective. This is how you should think about the strategies it's using.

这就是你针对这些问题创建有效工具的方式，也是你进行高效优化的方法。

This is how you create effective tool using towards those problems, and this is how you optimize it effectively.

能否为可能不熟悉的听众解释一下，您所说的'中期训练'是什么意思？大家都知道预训练和后训练，但在上下文中中期训练指的是什么？

Could you describe for folks who may not be familiar with it, what you mean by mid training? Because people are familiar with pre training, they're familiar with post training. But in the periodic what does mid training mean?

好的，抱歉用了行话。这个术语几年前出现是因为：我们已有预训练和后训练，但有时需要额外注入一些知识。

Yeah. Sorry for the lingo. So I think this this term came up years ago where it's like, well, we had pretraining. We had post training. But sometimes you need to put in a little bit more knowledge.

在搜索功能完善之前，存在信息时效性问题。预训练模型有知识截止点，只能获取当时互联网的快照，但用户需要更实时的知识。于是就产生了如何通过中期训练注入新知识的方法——中期训练本质上是将模型未掌握的新数据新知识继续用于预训练。

So before search worked really well, there was an issue of freshness. So we had pre trained models and they have a knowledge cutoff. So there's like a scrape of the Internet at that point, but users want more real time knowledge. So it's like, how do you get that in there and enter mid train? Mid train is basically you're taking new data, new knowledge that's not in the model, and you continue pre train.

嗯。这与标准的后训练不同，后训练通常侧重强化学习、监督学习，而中期训练的核心目标是为模型注入大量前所未有的知识。简单来说这就是中期训练的要义。

Mhmm. And this differs from standard post training, where post training typically is more reinforcement learning, supervised learning. And the mechanism is basic or the the goal of it is just put a lot of knowledge into the model that doesn't exist before. So that's that's med training in a nutshell.

在周期性背景下，这是否意味着本质上是从某个特定客户或行业的实验性实现中注入大量自定义数据？你们认为哪些是原子单元，即中期训练中能够提升模型在目前表现极差的问题上的能力的关键要素？

And in the periodic context, does that mean essentially going and injecting a ton of custom sort of data from a an experimental implementation in a in a particular customer or particular industry? What is the what are the sort of the the lines, the atomic unit that you guys think will of of mid training that that will improve the capabilities of the models on on problems that they're just terrible at today?

我的意思是，这其实就是所有知识。所以你可以有非常低层次的物理对象描述，比如晶体结构。你也可以有更高层次的语义描述，比如我是如何制造材料XYZ的。

I mean, it's just it's all all the knowledge. So it's like you can have very low level descriptions of physical objects, so like crystal structures, for instance. You can also have higher level semantic descriptions of like, well, this is how I made material x y z.

嗯。

Mhmm.

尝试将所有数据输入模型非常有价值。比如模拟数据、实验数据，这些目前都不存在。基本上就是将知识注入模型，并确保这些分布以某种方式相互关联。我的意思是，如果你只是简单混合分布A、B和C，无法保证泛化能力。你希望从这些系统中看到的是，引入其他数据集能够提升在其他数据集上的性能。

And trying to get all this data into the model is really valuable. So it's like simulation data, experimental data, none of this exists. And basically putting that knowledge into the model and making sure that these distributions are connected in some way. And what I mean by that is if you just sort of mix together distribution A, B, and C, there's no guarantee of generalization. What you wanna hope to see from these systems is the inclusion of this other dataset is improving performance on the other datasets.

所以这些就像是机器学习技术或需要解决的机器学习问题。但基本上，就是让它在物理和化学方面成为专家，弥补之前的不足。

And so these are sort of just like machine learning techniques or machine learning problems to solve. But basically, just make it an expert in physics and chemistry and where it was deficient before.

你们都知道，今年早些时候我在斯坦福物理实验室花了一些时间对许多这类模型进行评估，结果是这些模型在科学分析方面表现很差。因为它们没有被训练

You guys both know that I I spent some time running evals on a bunch of these models at the Stanford Physics Lab earlier this year, and the results were that the models are terrible at scientific analysis. Because they weren't trained

去做这些。

to do so.

因为他们没有被训练这样做。另一方面，许多研究通用模型的现有团队正在投入努力使这些模型变得更好。你们构建Periodic的方式有什么特别之处，能够利用基础模型的所有进展，还是你们必须从头开始一切，因此无法与当今主流模型的进步兼容？

Because they weren't trained to do so. On the other hand, many of the the existing research teams working on the general models are investing in trying to make these better. Is there something about the way you're building periodic that gets the draft off of all of that progress in the base models, or do you have to start everything from scratch and therefore not be able to be composable with advancements happening in the bay the mainline models today?

是的。我的意思是，我们受益于所有不同的进步。其中之一是大型语言模型（LLMs）正在变得更好。我们肯定从中受益，因为我们采用预训练模型然后进行微调，比如HighComputer。另一个是物理模拟工具正在变得更好。

Yeah. I mean, we benefit from all the different advances. So one of them is the LLMs are getting better. And we definitely benefit from them because we take a pre trained model and then mid train it, you know, HighComputer. Another one is the physical simulation tools are getting better.

所以DeepMind、Meta、微软、学术团体，他们正在开源新的模拟方法，使用机器学习预测属性的新方式。因此我们基本上可以利用所有这些。看起来机器学习在物理和化学领域已经产生了如此大的影响，我们预计这些改进会持续下去。

So DeepMind, Meta, Microsoft, academic groups, they're open sourcing new ways of simulating, new ways of using machine learning to predict properties. So we get to basically utilize all of those. And it seems like machine learning has made such an impact in the physics and chemistry fields that we expect these improvements to continue.

我认为另一件事是，当我们考虑智能体工具时，比如这是一个浏览器，这是一个Python，但越来越多的人将工具视为其他神经网络，其他智能体。所以如果你看很多物理代码，它并不是特别深奥。这不是竞赛编程。这有点像粗糙的脚本。但你可以依赖一些最好的系统，无论它们在哪些方面表现出色。

I think another thing is when we think about tools for agents, think of like, here's a browser, here's a Python, but increasingly people think about tools as other neural nets, as other agents. And so if you look at a lot of, like, physics code, it's not particularly deep. This isn't competition programming. This is kinda like hacky scripts. But you can rely on some of the best systems for wherever they spike on.

因此，神经网络作为这些智能体的工具，立即加速了我们的工作。所以你不需要

So neural net as a tool to these agents is something that immediately accelerates our work. So you don't have

复制一切。有一个历史模式，我们在这里讨论的物理科学中的许多基础研究，如物理、化学、生物学，历史上都是在大学实验室完成的。你认为大学生态系统在Periodic的未来中会扮演任何角色吗，还是你认为这些是完全不同的路径？

to replicate everything. There's a historical pattern that a lot of the fundamental research in the physical sciences that we're talking about here, physics, chemistry, biology, has historically been done at university labs. Is there a role at all that the university ecosystem you think will play in Periodic's future, or do you think these are just completely divergent paths?

绝对有。我的意思是，我们使用的许多模拟工具都是在学术界开发的。其中很多在欧洲，例如许多新颖的合成方法。所以我们肯定受益于许多这些非常深入的技术进步。比如所有的物理模拟工具都是这些复杂的代码，在我们团队中，我们并不真正知道如何高效地开发它们。

Absolutely. I mean, so much of the simulation tooling we use have been developed in academia. Many of it is in Europe for example, a lot of the novel synthesis methods. So we definitely benefit from a lot of these different very deep technical progress. Like for example all the physical simulation tools are these complicated portrait encode that in our team for example, we don't really know how to develop very efficiently.

但我们认为学术界与工业界实验室之间确实存在着非常深厚的联系。例如，最近许多大规模模拟都是在微软DeepMind和Meta等工业实验室完成的，但很多这些工具实际上是在学术界开发并传承下来的。因此，两者之间实际上存在着非常好的协同效应。

But we feel like there's definitely a very deep connection between academia industry labs. So for example, recently a lot of the large scale simulations have been done in industry labs like Microsoft DeepMind and Meta. But a lot those tools have been actually developed in academia and then passed on. So there's actually really nice synergy there.

我想再补充几点。就像你在评估模型进行科学分析能力时发现的那样，它们存在不足。这可能是因为训练这些模型的团队并没有将其作为直接目标。因此，我认为学术界和这些合作可以说，帮助我们明确哪些是重要的任务，比如如何进行这种分析。

I think I'd add a few other things too. So like you found when you were evaluating models on their ability to do scientific analysis, they were deficient. This was probably, I mean, not a direct goal for those teams training those models. So I think academia and these collaborations say, well, help us inform what are the important tasks. Like, how do you do this analysis?

我们希望在模型中融入哪些技能？一个技能可以是一个完整的分析，也可以是一个更小的基本单元，作为更大分析的一部分。但其次是如何思考的问题。有一位物理学家在研究我们模型的推理策略时表示，这全错了。

What skills do we wanna put in the model? A skill could be a full analysis or a skill could be like a smaller primitive as part of a larger analysis. But also secondarily is how do you think? So one of the physicists was looking at the reasoning strategies of what our model is. He's like, it's all wrong.

全错了。我们问，你是什么意思？他说，不对，应该更高层次地思考，应该从对称性的角度来思考。

It's all wrong. And we're like, what do you mean? He's like, no. This should be thinking higher level. It should be thinking in terms of symmetries.

这本书编码了更有效的思维策略。当然，你的强化学习需要奖励这些类型的策略，但考虑到一些顶尖科学家正在使用这些策略，它们很可能是有效的。这些正是产业界与学术界合作可以发挥巨大作用的地方，因为产业界对这些类型的分析、工具以及这种思维方式往往是盲目的。

This is this is the book that encodes the thinking strategies that will be more effective. And, of course, your reinforcement learning needs to reward those types of strategies, but given some of the most premier scientists are using these strategies, they're likely effective. And these are types of things where it's like an industry academic partnership can just be so powerful because industry just simply is blind to these types of analyses, these tools, as well as just this way of thinking.

是的。这也与工具问题有关联，因为语言非常重要，但在人脑中，我们还能看到其他视觉处理，比如几何处理。因此，尽管这些大型语言模型会不断进步，但它们很可能受益于独立的几何推理能力。目前，我们可以通过等变图神经网络或本质上就是几何工具的扩散模型来实现这一点，大型语言模型可以调用它们，从而既拥有语言方面的优势（非常适合合成配方），又具备几何方面的优势（非常适合表示原子和一般几何设计）。

Yeah. And there's a way of connecting that to the tooling question as well because language is very important, but then in the human brain, we also see other visual processing, like geometric. So it's plausible that while these LLMs will keep getting better and better, they'll actually benefit from having a geometric reasoning that's separate. So today we can do that with equivariant graph neural networks, we can do it with diffusion models that are kind of geometric tools by construction, and the LLM can call them so then it can have both the language aspect, which is very good for a synthesis recipe, but also the geometric aspects, is very good for representing atoms, just design geometries in general.

那么你们是如何考虑与学术实验室深化周期性风格合作的？

So how are you thinking about deepening periodic styles with academic labs?

是的，这对我们非常重要。因此我们在这个方向上有两大举措。其中之一是我们正在组建一个顾问委员会，其专业知识将涵盖从超导到电池化学再到物理学等多个领域。我们希望确保与这类长期研究方向保持紧密联系。

Yes, this is very important for us. So we have two major initiatives in this direction. One of them is we're starting an advisory board. This will be kind of expertise spanning from superconductivity to cell state chemistry to physics. And we want to make sure we're in touch with this kind of long term research directions.

大量重要的政府资金会投入这些研究组，我们希望在他们的重点与我们的需求之间建立紧密耦合。因此委员会包括超导专家，如斯坦福大学实验方向的沈志勋（ZX Shen），理论方向的Steve Kielson。我们还邀请了西北大学的Mercury Canacidis提供合成化学专业知识，以及Chris Wolverton负责高通量密度泛函理论（DFT）方面。此外还有曼彻斯特大学的Kostya（康斯坦丁·诺沃肖洛夫），他以发现石墨烯而闻名，将能就新颖奇异电子态和材料方面为我们提供建议。

A lot of important government funding goes to these groups and we wanna have a tight coupling between what's important for them and us. So this includes superconductivity expertise such as ZX Shen from Stanford on the experimental side, and Steve Kielson from the theory side. We also have Synthesis Expertise on the advisory board from Mercury Canacidis from Northwestern University, and Chris Wolverton on the high throughput DFT side. And then we have Kostya from Manchester University who is really well known for discovering graphene. So he'll be able to advise us on these novel exotic electronic states and materials.

我们的第二项举措是通过资助计划。我们真心希望支持学术界正在进行的一些卓越工作，其中部分研究并不适合工业界，最好在学术环境中完成。因此我们打算接受资助提案，支持那些有助于社区的研究，特别是与大型语言模型（LLMs）、合成代理、材料发现和物理建模相关的工作。也许在这次节目后，你们可以在节目说明中附上链接。

And our second initiative is going to be through a grant program. We really want to enable some of this amazing work going on in academia, and some of their work isn't a good fit for industry. It's best done in academia. So we want to kind of accept grant proposals and we want to enable and support the kind of work that's going to help community, especially in relation to LLMs, agents in synthesis, materials discovery, physics modeling. So, you know, maybe after this show, you can include the link.

好的。如果资助计划从今天开始开放，我们绝对会把链接放在节目说明里。没问题。太好了。

Yeah. We'll include them in the show notes if if grants are open starting today. Absolutely. Yeah. Great.

那么对于可能有兴趣加入Periodic的人，你们在寻找什么样的人才？

So for people who might be interested in joining periodic, what are you guys looking for?

首先，我们需要极度好奇的人，真正渴望在更深层次理解机器学习和科学，希望接触现实、推动科学进步。这必须是一种内在驱动力，但同时也要务实。我们尝试的任务极具挑战性，需要做事细致、以解决方案为导向、能快速达成目标的人。而且需要在某个维度上达到世界级水平——我们关注所有支柱领域：机器学习、实验专家、模拟计算，以及能带来创新的人，比如如何创建创造性的ML系统，如何为最先进的模型引入新工具或新思维，如何推进模拟计算并使其更稳健、更与实验可靠结合。

First off, someone deeply curious, someone who really wants to understand the machine learning, the science at a deeper level, who wants to make contact with reality, who wants to advance science. This has to be a driving thing, but also pragmatic. What we're trying to do is incredibly challenging, and someone who has very careful process and they to their get solution oriented, they get to goals quickly. And really someone world class along some dimension, we're looking across all these different pillars, so machine learning, experimentalists, simulation, and people who can bring some sort of innovation on how do you create a creative ML system, how do you bring new types of tools or new types of thinking to some of these state of the art models, someone who can advance simulations and make it more robust and more reliable with experiment.

是的。我想补充一点：Liam和我一直非常看重候选人的紧迫感，因为我们不希望这些技术等到十年后才出现。我们不希望大型语言模型（LMs）十年后才开始推动科学进步，而是希望尽快实现。因此，如果候选人对于改进物理系统、发现卓越材料、在超导领域创新怀有紧迫感，他们会非常合适。

Yeah. And maybe one more thing I'd add is Liam and I have been really looking for a sense of urgency in candidates because we want these technologies not in ten years. You know, we don't want these LMs to start improving science in ten years, but we want them ASAP. So if the candidate feels like a sense of urgency for improving these physical systems, discovering these amazing materials, innovating on superconductivity, they would be a good fit.

是的。如果你符合所有这些条件，请联系我们。

Yep. If you match all these, please reach out.

好的。听起来我们需要加快速度，扩大定期活动的规模，我们会把快递链接放在节目说明中。谢谢大家参与。

Alright. Sounds like we gotta amp up the speed, the scale of stuff happening at periodic, and we'll put the courier links in the show notes. Thanks for coming, guys.

感谢收听a16z播客。如果你喜欢这一集，请在ratethispodcast.com/a16z留下评论告诉我们。我们还有更多精彩的对话即将推出。下次见。提醒一下，这里的内容仅供参考，不应视为法律、商业、税务或投资建议，也不应用于评估任何投资或证券，且不针对任何a16z基金的投资者或潜在投资者。

Thanks for listening to the a 16 z podcast. If you enjoyed the episode, let us know by leaving a review at ratethispodcast.com/a16z. We've got more great conversations coming your way. See you next time. As a reminder, the content here is for informational purposes only, should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a sixteen z fund.

请注意，a16z及其关联公司可能也持有本播客讨论公司的投资。更多详情，包括我们的投资链接，请参见a16z.com/disclosures。

Please note that a sixteen z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a 16z.com forward slash disclosures.