Periodic Labs深度访谈:构建AI科学家与自主实验室
Synthesis Superintelligence: from Semiconductors to Superconductors — Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk
这是一篇关于AI for Science落地的一手工程复盘,详细拆解了从RL到自主实验室的完整工作流与取舍,对从事AI科研自动化的从业者极具参考价值。
It’s hard to believe that Periodic was only launched last September:
很难相信 Periodic 直到去年九月才刚刚推出:
One year later, it is considered one of the pre-eminent AI scientist labs, with dizzying talent density and astonishing progress in the autonomous lab buildout:
一年后,它已被视为顶尖的 AI 科学家实验室之一,拥有令人眩晕的人才密度和自主实验室建设方面的惊人进展:
Most people are familiar with the standard credentials of Liam and Dogus, but we found an incredible “talent slope” while learning more about Periodic, where each successive employee seems more impressive than the last:
大多数人熟悉 Liam 和 Dogus 的标准履历,但在深入了解 Periodic 的过程中,我们发现了一条令人难以置信的“人才斜坡”,即每位新员工的成就似乎都比前一位更令人印象深刻:
From building AI systems that reason over noisy physical experiments to creating laboratories where every instrument can become intelligent, Periodic Labs is betting that the next frontier of AI won’t come from simply training on more internet data, it will come from letting models experiment with the real world. In this episode, Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk join swyx and Brandon to explain why scientific discovery is fundamentally different from math and coding, and what it takes to build AI scientists that can actually discover new materials.
从构建能在嘈杂物理实验中进行推理的 AI 系统,到创造每台仪器都能智能化的实验室,Periodic Labs 押注认为,AI 的下一个前沿不会仅仅来自在更多互联网数据上进行训练,而是来自让模型在现实世界中进行实验。在本期节目中,Periodic Labs 的 Liam Fedus 和 Ekin Dogus Cubuk 与 swyx 以及 Brandon 一起探讨了为什么科学发现从根本上不同于数学和编程,以及构建能够真正发现新材料的 AI 科学家需要付出哪些努力。
We go deep on Periodic’s vision for “synthesis superintelligence”: reinforcement learning grounded in physical experiments, AI-powered materials characterization, simulations and density functional theory, high-throughput labs, and systems that learn from the entire process of doing science rather than only its published results. Liam and Dogus also explain why frontier models will still need experiments, why failed experiments may be some of the most valuable training data, what it means to give every piece of lab equipment “140 IQ,” and how autonomous experimentation could compress decades of scientific trial-and-error into months.
我们深入探讨了 Periodic 对“合成超智能”的愿景:基于物理实验的强化学习、AI 驱动的材料表征、模拟和密度泛函理论、高通量实验室,以及能够从整个科学研究过程(而不仅仅是已发表的结果)中学习的系统。Liam 和 Dogus 还解释了为什么前沿模型仍然需要实验,为什么失败的实验可能是最有价值的训练数据,赋予每件实验室设备“140 IQ”意味着什么,以及自主实验如何将数十年的科学试错压缩至数月之内。
We discuss:
我们将讨论:
- Why intelligence alone isn’t enough for scientific discovery
- How reinforcement learning changes when the environment is the physical world
- Why science requires reasoning under uncertainty, noise, and missing information
- Prediction, synthesis, and characterization in the materials discovery loop
- Why physics and materials science are still far from “solved”
- The “matter compiler” and Periodic’s goal of synthesis superintelligence
- Phase transitions, X-ray diffraction, and AI-powered materials characterization
- DFT, simulations, and why experiments remain the ultimate ground truth
- Room-temperature superconductors, new magnets, batteries, and more efficient compute
- Why quantum computing may not automatically solve materials discovery
- What it means for every piece of lab equipment to have “140 IQ”
- Why data quality and negative results matter more than simply throwing more compute at science
- Training models on the process of doing science rather than the final answer
- Why even future frontier models will still need to physically experiment
- Scaling autonomous labs across AI, chemistry, physics, and custom hardware
- How automated experimentation could massively increase the “surface area for luck” in discovering new materials
- 为什么仅有智力不足以实现科学发现
- 当环境是物理世界时,强化学习将如何改变
- 为什么科学需要在不确定性、噪声和缺失信息下进行推理
- 材料发现循环中的预测、合成与表征
- 为什么物理学和材料科学距离“彻底解决”仍有很长的路要走
- “物质编译器”与 Periodic 的合成超智能目标
- 相变、X 射线衍射与 AI 驱动的材料表征
- DFT、模拟以及为何实验仍是最终的真理基准
- 室温超导体、新型磁铁、电池以及更高效的计算
- 为什么量子计算可能无法自动解决材料发现难题
- 赋予每件实验室设备“140 IQ”意味着什么
- 为什么数据质量和负面结果比单纯向科学投入更多算力更重要
- 在科学过程而非最终答案上训练模型
- 为什么即使是未来的前沿模型仍需进行物理实验
- 在人工智能、化学、物理和定制硬件领域扩展自主实验室
- 自动化实验如何大幅增加发现新材料的“运气表面积”
Liam Fedus
- LinkedIn: https://www.linkedin.com/in/liam-fedus-26547811/
- X: https://x.com/LiamFedus
Ekin Dogus Cubuk
- LinkedIn: https://www.linkedin.com/in/ekin-dogus-cubuk-9148b8114/
- X: https://x.com/ekindogus
Timestamps
时间戳
00:00:00 Introduction
00:00:00 引言
00:02:49 AI and Reinforcement Learning in the Physical World
00:02:49 物理世界中的 AI 与强化学习
00:09:17 The End-to-End Materials Discovery Loop
00:09:17 端到端的材料发现闭环
00:13:28 Why Physics Isn’t “Solved”
00:13:28 为什么物理学尚未“被解决”
00:19:04 The Matter Compiler and AI Characterization
00:19:04 物质编译器与 AI 表征
00:30:22 DFT, Simulation, and Experimental Ground Truth
00:30:22 DFT、模拟与实验真实情况
00:42:27 Dream Materials, Compute, and Quantum Computing
00:42:27 梦想材料、算力与量子计算
00:45:57 Giving Every Lab Instrument “140 IQ”
00:45:57 让每台实验室仪器拥有“140 IQ”
00:50:02 Automating the Lab
00:50:02 实验室自动化
00:53:16 Data, Models, and Negative Results
00:53:16 数据、模型与负面结果
00:58:13 Training AI on the Process of Science
00:58:13 在科学过程上训练 AI
01:00:20 Why Frontier AI Still Needs Experiments
01:00:20 为什么前沿 AI 仍然需要实验
01:06:06 Scaling Autonomous Labs
01:06:06 扩展自主实验室
01:10:45 Building the Team and Deploying to Industry
01:10:45 组建团队并部署至工业界
01:17:01 From AI Copilots to Scientific Outcomes
01:17:01 从 AI 副驾驶到科学成果
01:20:57 Automating the Search for Superconductors
01:20:57 自动化寻找超导体
Transcript
文字稿
Introduction: Intelligence Must Meet Reality
引言:智能必须与现实相遇
Swyx [00:00:00]: Okay. We’re here at Periodic. We’re here today with Liam and Doğuş from Periodic. Welcome, and thanks for having us at yours.
Swyx [00:00:00]:好的。我们目前在 Periodic。今天我们要采访来自 Periodic 的 Liam 和 Doğuş。欢迎你们,也感谢你们邀请我们来这里。
Liam Fedus [00:00:11]: Yeah, great to be here, and yeah, thanks for coming in.
Liam Fedus [00:00:11]:是的,很高兴来到这里,也谢谢你们过来。
Swyx [00:00:13]: I want to start off with one of these quotes that I love from your website. It says, “Intelligence is necessary but not sufficient. New knowledge is created when ideas are found to be consistent with reality.” It seems so straightforward, but why is it non-obvious?
Swyx [00:00:13]:我想从你们网站上的一句我很喜欢的引言开始。上面说:“智能是必要的但并非充分的。当发现想法与现实一致时,新知识便由此产生。”这看起来非常直白,但为什么它并不显而易见?
Liam Fedus [00:00:29]: I think that’s sort of the thesis and premise behind Periodic that Doğuş and I thought when we were building this, which is you can’t just think your way to a solution. The universe is so complicated that in order to actually push the frontier of knowledge and to make progress, you need to create these conjectures and then actually see whether or not it holds. No amount of rereading that textbook or paper, thinking, or disappearing into a room is going to allow you to think through all possible experimental outcomes, expected and unexpected, and you really need this iterative process. And that’s our core belief for putting this together, and that’s why from the very beginning, we’re like, “We need to have the AI systems, the simulations of the physical world, but then also to build up the physical high-throughput experiment.” And we think a different type of intelligence emerges from that.
Liam Fedus [00:00:29]:我认为这正是我和 Doğuş 在构建 Periodic 时所秉持的核心论点和前提,即你不能仅靠思考就能找到解决方案。宇宙如此复杂,为了真正推动知识的边界并取得进展,你需要提出这些猜想,然后实际验证它们是否成立。无论重读多少教科书或论文,无论进行多少思考或把自己关在房间里,都无法让你穷尽所有可能的实验结果——无论是预期的还是意外的——你确实需要这种迭代过程。这是我们构建这一切的核心理念,这也是为什么从一开始我们就认为:“我们需要拥有 AI 系统、物理世界的模拟,但同时也要建立实体的高通量实验平台。”我们认为,从中会涌现出一种不同类型的智能。
Swyx [00:01:21]: I think it’s also notable, like with OpenAI and Google, you don’t have those resources in those big labs. You had to come out and do your own thing, because you’re kind of inventing your own playbook as you go along.
Swyx [00:01:21]:我认为这也值得注意,比如在 OpenAI 和 Google 这样的大实验室里,你并没有那些资源。你必须走出来做自己的事情,因为你们是在边探索边制定自己的策略。
Building an Interdisciplinary Lab for AI Science
为 AI 科学构建跨学科实验室
Ekin Doğuş Çubuk [00:01:34]: Yeah, and a team like this has never existed, I think. We try to bring together solid-state chemists, solid-state physicists, experimentalists, theorists, hardware engineers, LLM experts, computer scientists, and some of these technologies are very recent. The high-throughput experimentation, the robotic arms have only been tried in the last three years or so. The force field expertise, some of these force fields are also very recent. But yeah, we felt like having a lab is very important. Having this focus and bringing all these people together to work together is very important. And I guess the closest examples from history are places like Bell Labs, where incredible theorists and experimentalists and chemists worked together and achieved incredible things. So we are trying to do the same here, yeah.
Ekin Doğuş Çubuk [00:01:34]:是的,我认为像这样的团队从未存在过。我们试图将固态化学家、固态物理学家、实验人员、理论家、硬件工程师、大语言模型专家、计算机科学家聚集在一起,而这些技术中的一些非常新。高通量实验、机械臂仅在最近三年左右才被尝试使用。力场专业知识方面,其中一些力场也非常新。但确实,我们认为拥有一个实验室非常重要。保持这种专注并将所有这些人才聚集在一起协作至关重要。我想历史上最接近的例子是贝尔实验室,那里杰出的理论家、实验人员和化学家共同工作并取得了非凡的成就。所以我们在这里也试图做同样的事情,是的。
Physical-World Reinforcement Learning and Uncertainty
物理世界强化学习与不确定性
Brandon [00:02:17]: I want to talk about force fields in a bit, but
Brandon [00:02:17]:我稍后想谈谈力场,但是
Swyx [00:02:19]: Yes
Swyx [00:02:19]:好的
Brandon [00:02:19]: Before we get to that. Yeah what are they? But before we get to that, For context, some of our audience are AI engineers, and some are scientists. For the engineers: what’s the difference between working at a big lab like OpenAI and what you’re doing here? I think you’ve already hinted at this but maybe explicitly, how do you have to reframe your thought process?
Brandon [00:02:19]:在那之前。是的,它们是什么?但在讨论这个之前,作为背景信息,我们的听众中既有 AI 工程师也有科学家。对于工程师们来说:在 OpenAI 这样的大型实验室工作与你们在这里所做的工作有什么区别?我认为你已经暗示过这一点,但也许可以明确说明一下,你需要如何重构你的思维方式?
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力