Periodic Labs 的 Liam Fedus 与 Ekin Dogus Cubuk 谈从半导体到超导体的“合成超级智能”
Synthesis Superintelligence: from Semiconductors to Superconductors — Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk
Periodic Labs 联合创始人 Liam Fedus 与 Ekin Dogus Cubuk 在播客中提出“合成超级智能”路线：把强化学习建立在真实物理实验之上，用 AI 做材料表征、DFT 模拟与高通量自动化实验室，让模型学习做科学的全过程而非只看已发表结果。
译文尚不完整，完整内容请切换到原文。
一年后，它已被视为最杰出的 AI 科学家实验室之一，拥有令人眩目的 人才密度，并在自主实验室建设方面取得了 惊人的进展：
大多数人都熟悉 Liam 和 Dogus 的标准履历，但在进一步了解 Periodic 的过程中，我们发现了一条不可思议的“人才斜坡”，每一位后续员工似乎都比前一位更令人印象深刻：
从构建能够对充满噪声的物理实验进行推理的 AI 系统，到创建每台仪器都能变得智能的实验室，Periodic Labs 押注的是，AI 的下一个前沿不会仅仅来自用更多互联网数据训练，而是来自让模型在真实世界中进行实验。在本期节目中，Periodic Labs 的 Liam Fedus 和 Ekin Dogus Cubuk 与 swyx 和 Brandon 一起，解释为什么科学发现与数学和编程有着根本不同，以及要构建真正能够发现新材料的 AI 科学家需要什么。
我们深入探讨了 Periodic 对 “合成超级智能” 的愿景：以物理实验为基础的强化学习、AI 驱动的材料表征、模拟与密度泛函理论、高通量实验室，以及从整个科学过程而非仅从已发表结果中学习的系统。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之一，让每台实验室设备拥有“140 智商”意味着什么，以及自主实验如何能将数十年的科学试错压缩到几个月。
我们讨论了：
为什么仅靠 智能本身不足以 实现科学发现
当环境是 物理世界 时，强化学习会发生怎样的变化
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仍远未 “解决”
“物质编译器” 与 Periodic 实现合成超级智能的目标
相变、X 射线衍射与 AI 驱动的材料表征
DFT、模拟，以及为什么 实验仍然是最终的地面真值
室温超导体、新型磁体、电池与 更高效的计算
为什么 量子计算 可能不会自动解决材料发现问题
让每台实验室设备拥有 “140 智商” 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 科学实践过程 而非最终答案上训练模型
为什么即使未来的前沿模型也仍然需要 进行物理实验
在 AI、化学、物理和定制硬件 领域扩展自主实验室
自动化实验如何能大幅增加发现新材料的 “运气表面积”
Liam Fedus
Ekin Dogus Cubuk
时间戳
00:00:00 引言
00:02:49 物理世界中的 AI 与强化学习
00:09:17 端到端材料发现闭环
00:13:28 为什么物理学并未“被解决”
00:19:04 物质编译器与 AI 表征
00:30:22 DFT、模拟与实验真值
00:42:27 梦想材料、计算与量子计算
00:45:57 赋予每台实验室仪器“140 智商”
00:50:02 实验室自动化
00:53:16 数据、模型与负面结果
00:58:13 用科学过程训练 AI
01:00:20 为什么前沿 AI 仍然需要实验
01:06:06 扩展自主实验室
01:10:45 组建团队并部署到产业
01:17:01 从 AI 副驾驶到科学成果
01:20:57 自动化寻找超导体
文字记录
引言：智能必须与现实相遇
Swyx [00:00:00]： 好的。我们在 Periodic。今天我们和 Periodic 的 Liam 和 Doğuş 在一起。欢迎，感谢你们邀请我们来到你们这里。
Liam Fedus [00:00:11]： 是的，很高兴来到这里，也感谢你们过来。
Swyx [00:00:13]： 我想从你们网站上我很喜欢的一句话开始。它说：“智能是必要的，但不充分。当想法被证明与现实一致时，新知识就产生了。”这看起来如此直白，但为什么它并不显而易见？
Liam Fedus [00:00:29]： 我认为这就是 Periodic 背后的论点和前提，也是 Doğuş 和我在构建它时想到的，那就是你不能只靠思考得出解决方案。宇宙如此复杂，为了真正推动知识前沿并取得进展，你需要提出这些猜想，然后实际看看它是否成立。无论你重读多少遍教科书或论文、如何思考，或者把自己关在房间里，都无法让你想清楚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。这就是我们把它组合在一起的核心信念，也是为什么从一开始我们就说：“我们需要有 AI 系统、物理世界模拟，但也要建立物理高通量实验。”我们认为一种不同类型的智能会从中涌现。
Swyx [00:01:21]： 我觉得也很值得注意的是，比如在 OpenAI 和 Google，那些大型实验室里并没有这些资源。你必须出来做自己的事情，因为你有点像是在一边前进一边发明自己的打法。
为 AI 科学构建跨学科实验室
Ekin Doğuş Çubuk [00:01:34]：是的，而且我认为像这样的团队以前从未存在过。我们试图把固态化学家、固态物理学家、实验人员、理论人员、硬件工程师、LLM 专家、计算机科学家聚集在一起，而其中一些技术是非常新的。高通量实验、机械臂大约只是在过去三年左右才被尝试。力场方面的专业知识，其中一些力场也是非常新的。但没错，我们觉得拥有一个实验室非常重要。拥有这种专注，并把所有这些人都聚在一起协同工作，非常重要。我想历史上最接近的例子是像贝尔实验室这样的地方，那里有令人难以置信的理论家、实验人员和化学家一起工作，取得了令人难以置信的成就。所以我们正试图在这里做同样的事，没错。
物理世界强化学习与不确定性
Brandon [00:02:17]：我待会儿想谈谈力场，不过
Swyx [00:02:19]：是的
Brandon [00:02:19]：在我们谈到那个之前。是的，它们是什么？但在我们谈到那个之前，先给点背景，我们的受众里有一些是 AI 工程师，有一些是科学家。对工程师们来说：在像 OpenAI 这样的大实验室工作，和你们在这里做的事，有什么区别？我想你已经暗示过这一点了，但也许可以更明确地说，你必须如何重新构建你的思考过程？
Liam Fedus [00:02:49]：嗯，我觉得一个有趣的点是，现在我们的强化学习环境实际上源自环境，源自我们的物理实验室。我们的数据来自我们的物理实验室，而这在某种程度上是我们的终极真相。仅仅针对某些论文或教科书里已知的答案做优化是不够的，因为我们在超越那些，我认为这是最大的区别之一。但与此同时，还有诸如不确定性下的决策之类的问题。所以当你针对数学做优化时，它有很高的精确度。你并不是真的在处理方差、不确定性或异常测量，而这对我们的流程非常关键。例如，当我们进行材料发现循环时，从炉子里出来的东西并不会自带标签，对吧？甚至标注过程都可能是随机且有噪声的。有时不同机器之间还会有偏差。也许我们的遥测数据不足。所以我们以为我们是在这个温度下运行的，但实际上温度偏离了某个差值。而一个能够接收这些噪声数据并像科学家那样做出智能决策的智能，需要一套不同的推理策略。所以我认为有大量共同点，比如标准的中期训练、强化学习、工具使用智能体的构建、方差缩减、确保基础设施在训练和推理之间没有不匹配。但我们必须超越这些，真正思考在存在高度不确定性的情况下，如何做出准确的工作？如何极其高效地利用有限的数据集？所以我们在显著扩大规模，但仍然，这与可以任意增加更多环境或更多 rollout 的数字环境非常不同。我们没有同样的能力。所以样本效率是另一个关键点。
为什么实验是有噪声且不完整的
Brandon [00:04:48]: 这对我来说似乎有点抽象，而且这个主题在科学播客上已经出现了一段时间，但我认为当你谈论实验不确定性时，解释一下你正在做的具体实验过程会是什么样子，以及在进入 AI 部分之前，人类会如何处理这个问题，可能会有所帮助。你有没有一个具体的例子，说明你制造的一种材料，以及你在过程中会经历什么样的测量不确定性？
Ekin Doğuş Çubuk [00:05:17]: 我认为这有很多维度。所以其中之一是，我认为每个科学实验都必须进行某种降维，这与编码或数学非常不同。所以，当你做数学或编码时，所有上下文都可以提供给人类或 LLM。它是一堆公理，一些人们从中推导出的推论。它是一堆函数和 API 调用。所以你需要推理的一切都可用，然后你只需要非常聪明地弄清楚。在物理学中，如你所知，我们开始时拥有的原子比我们能在计算机上存储的还要多。所以显然，我们必须从代表系统的原始维度和比特数，转变为我们可以放入计算机的比特数。我想这就是热力学的原始前提。他们最初对蒸汽机如何工作感到非常困惑，但后来发现你可以找到五六个热力学变量，很好地解释正在发生的事情，这真是不可思议。这就是物理学的美丽之处。
Brandon [00:06:13]: 你有一个充满原子的房间，有 10²³ 或 10²⁷
Ekin Doğuş Çubuk [00:06:19]: 没错
Brandon [00:06:19]: 或者这个房间里有那么多原子，然而你需要温度和压力以及简单的变量，这就能告诉你一切。不，它告诉你大部分你需要知道的
Ekin Doğuş Çubuk [00:06:28]: 很多
Brandon [00:06:29]: 在大多数情况下需要知道，对吧？
Ekin Doğuş Çubuk [00:06:30]: 是的。
Brandon [00:06:30]: 所以假设你制造一种材料，你可以做类似看 X 射线结构的事情，而这件事是非常有损的投影，对吧？它实际上不是一个结构，或者它不告诉你真实的结构。它告诉你的是其中一些主成分。那么，人类如何利用这些简单的主成分，然后对它们进行推理？然后你如何将其扩展到 AI？
Ekin Doğuş Çubuk [00:06:58]: 所以在理论物理方面，我们已经确定能量、能量的平均值以及能量的涨落非常重要。很多热力学都是由此推导出来的。然后我们意识到反应势垒也很重要。动力学非常重要。所以人类可以尝试通过提出一些降维的描述符来推理这个极其复杂的系统。人们也非常喜欢原子尺度的图像。你会思考局部邻域是什么样子，它们是八面体还是四面体，这对化学家尤其有益。好了，这是理论方面。在实验方面，你会尝试收集尽可能多的数据。所以你会有一个实验记录本，写下你看到的一切等等，但当然要明白你会遗漏很多。所以就像 Liam 提到的，会出现所有这些问題。例如，你的炉子会随时间退化，因为当你使用炉子时，你烘烤的一些东西会蒸发并覆盖加热元件。所以每次使用炉子，它都会变得越来越糟。另一个问題是你有一个炉子，但温度并不完全均匀，所以你在炉子中放置的位置会影响结果。我们还没有这个问題，但也许有一天会有。光学仪器很容易受到振动的影响。所以每次有人走动。我记得我读博士时，一个实验室有这个大问題，因为有时他们的结果会与其他时候非常不同，他们最终发现是在夜间的某个时间，楼上有人在走动，那种振动影响了激光装置。所以是的，科学非常困难，但这就是它特别之处，对吧？Liam 和我一直在讨论的另一件事是，目前很多改进都集中在数学、编码和理论计算机科学上，因为这些对 LLM 来说更容易。但在现实生活中，大多数需要智能的事情实际上更像科学。有很多不确定性，有很多噪声，有很多缺失的上下文，但你必须成为那个智能体，弄清楚下一步该做什么。
Liam Fedus [00:08:53]: 我认为一个非常有趣的点是，做数学、做理论计算机科学、做这类事情的最佳推理策略，可能并不是做科学的最佳推理策略。
Ekin Doğuş Çubuk [00:09:04]: 是的。
材料发现循环：合成与表征
Brandon [00:09:05]: 当你的输出是一个有噪声的晶体结构，或者一些我不知道的集体变量之类的，你如何定义 RL 奖励函数？就像。
Liam Fedus [00:09:17]: 也许，我会讲这个循环中的一个部分，然后深入探讨一下。所以，作为材料发现循环的一部分，我们首先得弄清楚要制造什么。这涉及到原子组合的稳定性，以及我们是否预期它具有我们想要的特性。所以我们不只是想要一种新颖的原子构型。我们希望它们组合在一起时，能具有我们感兴趣的特性。接下来是如何合成它？你实际上怎么制造这个东西？这同样不是 trivial 的。所以即使你知道某种东西能保持稳定，实际的过程，即需要什么样的处理条件才能真正实现它，也是高度非 trivial 的。但同样，这个东西。一旦你完成了这两个步骤并制造出了某种东西，它并不会带着标签出来，所以你必须对其进行表征，弄清楚你制造了什么。因此在这种情况下，与其考虑强化学习环境，即我们只是启动一个实验，等几天，然后尝试根据你是否找到了室温超导体来进行更新，这完全不可行。
Brandon [00:10:15]: 是的。我正想问这个。是的，你不能让你的 GPU 闲置一周之类的。
Liam Fedus [00:10:19]: 是的。是的，这完全不可行
Brandon [00:10:20]: 是的
Liam Fedus [00:10:20]: 因为你没有足够的智能体来充分降低方差。在不同仪器之间的 rollout 时间。你在等待。实际合成或炉子时间存在一些物理限制。所以它太慢了，噪声太大。因此，我们思考 AI 项目的方式基本上是围绕我们产生的这组数据构建智能体。所以在表征的情况下，你寻找的是奖励识别实际存在的相的强化学习环境。你能否根据原始实验数据有效地拟合？所以我们这样做的方式是，我们会向材料发射 X 射线。X 射线的频率或波长大致可测量或与原子间距相当，因此你会得到漂亮的衍射图案，这有点像晶体结构的指纹。这样做，识别实际存在的东西是高度非 trivial 的。所以我们早期的一些 AI 工作基本上是在构建系统来做这件事。这也让我们能够更快地推进，因为如果你运行这么多实验，你很快就会在理解所产生的东西的能力上遇到瓶颈。但同样，现在强化学习环境相当直接，即你将奖励识别存在的相。你将惩罚虚假的相或任何没有真正化学合理性的东西。所以这就像一个简洁的小例子，你可以为不同的部分这样做。当你把这些拼接在一起时，那就是端到端的发现循环。
基于新颖实验数据训练，而非记忆答案
Brandon [00:12:02]: 好的。你
Liam Fedus [00:12:02]: 也许另一个
Brandon [00:12:03]: 是的。哦，抱歉
Liam Fedus [00:12:03]: 另一部分将是，随着我们积累大量实验数据，你可以把它看作是对世界状态进行时间戳标记。所以你可以说：“在这个日期，这是我们截至那时的实验证据。”然后你可以构建强化学习环境，说：“好吧，给定这些实验证据和选择，科学家做出的下一个选择是什么，或者那个实验的结果是什么？”这真的很有趣，因为这允许我们进行一些在外部更难做的项目。因为如果预训练模型已经记住了其中一些数据，然后你试图在此基础上进行强化学习，如果它已经知道答案，而你构建依赖于该答案的强化学习任务，它就可以假装工作。所以，然后，因为它就像，哦，好吧，它已经知道答案了，所以它实际上不必做那些困难的物理推理。它不必实际进行计算模拟。它得到了正确答案，然后你强化那个策略，你提升那些推理策略的权重。那些推理策略不会泛化到新系统，所以对我们没用。但随着我们积累实验数据并将事物拼接在一起，拥有整个过程的谱系，这使我们能够构建在其他地方不可能的新类型环境。
为什么已知物理学并不意味着完美模拟
Swyx [00:13:27]: 我想问我的问题
Brandon [00:13:28]: 说真的
Swyx [00:13:28]: 但我担心我会把它带偏。我还是直接问吧，因为我，再次，我来自非科学家，但工程方面。我熟悉机器学习方面，但不熟悉物理方面。好吧，这有几个版本，但我认为基本问题是，我们不是已经弄清了大部分物理定律吗？为什么我们还没有完美的模拟器？
Brandon [00:13:50]: 哦，这是个好问题。
Swyx [00:13:52]: 就像，这很基本，就像。他就像，“哈，那很可爱。”但就像
Brandon [00:13:56]: 不，这是个好问题
Swyx [00:13:57]: 老兄，我有所有这些物理书。你什么意思，你什么意思我们还没完成？
Ekin Doğuş Çubuk [00:14:03]: 是的，所以这里有几件事在发生，对吧？所以
Swyx [00:14:07]: 是的
Ekin Doğuş Çubuk [00:14:07]: 我认为首先，我们绝对没有完成任何物理定律。即使是
Swyx [00:14:12]: 在量子尺度，好吧，但是。也许在非常大的尺度，但在人类尺度，在材料尺度，我们，我们。还剩下什么？
Ekin Doğuş Çubuk [00:14:22]: 我们没有。所以这是个非常好的问题。而且非常有趣的是为什么我们还没完成，对吧？所以有一个非常常见的故事人们谈论。当狄拉克在弄清楚量子力学时，大约1920年代末，他写了一本关于量子力学的教科书，他把它框定为一切都完成了。然后有一句话人们喜欢引用。我不确定实际上有多准确，但显然狄拉克说，“剩下的都是化学。”意思是说他能解出氢原子
Swyx [00:14:50]: 把所有东西组合起来就行。
Ekin Doğuş Çubuk [00:14:51]: 是的。他也许能解决一维氢原子的链，但当前他无法解决，比如说，氮与氧的相互作用，但那没关系，那只是化学。
Swyx [00:15:00]: 是的。
Brandon [00:15:00]: 物理学家真的很喜欢单原子代表。非常简单的系统。
Ekin Doğuş Çubuk [00:15:05]: 简单系统。
Brandon [00:15:05]: 而这就是你能解决的全部。
Ekin Doğuş Çubuk [00:15:07]: 球形奶牛。
Brandon [00:15:07]: 是的，球形奶牛。
Ekin Doğuş Çubuk [00:15:08]: 但事实证明，我认为我们在过去一百年学到的是，首先，那并不正确。它并不像只是氢的简单扩展。而且它也不仅仅是。其余的不只是化学。那里实际上有很多物理。我想我们仍未解决的问题之一就是高温超导。但还有很多其他
Swyx [00:15:25]: 对你来说高是175、200？
Ekin Doğuş Çubuk [00:15:28]: 哦，当人们说高温超导时，他们指的是非常规超导。所以有常规超导，那主要是由电子-声子耦合驱动的。在那种情况下，你可以看到同位素效应。如果你取相同的系统，但只是其中一个元素的重量不同，你可以看到超导温度以你预期的速率下降。所以那个是常规的。但像铜氧化物这样的高温超导体，比如那些高于77开尔文的，比如93开尔文，事实证明不遵守那种物理。但我们不知道它们遵守什么物理。它们只是不可思议的超导体。但不仅仅是那样。所以那绝对是这些非常流行的、我们还没有理论的主题之一。但还有很多其他我们不了解的事情。例如，即使是简单量子力学系统中的强关联，我们目前也无法模拟。密度泛函理论是一个非常强大的工具。它在某些事情上非常准确，但只要。一旦有强电子关联，它实际上就无法捕捉某些效应。有太多东西需要弄清楚。这就是为什么我真的很兴奋AI被应用于这个领域，因为从理论上、计算上和实验上，都有太多东西需要弄清楚。我们在这里所做的每一项改进都应该改善人类生活，因为，我们越能理解材料、固态物理，我们就能为它们制造更好的设备。
Brandon [00:16:41]: 有菲尔·安德森的一句名言，就是“多即不同”，这基本上是说你可能理解某事物的所有基本定律，但当你添加许多东西时，它们的行为在性质上截然不同，与简单物理应该告诉你的不同。所以理解许多事物如何一起行为，看起来应该很简单。起始定律很简单，但集体行为如此复杂，以至于真的很难建模。
Ekin Doğuş Çubuk [00:17:07]：它是涌现的、普适的，这太疯狂了。我们在深度学习模型中也看到了这一点，对吧？我们在深度学习中到处都能看到幂律。我们并不理解，但它非常让人联想到物理学，那里也存在涌现和普适性。
Brandon [00:17:20]：物理学界已经有很多关于神经网络中涌现现象的理论论文。
Ekin Doğuş Çubuk [00:17:24]：是的，绝对是这样。
Brandon [00:17:25]：是的。
合成预测与物质编译器
Swyx [00:17:26]：然后我的另一个追问也是关于这个过程，我们就叫它，嗯，过程和最终结果吧。我们就这么说吧，你想要针对材料实现这些目标属性，然后你必须找出达到目标的工艺过程。把这些东西分开来考虑是否值得，这样你就可以……你只需训练一个模型，给定任何你想要的理论最终状态，它就能完美地逆向工程出任何工艺过程？这有意义吗？
Ekin Doğuş Çubuk [00:17:54]：也许回到我们之前关于这个的讨论，既然原子数量超过阿伏伽德罗常数，我们永远无法知道所发生事情的确切情境。我认为这可能就是为什么我们永远无法完美地逆向工程一切的原因之一。但我们只是试图进行足够好的逆向工程，基本上能够改善材料的性能。
Swyx [00:18:13]：是的。
Brandon [00:18:13]：你问的是是否存在一个正向模型，你输入一个输入，就能预测输出，而你想要理解
Swyx [00:18:19]：是的。
Brandon [00:18:20]：你想要创建一个类似代理模型的东西，它能够理解
Liam Fedus [00:18:22]：输入是那个的最终状态，即原子的这种构型，然后预测是
Swyx [00:18:27]：因为那样你就可以，你几乎可以把工作分开。你可以让一个团队做工艺方面，另一个团队做属性方面，然后
Liam Fedus [00:18:33]：是的
Swyx [00:18:33]：你知道，就让它们赛跑。
Liam Fedus [00:18:36]：是的，我认为在这些方面我们可以取得很多进展。把这些智能体工作流拆分到不同的领域，我认为在合成预测方面我们也可以做很多实证工作。
Swyx [00:18:49]：是的。
Liam Fedus [00:18:49]：而接收那种原子构型，你能使用什么工具？如何，鉴于这些实验证据，鉴于先前的文献、先前的论文，你实际上如何达到这些东西？而最终你是在测试，你是否真的能够复现它？
Swyx [00:19:04]：我之所以这么说，也许也是从公司的角度来思考。你知道，就像台积电在某种程度上是半导体的晶圆厂，但他们不设计半导体，你可以成为材料领域的台积电，人们来找你说，“嗯，这是我想要的，”然后你就能找出工艺过程，然后
Liam Fedus [00:19:21]：是的，就像一台物质编译器。
Swyx [00:19:23]：是的。
Liam Fedus [00:19:24]：所以我想，给定这些需求
Swyx [00:19:26]：很有《星际迷航》的感觉。
Liam Fedus [00:19:27]：是的。但问题是，给定这些要求，这真的是一组合法的配置吗？这能存在吗？
Swyx [00:19:33]：它是一个好的目标函数吗？我不知道。它太大了。
Liam Fedus [00:19:37]：是的。我们有个使命叫合成超级智能，所以我觉得这跟它是一致的。
相变、X 射线衍射与 RL 奖励
Swyx [00:19:44]：是的。
Brandon [00:19:44]：我想深入追问一下你刚才提到的一点，就是你谈到了相。我觉得这正好呼应 Doğuş 关于什么是涌现的评论。相变这个概念对很多听众来说可能比较陌生。所以你能解释一下什么是相变，以及为什么它对 RL 环境这类东西来说是一个有用的信号吗？
Ekin Doğuş Çubuk [00:20:08]：相变非常迷人。我觉得人们生活中最容易产生共鸣的可能是冰融化，或者水沸腾。你给冰升温，它看起来还是冰，行为也像冰，但到了某个温度，它就不再升温，然后变成液体。于是它突然从固态相变成液态相。另一个跟我们非常相关的当然是伊辛模型。我觉得计算机科学和数学家可能用不同的名字研究它，但基本上，你可以有自旋向上和向下，如果两个都向上，或者一个向上一个向下，它们有不同的相互作用项。在足够高的温度下，它们通常只是随机地向上或向下。熵占了上风。你降低温度，它看起来还是一样。但到了某个点，它们突然开始完美地对齐。是的，相变。相变非常好，因为它让物理学家研究某些东西时更轻松。它们往往具有某些空间关联，这对我们也很有帮助。好，所以在我们的实验室里，相变通常是因为我们混合不同的前驱体，基本上就是晶体，然后我们升高温度或做其他事情促使它们反应，然后原子开始反应，形成这种新晶体。这会体现在 XRD 图谱中，因为通常如果原子的几何结构发生剧烈变化，XRD 图谱也会变化很大。但在典型的实际材料发现活动中，真正发生的是：你第一次尝试某样东西时，它不成功，结果并不是清晰的是或否，而通常是混合相。它通常会有一些前驱体。它可能有一些非晶相，然后还会有一堆你也许并不想制造的相。这就成了一个挑战，也是为什么我们发现很难把模拟和 AI 结合起来做表征。模拟总是可以说：“哦，这个相看起来不像我们预测的，但这是它的一个轻微变体。让我现在对它做力场计算，看看它是否仍然稳定。”或者 AI 可以说：“基于之前的实验和这一次，这大概不是我们想制造的相，所以我们改一下条件。”
Liam Fedus [00:22:18]：是的，而且在某些情况下
Swyx [00:22:19]: 对听众来说，XRD 就是 X 射线衍射，你已经描述过了。
Liam Fedus [00:22:22]: 是的。在某些情况下，存在一种以前从未被记录过的相。它在任何论文或数据库中都不存在。因此，AI 实际上必须去使用这些工具，说：“那么，什么样的原子构型才能真正解释这些相？”但这对于指导科学过程变得极其重要，因为假设我们心中有一个目标相。我们真的需要像爬山一样去优化它，以便获得更好的测量结果。所以如果它是百分之一，而我们想，实际上，我们想提高这个相的纯度，我们就需要对这个实验数据有一个非常好的测量，来说明什么存在、什么不存在。
Brandon [00:22:57]: 所以优势之一就是你有一个可以观察的独特变量。回到关于如何处理不确定性的说法，你的答案中似乎有一部分是让你可观察的特征变得有些明确无误。那里没有解释的余地。是这样吗？这就是背后的逻辑吗？还是说这实际上只是你需要找到一个新的相，所以这才是你关心的东西？
Liam Fedus [00:23:20]: 我认为在很多这些情况下仍然存在模糊性。一个非常笨的办法是重复实验。
Brandon [00:23:25]: 是的。
Liam Fedus [00:23:26]: 我们当然会做重复实验。
Brandon [00:23:27]: 是的。是的。
Liam Fedus [00:23:28]: 但不，我认为那里——你知道，这是一项具有挑战性的任务，因为它不只是一个完全确定性的过程，而且存在模糊性以及其他问题。两个不同的相实际上可能与同一个图谱一致。
Brandon [00:23:42]: 是的。
Liam Fedus [00:23:42]: 所以我认为这就是为什么系统使用一些化学直觉来说“好吧，考虑到合成条件，考虑到这些先验知识，这将是极不可能的”非常重要，这可以用来帮助消除歧义。
化学先验、多模态测量与模糊性
Swyx [00:23:58]: 所以你是在某种程度上注入先验，基于你所期望的。
Ekin Doğuş Çubuk [00:24:03]: 是的。我认为热力学是最大的先验，对吧？
Swyx [00:24:06]: 确实。
Ekin Doğuş Çubuk [00:24:07]: 然后物理学也是一个很大的先验。我们在这里真正受益的另一个方面是多模态或材料表征。所以我们可以做 XRD，是的，有些相在 XRD 图谱上可能看起来相似，但我们还可以测量它们的电学性质、磁学性质。我们可以测量它们的形貌，比如某种电子显微镜。然后一些在 XRD 上看起来相似的相，在其中一些维度上会看起来不同。这种多模态确实很有帮助。而且，AI 在这里真的很有帮助，因为人类的上下文和计算能力都是有限的。所以如果你同时给人类 10 种不同的模态，说“一致地分析所有这些”，这有点困难。但对 AI 来说，这实际上并不是超级智能才能做到的。这很棒，是的。
Liam Fedus [00:24:46]: 是的，这就像超级智能，它能进行大量的计算。它能审视一切。这就引出了那句关于在不确定性下做出更好决策的说法，现在它正在纵向地、跨不同实验、跨重复实验，将来自许多不同仪器的信号拼接在一起，然后得出这个更底层的更好模型。
Swyx [00:25:08]: 是的。
Liam Fedus [00:25:08]: 而不是仅仅过度依赖来自单一仪器的单一测量结果。
信息限制、遥测与麦克斯韦妖
Swyx [00:25:12]: 让我就，这么说吧。你知道，我想你之前也说过，数据量就是比任何合理的计算机或流程能容纳的都要多。所以到了某个时候，即使你拥有所有这些重复实验，即使你大概有十种不同的方法来测量同一件事，你也必须丢弃数据。因为万一你的某个温度测量设备出错了呢？那你什么时候把它扔掉？你怎么决定该怎么做？作为一个渴望数据的机器学习从业者，我只想要一切，对吧？然后我就，你知道。
Ekin Doğuş Çubuk [00:25:37]: 所以，是的。
Swyx [00:25:37]: 摇摇摆摆，学习一切。我不知道我不知道什么，就把它扔进去，扔进去，让机器去处理。
Ekin Doğuş Çubuk [00:25:42]: 是的。所以我们不丢弃任何数据。我想也许我的意思是，如果上帝在观察这个实验。
Swyx [00:25:50]: 是的。
Ekin Doğuş Çubuk [00:25:50]: 实验中存在的数据比人类能塞进计算机的还要多。不幸的是，我们这些凡人无法观察到所有的原子。所以，即使数据量比我们能存储的要多，我们实际上也无法获取那些数据。例如，我无法追踪所有原子的位置。所以不幸的是，我们仍然处于这样一个阶段：我们收集的任何数据都非常宝贵，我们不会删除任何数据。是的。
Brandon [00:26:11]: 也许有一种思考方式，把它和人工智能方面联系起来，就是想想探测，比如对基础模型进行线性探测，对吧？这个模型拥有所有这些数据，所以上帝拥有材料正在发生什么的完整图景，而实验就是这个小小的线性探针，它只有八位左右的输出。所以你把这个巨大的东西粗粒化，缩减成一点点信息，然后现在，要么传统上是人类的工作，要么现在是你的智能体的工作，就是从这八位信息中重建。
Liam Fedus [00:26:43]: 这是个类比，是的。
Brandon [00:26:44]: 内部实际正在发生什么。
Liam Fedus [00:26:45]: 这是一个类比，你无法完全看清每个参数权重，也无法看清该输入的每个激活值。是的，你得到的是它的某个子集，或者整个东西的一些粗粒化特征。
Ekin Doğuş Çubuk [00:26:58]: 是的。但所以你可能会发现。你很可能确实会发现，麦克斯韦妖的论点非常有趣，因为它那种。
Swyx [00:27:05]: 你能，你能再背诵一下吗？因为我不。他之前提到过。我老是忘记。
Ekin Doğuş Çubuk [00:27:10]: 所以，在19世纪，人们意识到一个系统的熵必须增加，或者对于一个封闭系统来说。
Swyx [00:27:18]: 是的。
Ekin Doğuş Çubuk [00:27:18]: 或者保持不变，但它不能下降。
Swyx [00:27:20]: 是的。
Ekin Doğuş Çubuk [00:27:21]: 这是热力学第二定律。有一个思想实验问，如果有一个非常微小、无所不知的恶魔，能够观察原子，每当一个原子能量低时，就打开一扇门让它进入另一个房间，并一直这样做以降低系统的熵，这会打破热力学第二定律，那会发生什么？很长一段时间里，我认为对于为什么会这样，并没有一个非常令人满意的答案。但后来在1900年代，对我们来说更近些时候，有一篇更近期的论文，Landauer提出了一个论点：你必须花费能量来删除信息，而麦克斯韦妖通过做这件事基本上会接触到如此多的信息，以至于他必须删除一些。而当他删除信息时，他必须花费能量，这反过来会确保熵增加。所以，尽管我们还没有达到可以成为麦克斯韦妖的水平，我认为你的问题非常有远见，指向了一个我们存储了太多数据以至于必须删除一些的未来。
Swyx [00:28:17]: 是的，不，我并没有，我显然没有考虑到那种程度。这确实让我想起了空调。空调差不多就是这样工作的。但另外，我认为这只是一个问题：你为什么不直接买下世界上所有的传感器，然后过度地、荒谬地过度装备、过度复制一切呢？
Liam Fedus [00:28:33]: 哦，我们就是这么做的。
Swyx [00:28:34]: 对吧？是的，好的，是的，好的。这就对了。
Liam Fedus [00:28:35]: 是的，没错。
Swyx [00:28:35]: 好的，这样就行得通。对。好吧。
Liam Fedus [00:28:37]: 是的，我认为，我认为这是一个
Swyx [00:28:39]: 确认了。
Liam Fedus [00:28:39]: 是的，没错。你能看到单个原子吗？那是不可行的。但不，我们添加了大量的遥测数据，这基本上减少了隐藏变量的数量，如果你愿意这么说的话。
Swyx [00:28:49]: 是的。是的。只是出于好奇，你们担心你们只是在加利福尼亚吗？你们不想在尼泊尔和澳大利亚做这个吗？
Ekin Doğuş Çubuk [00:28:59]: 是的，我们可以。我们有计划开设更多实验室。
Swyx [00:29:02]: 因为显然重力、你面向哪里、你在哪里。你在哪里很重要。
Ekin Doğuş Çubuk [00:29:07]: 而且专业知识也会变化，对吧？我们这里有某些类型的专业知识。例如，斯坦福物理系、工程系有某些类型的专业知识，这对我们真的很有帮助，因为我们可以从那里招聘。我们的研究人员可以合作。正如你所说，地理位置可能很重要，但不同的城市在物理、化学方面可能有不同的专业知识。
Swyx [00:29:27]: 是的。哦，我们有Zoom
Ekin Doğuş Çubuk [00:29:28]: 是的。
Swyx [00:29:28]: 为此。
Ekin Doğuş Çubuk [00:29:29]: 是的。但仍然
Swyx [00:29:31]: 这不像你的实验会受到你所在位置的影响。
Liam Fedus [00:29:34]: 是的。是的，Periodic的未来不是一个单一的实验室。
Swyx [00:29:38]: 是的。你得坐上你的私人飞机去太空，你知道的。
Liam Fedus [00:29:42]: 对。是的，我认为有必要建立这种实验室网络。而且，就像 Doğuş 刚才指出的那样，基于专业知识或材料限制，每个实验室都有其最佳位置。有许多不同的限制条件，比如许可。
将模拟与物理实验相结合
Brandon [00:29:56]: 好的，那么其中之一是。我一直好奇的一件事是，你如何处理计算工具？或者你如何将计算工具整合到决策工作流程中，特别是当计算和实验并不一定吻合时？从高层次理解的角度来看，这也涉及理论。那么，这些是如何协同工作的呢？
Ekin Doğuş Çubuk [00:30:22]: 是的。所以，这里非常有帮助的一点是，有些事情在模拟中更容易做，也更准确。有些事情在实验中更容易做，也更准确。一般来说，当然，实验更准确，但某些测量在实验中非常难以进行。所以你只能做一个糟糕的版本，得到不准确的数据。让我给你举几个例子。其中之一是，我们使用密度泛函理论来估算材料的生成焓，以及
Brandon [00:30:51]: 抱歉，稍等一下。你能解释一下密度泛函理论和生成熵吗？
密度泛函理论与生成焓
Ekin Doğuş Çubuk [00:30:55]: 是的，当然。
Brandon [00:30:56]: 抱歉，是焓，对。
Ekin Doğuş Çubuk [00:30:57]: 是的。所以，密度泛函理论可能是材料领域最常用的模拟方法。它源自 Hohenberg–Kohn 定理，Kohn 因此获得了诺贝尔奖。他。当他意识到这一点时，当然，量子力学非常昂贵，因为你基本上要为每个波函数处理这个指数级的希尔伯特空间，所以你必须解决这个非常困难的问题，尤其是随着系统尺寸的增长。Kohn 和 Hohenberg 意识到，事实证明你不需要在指数空间中进行计算。你只需要考虑电荷密度，至少对于量子力学系统的基态而言。所以这个定理其实相当简单。我想甚至一个物理学本科生都能理解。但核心思想是，量子力学系统基态的所有性质都是电荷密度的泛函，这很了不起，因为电荷密度只是一个三维对象，而希尔伯特空间和波函数是指数级的。所以这是一个非常有趣的观察。然后 Kohn 发表了另一篇论文，我想是和她的博士后 Sham 一起，Kohn–Sham 波函数，提供了一种实用的方法来求解材料量子力学性质的一些近似。如今，我们在实验室内部和外部都大量使用它。然后生成焓是我们可以进行的测量之一，用来理解材料的能量。基本上，我们想问，这种材料具有的能量是多少？因为如果这个能量高于这些原子可以形成的其他材料，那么这种材料就不太可能被制造出来。所以稳定只是意味着它处于其他材料及其能量的凸包上。这样够清楚了吗，还是我们应该。
Brandon [00:32:42]: 也许一个简单的高层次观点。所以密度泛函理论是一种方法，把在电子数量上呈指数级困难的东西
Ekin Doğuş Çubuk [00:32:51]: 对
Brandon [00:32:52]: 或者在你的系统尺寸上，然后通过引入一些误差，近似地将其降低到通常可能是 n 的三次方近似。所以现在你可以计算那些，你知道。你现在可以计算那些原本无法计算的东西，但要以某种代价，对吧？如果你能完美地做到，你就会——我们就不需要实验室了。但我们做不到。这可能是对我们实际上第二集的一个很好的呼应。Heather Kulik 曾说过，她有一句名言，“材料领域没有 AlphaFold，不仅从计算角度来说，而且真实基准根本就不存在。”就像，我们并不真正知道材料的晶体结构是什么样子，而 DFT 通常是我们最好的途径。
Ekin Doğuş Çubuk [00:33:34]: 是的，我想是这样。也许我要补充两点，DFT 从定义上讲不一定必须是近似。Hohenberg–Kohn 定理表明它
Brandon [00:33:42]: 它不必是
Ekin Doğuş Çubuk [00:33:42]: 它可以是精确的。但我认为，正如你指出的，交换关联泛函我们目前还无法获得。然后对于基于电荷密度的模型，我们不知道动能泛函是什么。而且即使我们有完美的 DFT，我仍然认为我们需要实验室，因为我们。即使我们有完美的 DFT，我们也不能把 10 的 23 次方个原子放进计算机里。
Brandon [00:34:03]: 是的。
Swyx [00:34:04]: 你对那个数字非常着迷。
Brandon [00:34:06]: 但是。不。好吧，好吧。是的。但是随着计算能力呈指数级增长，你可以认为在 n 立方缩放规模下，最终原则上你可以计算事物。你只需要等到暂停足够长的时间。
Ekin Doğuş Çubuk [00:34:17]: 哦，是的。我们需要大量的计算机，是的。
Brandon [00:34:19]: 但是是的。但是只要
Ekin Doğuş Çubuk [00:34:20]: 这需要一段时间。
Brandon [00:34:20]: 是的。但是只要 n 立方是一个好的缩放，你就可以，你可以。
Swyx [00:34:24]: 作为非科学家，我会提出一个观察，我不知道你是否想讨论或什么的。我来自金融领域。我们有高斯 copula，这是一种为信用违约互换定价的方法，通过做相关性并将一切简化为一个单一内核。而且它非常。感觉在精神上也类似于 VAE，就像将所有这些事物浓缩成一个单一参数。我想知道是否到处都有同样的技巧。你的听起来比我的更高维，VAE 就像一个 sigma 的东西。但听起来一样？
Liam Fedus [00:35:01]: 是的，电荷密度文件仍然是一个相当大的文件。
Swyx [00:35:04]: 是的。对。你有点含糊其辞，但已经足够好了。这是，这是最高位。是的。
Ekin Doğuş Çubuk [00:35:10]: 我个人不知道有更好的方法来预测新材料的稳定性。它肯定不完美，但肯定比我所能想到的其他方法更好。是的，所以回到你的问题，我们使用它们的原因是对某些事情来说模拟比实验更容易，而且它们很好地互补。然后我们可以在循环中这样做。我们觉得模拟本身永远不够，但在模拟、AI 和实验的循环中，我认为我们可以比以前更快地取得进展。
当模拟失败时：微观结构和校准
Brandon [00:35:41]: 模拟让你做的一件事是更快地扩大规模。因为现在计算能力更容易上线
Ekin Doğuş Çubuk [00:35:48]: 是的
Brandon [00:35:48]: 比引入新实验室更容易。你如何避免。或者假设你扩大了一堆 DFT，你如何避免让你的模型过度依赖它，并且仍然基于现实世界最终是你关心的真相？如果你。你知道，你如何平衡这一点？
Ekin Doğuş Çubuk [00:36:05]: 嗯，我们也扩大了实验室。
Brandon [00:36:06]: 是的，你扩大了实验室。
Ekin Doğuş Çubuk [00:36:07]: 对，是的。
Brandon [00:36:07]: 但你可以。我可以想象在现代计算能力下将 DFT 扩大到数百万、数亿次，而你的实验室，我想象，每天数百、数千次？我不知道。仍然相当。嗯，即使在真正高通量的情况下，相比之下你仍然相当受限，对吧？
Ekin Doğuş Çubuk [00:36:27]: 是的。一方面是人力，但 LLM 也相当清楚 DFT 的局限性。例如，即使我们可以像你说的那样，用 DFT 进行一亿次试验
Brandon [00:36:39]: 是的
Ekin Doğuş Çubuk [00:36:39]: 我们实际上从未能尝试它们的微观结构。微观结构这个概念是指，DFT 通常模拟完美晶体，但晶体实际上通常并不完美，而它们的微观结构，即中等尺度上的结构，确实会影响性质。当然还有其他问题，比如超导温度等某些性质无法用 DFT 轻松模拟。再说，即使你用 DFT 做一亿次试验，也不够。所以我们必须依赖启发式方法、实验等等。
Liam Fedus [00:37:08]: 所以从所有这些计算到实验室实际执行之间，存在一个巨大的筛选过程。
Swyx [00:37:13]: 这是由人工介导的，还是 LLM，还是两者混合？
Ekin Doğuş Çubuk [00:37:17]: 可以是混合的。
Swyx [00:37:17]: 是的。
Ekin Doğuş Çubuk [00:37:18]: 例如，很长一段时间里，Materials Project，也就是最好的开源 DFT 数据库，使用实验校准来校准 DFT。所以他们实际上会用实验数据来校准 DFT。我们在这里内部也这样做。每当我们对某个化学体系感兴趣时，你会得到实验、得到模拟，然后你可以对它们进行校准。
为什么无机材料与生物学不同
Brandon [00:37:37]: 这是另一件事。我和很多生物学家朋友聊过，当你提到材料时，他们总是非常困惑，你不能只做一个 XRD 结构就真正知道结构是什么。因为如果你在生物学领域，你可以查看蛋白质的晶体结构，通常或多或少能以亚埃级、好吧，埃级精度重建。材料 XRD、X 射线衍射——X 射线衍——不
Ekin Doğuş Çubuk [00:38:05]: 衍射。
Brandon [00:38:06]: 衍射，抱歉。材料的实验 XRD 与，比如说，生物学领域相比。你丢失了什么信息，为什么这有点像有损投影？
Ekin Doğuş Çubuk [00:38:16]: 回答这个问题很难不显得疯狂。但你不觉得吗，有机化学和生物学几乎源于一个更低的 VC 维度，就像更低的柯尔莫哥洛夫复杂度模型？在那里，你几乎可以表示为一维序列，而且它几乎可以被编译。所以从某种意义上说，它在柯尔莫哥洛夫复杂度意义上更简单。
Swyx [00:38:38]: 一维？
Ekin Doğuş Çubuk [00:38:40]: 就像 DNA 只是
Brandon [00:38:43]: 是的。
Ekin Doğuş Çubuk [00:38:44]: 或者 RNA。
Brandon [00:38:44]: 我认为大多数人会认为它更多是二维的，但
Ekin Doğuş Çubuk [00:38:47]: 好吧，是的。我不是生物学家，所以
Brandon [00:38:49]: 是的。
Ekin Doğuş Çubuk [00:38:50]: 我确信他们是对的。但就无机化学而言，这简直太疯狂了。它显然不是来自那种低柯尔莫哥洛夫复杂度的模型，因为三维无机晶体可以容纳金属、绝缘体、超导体、金刚石。差异真的非常惊人。人们尝试过，但我们始终找不到一维表示。为三维无机晶体找到SMILES字符串非常困难。而且原子之间相互作用的方式相当奇特，比如共价键、离子键、金属键。不过，是的，这非常有趣。我有时会想，你有没有研究过固态电解质的问题？为什么人们不能用固体替代电池中的液体电解质？会形成所有这些枝晶。基本上，锂会形成这些结构，进入固态电解质并使其开裂。你看着这个，会觉得与生物学相比这太原始了。在生物学中，这些系统能实现如此复杂的纳米技术——技术进步，而我们甚至无法让两个界面协同工作，因为锂会破坏它。是的，但我想无机材料可以非常耐高温。你可以造航天飞机。你可以造硅，实现摩尔定律。所以，是的，这是一种权衡。
Brandon [00:40:05]: 是的，我们和其他几位嘉宾讨论过这个。在我看来，一个很大的区别是，生物学中我们有一套工具包，基本上可以借用数百万年进化的成果
Ekin Doğuş Çubuk [00:40:19]: 是的
Brandon [00:40:19]: 那已经给了我们所有工具，我们只需复制它们。但进化也限制了蛋白质等的VC维保持在相当低的水平。像细胞这样的大规模系统涌现出的超越行为可能复杂得多。
Ekin Doğuş Çubuk [00:40:36]: 是的，还有意识。
Brandon [00:40:37]: 是的。是的，但也许更具体地说。如果你只看一个X射线，你有一个完美晶体，你对它做XRD，你得到光谱。那实际上仍然不能唯一地告诉你晶体结构是什么，对吧？
Swyx [00:40:54]: 不能。
Brandon [00:40:55]: 所以，那是
Swyx [00:40:56]: 什么？
Ekin Doğuş Çubuk [00:40:57]: 那只是平均值。
Brandon [00:40:58]: 是的。
Ekin Doğuş Çubuk [00:40:58]: 那只是平均行为。但比如，歧化反应是非常复杂的事情，对吧，有可能你有一个完美晶体，平均来看它呈现这个完美的XRD。但实际上，还有另一种模式在发生，原子稍微向右或向左偏移，但平均来看它们在中间。是的，生活艰难啊，是的。
Brandon [00:41:16]: 而对于蛋白质，你基本上知道存在某种底层蛋白质遵循某些规则，所以你可以将前向模型与之结合。
Liam Fedus [00:41:23]: 对。然后你可以逐步走向单晶XRD，那会有帮助。
Swyx [00:41:29]: 当我们创办 Science Pod 时，科学中的人工智能、数学中的人工智能、物理学中的人工智能，所有这些元素都存在。我们之间有一种争论，哪个更难。我认为你可以从科学角度论证，材料科学是最难的。
Liam Fedus [00:41:43]: 嗯，我们选择它并不一定是因为难度。实际上，在很多领域它反而更容易。对于大类材料，我们有很好的模拟器，而在生物学中，模拟一个细胞、一个器官或一个完整生物体极其困难。所以这是一个巨大的优势。
Swyx [00:42:01]: 是的。这难道不有趣吗？在最微观的层面，它是一维或二维的，但随后它只是多了许多数量级的尺度和结构，而这实际上引入了复杂性。这很奇怪。而材料中的复杂性可能更多是微观结构。
Liam Fedus [00:42:20]: 是的，我还是——对于材料，你在不同的长度尺度上仍然会遇到不同的复杂性。所以是的，微观结构以及更远的尺度。
梦想材料：超导体、磁体和电池
Brandon [00:42:26]: 是的。
Swyx [00:42:27]: 是的。就稍微换个轻松的话题。就科幻而言，假设你可以发明任何你想要的材料。什么更有价值？你知道，我这里有一个清单。显然有室温超导体。但还有，你提到了电池。我一直认为电池。只要比锂离子电池更好，你就很厉害了。这东西已经存在一百年了。最后一个就是碳纳米管，主要用于太空电梯。我不知道，像。就是这三个浮现在脑海中。在你的领域里，人们谈论的梦想材料有哪些吗？
Ekin Doğuş Çubuk [00:43:03]: 绝对是超导体
Brandon [00:43:05]: 是的
Ekin Doğuş Çubuk [00:43:05]: 磁体。
Brandon [00:43:06]: 对。
Ekin Doğuş Çubuk [00:43:06]: Especially if we can lower the dependence on rare earth or transition metals that are hard to source, like cobalt, reducing cobalt in batteries. One of the things I think is very fundamentally important is, if we can get close to Landauer limit for compute energy efficiency.
The Landauer Limit, Reversible Computing, and Energy
Brandon [00:43:24]: Can you, explain Landauer limit?
Ekin Doğuş Çubuk [00:43:26]: Yeah. So if you look at how much energy we spend per, like FLOP or compute, it has followed the Moore’s Law type behavior, but it’s gotten exponentially more efficient. So we’ve been spending exponentially less energy per FLOP or compute. I haven’t checked recently. This was the case ten years ago when I last
Brandon [00:43:44]: Yeah, that’s great. I remember when I first learned about it, looked at that, and we were, what ten, twenty orders of magnitude away from it or something.
Liam Fedus [00:43:50]: Beyond the limit, yes, right.
Brandon [00:43:50]: And the last time I looked, it’s. We’re weirdly close. We’re within a factor of I don’t know three or four orders of magnitude away.
Ekin Doğuş Çubuk [00:43:59]: Yeah.
Brandon [00:43:59]: It’s actually like
Liam Fedus [00:44:00]: I just exponentially rolled out over many years.
Brandon [00:44:01]: Yeah, exponentially over many years are
Ekin Doğuş Çubuk [00:44:03]: Four or five, yeah.
Brandon [00:44:04]: Yeah, pretty impressive.
Swyx [00:44:05]: But in your lifetime, you
Brandon [00:44:06]: Yeah, no I. Yeah, since, just since I discovered this, what, ten or fifteen years ago? Yeah, no.
Ekin Doğuş Çubuk [00:44:11]: And one of the things we have to do is dissipate heat, because as these, computations are happening, it’s producing heat, which has to happen, but then we have to dissipate it. So if we can get close to the Landauer limit for how efficiently we do computation, that’s amazing for humanity, right? Because that means now we’re doing computation, which is probably one of the most fundamental things we do as humanity, but as efficiently as possible from energy perspective, which is a real occurrence in the universe.
Swyx [00:44:34]: That get affected by quantum computing? I’m just going to throw it out there. Theoretically, massively, embarrassingly parallel compute.
Ekin Doğuş Çubuk [00:44:41]: So it definitely gets affected by reversible computing. But honestly, I’m not an expert. I don’t understand. So if you can do reversible computing now, you’re not. You don’t actually have to spend energy to do computation because you’re not actually deleting any information. It’s reversible. But I don’t know if that works. I don’t know. Quantum computing, I assume, has to obey these laws somehow because they are still, like. They have to obey thermodynamic
Brandon [00:45:06]: It’s unitary, so
Ekin Doğuş Çubuk [00:45:07]: Yeah, exactly
Brandon [00:45:07]: It’s reversible, so for now.
Swyx [00:45:10]: One of my most memorable conversations with Elad Gil was he actually was a very skeptical person about quantum computing.
Ekin Doğuş Çubuk [00:45:15]: I see.
Swyx [00:45:15]: He’s like, he’s like, “Even if you had it today, there’s no applications.” I’m like, “Whoa.”
Ekin Doğuş Çubuk [00:45:19]: I
Swyx [00:45:19]: It breaks, it breaks the it breaks the security RSA. That’s it.
Ekin Doğuş Çubuk [00:45:22]: I agree with that. Like for them, people talk about how if you had a quantum computer today, you could simulate things so much better. And I asked them, “Okay, let’s accept that. What would you simulate?” Like, you’re still simulating perfect crystal. You still have the issues that DFT has if DFT was perfect in its prediction ability. So yeah, I do think people are kind of glossing over some things because quantum computing is so exciting. If we can compute in a quantum logic space instead of classical logic, it’s just so exciting that people are glossing over what it would actually do when it’s made. Maybe that’s fine because maybe once it’s made, it will do amazing things we can’t even imagine.
Putting Intelligence Into Every Laboratory Instrument
Swyx [00:45:57]: Good. Okay. I was going to move to the automating the lab side, where you’ve closed the loop on your experimentation. You can talk about robotic arms, talk about. But it’s basically just everything you’ve done here. My favorite quote from you was that every piece of equipment in your lab is going to have a hundred and forty IQ. Okay, what does that mean?
Liam Fedus [00:46:17]: In the early days, as we were scaling things up, we realized that we had huge bottlenecks imposed just by operating machinery. So, for one set of machinery, we would have technicians and scientists looking for particular morpho- morphology and trying to see, okay, what actually were we making, see if this is consistent with our intentions. And we really quickly, as we scaled up the lab, came into these bottlenecks where it just wasn’t keeping up. So we started doing some programmatic approaches to capturing the data on SEM. And the data that was being surfaced from this very simple program where it sort of takes a field of view, captures, zooms in captures, was just not all that useful. It was a little too dumb for what we really wanted to get. And so at that point, it was then, really pertinent to build AI systems directly onto the machines to start controlling these things. And now they have the full context as to what we were trying to achieve. So what was the intent of the experiment? What were we trying to synthesize? What were some other experimental evidence? And it’s actually looking, in the machine and capturing that data. And this is really valuable now because we’ve been talking about these hidden variables not being able to capture everything, but if you can do more intelligent data capture at the time of that experiment, your data for future AI systems and future computational predictions is that much better. You’ll have just a richer set of data. And so that’s sort of what we mean by just, everyth- everything on the lab has to be incredibly intelligent to just make the data as useful as possible. So that was, that was sort of the motivation and inspiration behind that.
Swyx [00:47:57]: What’s the state of the art there in terms of putting intelligence on every device, right?
Liam Fedus [00:48:02]: Well, I think one interesting aspect, is what’s the latency
Swyx [00:48:08]: Exactly
Liam Fedus [00:48:09]: Of controlling that instrument?
Swyx [00:48:10]: Because there’s cloud latency, but then there’s also device compute, which
Liam Fedus [00:48:13]: That’s right
Swyx [00:48:13]: Also.
Liam Fedus [00:48:14]: And then there’s sort of a time scale associated with different physical processes. And if calling out to an API or something is simply too slow, so the amount of reasoning or tokens or tool calls, is just not matched to the latency of that actual process, then that’s infeasible.
Swyx [00:48:33]: Okay. That is right.
Liam Fedus [00:48:34]: It’s almost like, self-driving cars, right? So.
Brandon [00:48:37]: I. Real quick. I just am actually surprised to hear that because the latency I’d imagine for reasoning seems small compared to a lot of these things take hours to run, right? Or maybe, you know. Or maybe I. Maybe your experiments are much faster or high-throughput or something. I’m just surprised to hear that, actually.
Liam Fedus [00:48:52]: So ultimately, the experiments do take, many hours, days, but there could be particular steps where latency would matter.
Ekin Doğuş Çubuk [00:49:00]: Yeah. You might want to have finer control. Another thing is, how expensive it is because the same reason some of these models can be really slow in analyzing data also makes them very expensive. And then finally, the human patience. If a human wants the analysis of a certain XRD pattern and they have to wait two hours before they get a good result, that’s very different, I think, than if they can get it in two minutes.
Brandon [00:49:25]: Oh, okay.
Liam Fedus [00:49:26]: Yeah. And then the reasoning, it’s. You know, you’re not doing a single call to that model, right? You’re doing, potentially many tokens, to kind of get to these patterns.
Ekin Doğuş Çubuk [00:49:35]: Run simulations in the loop.
Liam Fedus [00:49:37]: Exactly.
Ekin Doğuş Çubuk [00:49:38]: Deep research in the
Liam Fedus [00:49:39]: It’s not a single tool call. It’s not a single inference thread. So then the latency can blow up quite a bit.
Brandon [00:49:44]: I see what you mean.
Liam Fedus [00:49:45]: Yeah.
Brandon [00:49:46]: How much of your latency is tool calls? And I assume tool calls is largely DFT or computational- … and therefore. Yeah. So how much of your latency is derived from those versus actual reasoning?.
Liam Fedus [00:49:59]: It’s really process dependent.
Pragmatic Robotics, Automation, and Lab Reliability
Brandon [00:50:01]: Okay. Yeah.
Liam Fedus [00:50:01]: Yeah.
Swyx [00:50:02]: And then the other thing I think about is also, I guess, building up from small things to bigger things, where I assume that the general temptation or the typical development is incremental, where everything is human, operated, and then you find ways in which to automate it, and then you sort of build up from there. I worry that sometimes that is the way that people evolve things, but that’s a local minima, optima.
Ekin Doğuş Çubuk [00:50:25]: Like short-horizon optimization.
Liam Fedus [00:50:26]: Right.
Swyx [00:50:26]: Exactly. When actually you should get a humanoid and just put them in there.
Liam Fedus [00:50:31]: We. I think we’re of the opinion that solving humanoids would actually be slower to kind of getting to some of our goals.
Swyx [00:50:37]: Just checking. Again a lot of this is just like you do this every day. We, like
Liam Fedus [00:50:41]: Yes
Swyx [00:50:41]: See this, but we. I don’t know the reality of the situation. A lot of people having humanoids.
Liam Fedus [00:50:46]: Yeah. I think maybe one process is, by having a mix of humans and automation, you can identify really quickly what are some of the bottlenecks in the experimental process, and you start alleviating those bottlenecks one at a time. And so for example, if there’s some really tricky dexterity task that humans are excellent at but you’d have to spend months automating machine, maybe don’t spend a ton of time there. And maybe a more promising thing is something very routine that takes up a huge amount of scientists’ and technicians’ time, and it’s easy to automate. So I think it’there’s this pragmatism to it. But ultimately what we want from the lab is a huge quantity of data, high-quality data, diverse data, and those are our goals. And full autonomy is a non-goal. It’s sort of in the service that we use automation in the service of achieving the goals on the data.
Brandon [00:51:39]: Yeah.
Liam Fedus [00:51:40]: But also, I think another aspect to automation is, robots will just be. Make fewer mistakes potentially in the lab. And it’s been really helpful having AI systems with a full view over all of our data, because sometimes if there’s a permutation in data, then it can identify it. So for example, one of our steps at one point had, a cyclic error because one of the machines was loaded incorrectly, and the patterns were inconsistent. So the AI was reading through these things and says, “Well, given what was run, this is not expected.” And it’s looking at this basket of data longitudinally, and then it realized, “If I do this cyclic permutation and reverse it, everything is consistent.” And then we were able to go back to the physical infrastructure and understand that a mistake had been made in the loading. And I think managing data quality is so foundational to doing AI in the physical world. And these are the types of things that we’re building in. And so you can improve the operating process, but at that point you’re like, “Okay, this is another great opportunity for automation. How do we make this just so reliable, so durable that we never have those types of mistakes again?”
Ekin Doğuş Çubuk [00:52:51]: Yeah, the reducing the noise floor of experimental data I think is very important. Most experimental data in the literature have such a high noise floor that it’s actually usually worse than DFT accuracy. Because, when different people do experiments, different labs, different parts of the world, different times, it really introduces so much fluctuation to the result. So we are hoping that by standardizing these workflows, one of the biggest benefits will be that the noise floor will be lowered.
Open Models, Proprietary Data, and Compute Efficiency
Swyx [00:53:16]: You’ve been pretty public about how you use open source models and fine-tune them for stuff like this. Is it better. Is it basically. I can imagine a situation where, you have maybe the dumber models do one task, and then you optimize for that task, and then sort of like a generalist frontier model that supervises everything. Is that a good mental model to have? Are there more stages to this that I can think about?
Liam Fedus [00:53:37]: We basically use a mix of open source models and closed source models. You know, we’re huge beneficiaries of this from our simulation side, building up the ML stack, the simulation stack. We don’t need to push on that axis. But there’s a lot of areas where, the latency is too high, the cost is too high. But in some cases we can actually. You know, by having access to this data, we can actually push beyond the frontier of what these systems can do under even some the highest reasoning efforts. And so it’s not necessarily just about cost or Pareto efficiency, but in some cases you can push beyond it when you have access to data that no one else has. One, way to characterize this is in terms of compute efficiencies. So by having access to this data, you can be that much more compute efficient compared to some of the frontier models. And that’s been able. Been a really instrumental thing in our program.
Swyx [00:54:33]: Yeah. I’m so. Data quality or data access as a trade-off of compute I think there’s some amount of exchange rate of dollars for compute, for dollars for data that people. I feel like the pendulum might be swinging now towards data. Obviously you would, you would agree with that, but like.
Liam Fedus [00:54:50]: Yeah, especially like, talking about the noise floor of experimental data. We need to have high-quality data. If you throw a huge amount of compute against a bunch of noise, you’re not going to have a good thing emerge.
Swyx [00:55:00]: And it’s harder for you because you actively try to be sparse. You actually try to get null results and learn from that and
Liam Fedus [00:55:06]: That’s right.
Swyx [00:55:06]: Yeah.
Negative Results and Learning From Failed Experiments
Brandon [00:55:07]: Yeah, you’ve Yeah, you’ve talked about null results in several other venues. What does a null result look like for materials, and how do you use that effectively? Especially when I think your overall signal is probably quite sparse usually in terms of success.
Liam Fedus [00:55:19]: A null result could be we intended to produce some structure, and then all of the evidence points to us not producing that structure.
Brandon [00:55:28]: Do you ever intend to not produce a structure? Of course.
Liam Fedus [00:55:30]: Yeah.
Brandon [00:55:30]: Okay.
Liam Fedus [00:55:31]: Absolutely.
Brandon [00:55:31]: Okay. Cool. You do the negative controls
Liam Fedus [00:55:34]: Absolutely, yes.
Brandon [00:55:34]: Regularly. Okay.
Ekin Doğuş Çubuk [00:55:35]: There are impurity cases that will kill your property. Yeah.
Brandon [00:55:39]: Okay.
Ekin Doğuş Çubuk [00:55:40]: Can be toxic.
Liam Fedus [00:55:41]: Yeah, exactly. 100%.
Brandon [00:55:43]: Okay.
Liam Fedus [00:55:43]: And a negative result, in some cases is actually. You know, we intended to make this, the following thing, and then we’re actually able to identify a new structure that hadn’t previously been identified. So it’s. Negative in some respect. We intended to do something else, but something else emerged from the data.
Ekin Doğuş Çubuk [00:56:02]: Yeah. And also in general when you’re training a machine learning algorithm, especially like at a very basic level, if it’s a classification algorithm, if you don’t have negative samples, you can’t really train if everything is positive.
Brandon [00:56:13]: Yeah. Yeah.
Ekin Doğuş Çubuk [00:56:13]: And this is a particularly bad problem in material science because people usually publish crystals they could synthesize, but they usually don’t publish if they fail to synthesize a crystal. Sometimes they might. And also we never know really if a crystal is not synthesizable ever, right?
Brandon [00:56:30]: It could just be a skill issue.
Ekin Doğuş Çubuk [00:56:31]: Exactly.
Brandon [00:56:31]: Yeah. Yeah.
Ekin Doğuş Çubuk [00:56:31]: It could be a skill issue.
Brandon [00:56:32]: Especially with the literature.
Ekin Doğuş Çubuk [00:56:33]: Synthesis method issue, technology. Yeah. So, I think it really helps us when we do our own experiments and get negative results in the context of what we tried, so then we can even train a classification algorithm.
Liam Fedus [00:56:46]: I think there’s also another valid thing too
Brandon [00:56:48]: Yeah
Liam Fedus [00:56:49]: Of when you have this sort of string of negative results and then finally through process iteration, you’re able to get to that positive result, it’s a really interesting set of I’ll call it process engineering-type data. So it’s like through iteration, how did you actually get to that correct result? Because so much of material science has this ambiguity in actually how something was made. And so, yeah. And so there’s actually. Even if there’s a known material, it can be highly non-trivial to replicate that. Some things are, in like, a high school textbook, but other things are really at the frontier and we’re building up that know-how as well. And then building up a system that, given this string of negative results, how do you actually get to that
Brandon [00:57:35]: Yeah
Liam Fedus [00:57:35]: Positive case?
Brandon [00:57:36]: And going to the classifier result, I would almost assume that if you are doing everything in-house, most of your results will actually be negative rather than positive, which is kind of ironic because the literature only gives you positive results. So if you are sort of cold starting this, it seems like the problem is actually the reverse, that you
Ekin Doğuş Çubuk [00:57:52]: Yeah
Brandon [00:57:52]: Have an abundance of negative results and not enough positive.
Ekin Doğuş Çubuk [00:57:56]: And as you said if we were doing per experiment labels, it would be mostly negative.
Brandon [00:57:59]: Yeah.
Ekin Doğuş Çubuk [00:57:59]: But if we were doing per campaign and assuming campaigns end when we succeed, then it could be a bit more balanced.
Training on the Process of Science, Not Just Its Results
Brandon [00:58:08]: I see. So your reasoning traces could almost be over entire campaigns
Liam Fedus [00:58:12]: Absolutely.
Brandon [00:58:12]: And not. Okay.
Liam Fedus [00:58:13]: This is, this is the key thing. This type of data basically doesn’t exist anywhere else, and we spend so much of our time getting the full lineage of the scientific process into the model. And so tracking all this data, like the conversations, the intuitions, what was executed in the lab, where the computations run, where was the code written, stitching all this together is so valuable. And I think the overall goal is rather than training on the final output of science, you’re training on the process of doing science.
Brandon [00:58:44]: This feels very much like you. What you would try to do with RSI, where you are having a model trained to train better models. You’re now having a model trained to make better experiments. But
Liam Fedus [00:58:56]: Absolutely
Brandon [00:58:56]: The difference is, it’s not going back into the core model for. So it’s
Ekin Doğuş Çubuk [00:59:00]: Unless the physics models you’re developing are improving chips, which then improve AI. So this is the wider
Brandon [00:59:07]: The bigger. Yeah, the higher level RSI.
Liam Fedus [00:59:09]: That’s the big data flywheel. Yeah.
Brandon [00:59:10]: But does that mean that. Is it plausible that in Fable 6 or something, or GPT-8 could just have a. Which has been tuned on these sort of higher order, reasoning traces, could actually just do this without any
Liam Fedus [00:59:30]: Mm
Brandon [00:59:30]: Logic? Because it’s sort of the same thinking process, right?
Liam Fedus [00:59:33]: Yeah. I think there’s, decision-making under uncertainty. Obviously in machine learning, there’s noise, in running the AI loops. But we do think there’s a different set of challenges when you’re actually interfacing with the physical world. But also I think there’s another piece too, which is getting the compression of everything into weights is still really valuable. If inference time reasoning was sufficient, all of the frontier labs would have stopped training at GPT-4 and were like, “Okay, every. From now on out, we’re going to get really good at inference time improvements.” and so we think that by. You know, because of the differences between, physical sciences and machine learning, getting that compressed into our own weights will lead to different types of systems and different types of capabilities.
Ekin Doğuş Çubuk [01:00:20]: But also, we feel like even if Fable 7 gets really good at. Even better than what it is today, it will still have to run experiments to get results. And the reason for it is, right machine learning is really good at what it’s been trained on but scientific discovery is almost by definition what you haven’t been trained on. And that’s why we’re building these labs, so that, whether open models or closed models can use these labs to tinker with the universe, because we don’t feel like you can make a big discovery without trying things.
Liam Fedus [01:00:50]: Yeah, there’s
Brandon [01:00:51]: Yeah
Liam Fedus [01:00:51]: There’s not going to be like. No one’s going to zero shot the room-temperature superconductor.
Ekin Doğuş Çubuk [01:00:54]: Think theoretically
Brandon [01:00:55]: That would be pretty cool. Yeah. As someone who has worked closely with wet labs before, certainly I am more skeptical about zero shotting scientific results than some.
Liam Fedus [01:01:05]: Yeah.
Brandon [01:01:05]: But it is something that I think, some people might ask, so.
Liam Fedus [01:01:08]: Yeah. I think it’s really important to distinguish the results we see in math and theoretical physics from the physical world. Right? Like
Brandon [01:01:18]: Yeah
Liam Fedus [01:01:18]: These are two different things.
Brandon [01:01:19]: Two very different things.
Liam Fedus [01:01:20]: Yes.
Ekin Doğuş Çubuk [01:01:20]: I’m not even theoretical physics. So far it’s been theoretical computer science
Liam Fedus [01:01:22]: Yes
Ekin Doğuş Çubuk [01:01:22]: Coding and math.
Liam Fedus [01:01:23]: Right.
Brandon [01:01:24]: Yeah.
Ekin Doğuş Çubuk [01:01:25]: Maybe theoretical physics is next.
Liam Fedus [01:01:26]: Don’t come. Yeah.
Brandon [01:01:28]: Yeah.
Levels of Scientific Abstraction and the Human Role
Ekin Doğuş Çubuk [01:01:29]: Oh, yeah.
Swyx [01:01:29]: Can I get a sort of mental model of the layers, levels of extraction that you can go? So for example, in terms of automation, and I’m just on this theme again where we talked about the campaign, talked about individual essays and running tests and how they’re. Humans are bad at it, so we should stop humans from doing it. Are there others? So for example, one level higher than campaign could be, a physics theory that you’re testing? Or one level lower than campaign is what? And like, I just. I like to think about it from that point of view and Think about it from like, okay, well, this is an API call now. You never have to touch this again. And this one is. Yeah, no, this is still ninety percent human. And maybe we can sort of draw the map of the territory that way. Is there — are there other levels?
Ekin Doğuş Çubuk [01:02:17]: I can tell you some levels. I feel like you asked a good question. I don’t have a very systematic answer, but let’s talk about some levels. So one level is the atomistic structure. There’s an abstraction, right? There’s no perfect atomistic structure in anything we do, but it’s one approximation. Another one is the continuum model. So this one is not atomistic anymore, but it’s, a continuous mesh representing a material.
Swyx [01:02:39]: Okay.
Ekin Doğuş Çubuk [01:02:39]: And then in a different dimension, there’s the thermodynamics. Assuming, you do this experiment forever, what would be the final state? And then there’s another layer of abstraction, which is kinetics, acknowledging that we’re not doing this experiment forever, so time matters. So how quickly will a reaction happen or not, even if it’s lower energy or not? So thermodynamics, kinetics, atomistic, continuum. What else? Are there other levels of abstraction?
Liam Fedus [01:03:04]: Well, I think your point about okay, there could be an overarching new theoretical advance that informs multiple campaigns right?
Swyx [01:03:14]: It makes a difference. Kind of like as an investor, I want to go here’s the bottleneck, guys. And when we get this, we get everything else. I don’t know.
Ekin Doğuş Çubuk [01:03:22]: To me, from the beginning, our hypothesis has been that one of the big bottlenecks is automated characterization, because it’s not that hard to mix powders to get to try stuff. But if you can’t characterize and analyze it and then decide what the next step should be intelligently, you don’t really benefit much from mixing powders randomly. So we focused a lot on automated characterization, closing the loop with simulations, because that’s how you try a lot of things, make an informed decision, and then decide what to do the next day.
Brandon [01:03:51]: Maybe as a follow-up question, what is how do humans live in this loop? Where at what points do you pull drop the human in? I’m guessing you probably
Swyx [01:04:02]: You could, you could randomly drop in at any part
Ekin Doğuş Çubuk [01:04:04]: Of course.
Swyx [01:04:04]: And add value.
Liam Fedus [01:04:05]: Yes.
Swyx [01:04:05]: But where do you decide to spend your time?
Brandon [01:04:08]: It seems like it might be all levels. You have you have people in the lab who are physically moving materials, and then you also have scientists who are guiding decisions about what campaign, and then you have humans who are. What’s that? I’m just asking a question for you. Yeah.
Liam Fedus [01:04:21]: Yeah. One hundred percent. I think it’s always shifting, too.
Brandon [01:04:23]: Yeah.
Liam Fedus [01:04:24]: So, I think even just using the guidance of campaigns, in the early days, that was fully human-driven. Now, increasingly, it’s a mix of AI-driven and human-driven. And then Doğuş was saying early on we identified that characterization was a big bottleneck. In the early days, it was human-driven, and the scientists were very much overwhelmed, and the balance has shifted significantly towards AI-driven. And this has allowed the scientists who were previously spending all of their time doing these refinements to now elevate their work to something else. But I think we’re. It’s really sort of a function of time. There’it’s always changing at each of these levels.
Search, Optimization, and Design of Experiments
Swyx [01:05:04]: Yeah. In terms of just general search, are there things that perform well in let’s say, the physical world that don’t perform well in the other worlds or vice versa? Like so evolutionary search is pretty popular in LLMs. I don’t imagine it works well here.
Ekin Doğuş Çubuk [01:05:21]: So especially in simulations, people have been using evolutionary search, for a long time. It’s pretty good at structure prediction, for example, finding low energy structures.
Swyx [01:05:31]: Okay, so it works. Yeah.
Ekin Doğuş Çubuk [01:05:33]: And then process engineers in semiconductor industry will do DOE, design of experiments, and they’ll often use some kind of zeroth order whether it’s Bayesian optimization or evolutionary search or. I don’t think they use RL, but similar idea, yeah.
Swyx [01:05:47]: You have to have some kind of algorithm
Ekin Doğuş Çubuk [01:05:48]: Yeah
Swyx [01:05:49]: To just figure it out.
Ekin Doğuş Çubuk [01:05:50]: And all zeroth order optimization behaves the same anyway, right? So.
Scaling Labs and Reusing Data for Reinforcement Learning
Brandon [01:05:54]: Yeah. Okay, I want to go back to the question I had earlier, which was one of scaling. And so I-I’m curious, what is your vision for scaling the lab? So there’s different ways. Coming from the biology world there’s a lot of cool tricks you can do in biology. You can tag things, you can use DNA sequencing for all sorts of interesting readouts. And if you can map your complicated assay onto sequencing, you can now scale to millions or hundreds of millions or something. Can you do tricks like this? Ultimately, every assay has its own runtime. I’ll, I’I have a favorite blog post I would want to. I’ll quote in the show notes. But what does the runtimes look like for these labs? Is it. Is your scaling just for mater-materials just fundamentally linear in terms of you do you just need more resources to do more things, or can you paralyze things in clever ways and combine things, and are there tricks there?
Liam Fedus [01:06:52]: Well, I think one way to think about it is, the RL environment is not quite trivially run the experiment, wait for that agent to fully roll out, doing the calculations, going through all the instruments to the final outcome. We basically start producing data across all the different instruments and all these different campaigns. And one way you can begin to expand that for machine learning is now reinforcement learning environments can be constructed based on again dates for the experimental campaigns or computational campaigns at that point. Or you can also, start taking, subsets of instruments, and you’re like, “Okay, we’re going to create reinforcement learning environments on the basis of this instrument alone.” Or maybe it’s, across these different instruments to kind of get to some reward state. So that’s a way where. This finite basket of data, when viewed in different ways, can then be expanded, for training purposes. So it doesn’t get at the actual experiment run, but from an ML perspective.
Ekin Doğuş Çubuk [01:07:55]: Maybe I can give you some analogs for the bio examples you gave. So, one thing people have tried is combinatorial sputtering. I don’t know if you’ve ever heard of this. You have these sputtering targets, and you create this compositional gradient. So when you look at the final product, because there is gradient from each precursor, it creates different crystals at different spatial locations. So in one go, you maybe try hundreds, thousands of crystals. That’s an that’s similar, right, to your example.
Liam Fedus [01:08:23]: Oh, yeah, that makes sense.
Ekin Doğuş Çubuk [01:08:24]: And as far as I know, this hasn’t worked very well so far because it turns out there is diffusivity, so things move around. Another one that we thought about, but we haven’t done yet, but a lot of people talk about this is superconductivity measurements are a bottleneck because unlike XRD, there’s no high-throughput superconductivity measurement, and it can take one hour per measurement. So people have thought about taking the different candidates, mixing them all into one sample, and putting it in. And if, it has superconductivity, you know that one of the sources that. And then you can do a bit like how people are doing COVID testing to. So there are ideas like this exist. Some of them work well, some of them don’t. Yeah.
Custom Hardware, New Labs, and Scaling Ambition
Brandon [01:09:00]: So you, I think, are about to announce a big fundraise. How are you thinking about, scaling?
Liam Fedus [01:09:06]: We’ve been doing our design of our labs for a while, and so I think the resources allow us to continue to scale up the labs significantly, but also the compute, both from the AI side and the computational side. But I think really importantly too, it’s what are the new types of labs we’re able to build? So I think we’ve been proving out this loop, in our initial labs here in Menlo Park, but we’ll be continuing to expand to new types of labs and repeat the same process.
Ekin Doğuş Çubuk [01:09:33]: Yeah, like the kind of lab we’re trying to build, I think haven’t been built before at this scale or at this approach. So we have learned a lot from our own build-ups so far, and those lessons guide us for the next version and the next version. And then we can, as you said increase the scale, the ambition, the quality of instruments. Some instruments are very expensive. As we get a really good understand for what we get a lot of return from we can invest into those more.
Liam Fedus [01:09:58]: And it makes. Hardware engineering has become so core to it. So for example, in the early days, just for speed, we’d buy some off-the-shelf instruments. And as we kind of push the scale of the lab
Brandon [01:10:08]: You make your own.
Liam Fedus [01:10:09]: We have to make our own. So we realize, okay, for this instrument, it’s pretty fast. These components are actually pretty quick, but this weighing machine actually becomes the bottleneck. Or there can be things pertinent to data quality where, well, the resting position of this robotic arm actually is above the plate of previously mixed things, so there’s a contamination risk. So in our hardware design, we’re going to redesign it so that the resting position lies away from the-those things. These are the kind of subtle details that allow us to kind of push the noise floor down for our experimental campaign. So, basically take these learnings and scale it up.
Assembling the Team Behind Synthesis Superintelligence
Brandon [01:10:45]: You ship your org chart. You have many teams. I think the. We were surprised when we looked at your jobs page, and we were like, “Okay, we don’t think we fully map what Periodic does.”
Liam Fedus [01:10:56]: Maybe some organizing principles, so to kind of make sense of the job page. So we hire for AI research and infrastructure, computational, experimental roles, and then hardware engineering roles, and then product roles.
Ekin Doğuş Çubuk [01:11:10]: We want to achieve synthesis superintelligence, right? And we feel like for that, we need the right chemistry expertise, the right physics expertise, simulation theory, and thin films, powder. So hardware engineering these seem like very diverse roles, which they are, but they’re all actually, coherent in what they’re trying to do, which is a synthesis for intelligence. And this also applies to the LLM researchers, the infra, even the product engineers, because the product engineers kind of make sure all the research gets made into a product that the experimentals in the lab can use.
Liam Fedus [01:11:42]: Yeah. So in order to do the end-to-end loop with the physical world, it really requires your ability to have agency over the physical world to build these labs, to run the campaigns effectively, to create AI systems against that, to make them good users of the computational tools. That’s like the difficulty, but also the opportunity of Periodic. It’s like this group of people has just never been brought together before. It’s irreducibly a multidisciplinary problem.
Ekin Doğuş Çubuk [01:12:07]: Yeah. And one other guiding principle for us that is really important for us is we want the people who are very good at what they do and experienced to do it hands-on. You know the modern life has gotten us to this place where, especially in academia, when someone is really good at research in some area, we give them so much responsibility for grant writing, teaching, that they stop having the time to do research in that area hands-on. But if you look back at Bell Labs, IBM, institutions that made really good progress, it was really experienced people doing hands-on work. Like Bardeen was in the lab every day, even though he’s a theorist. Alex Müller was in the lab doing experiments, even though he was the lab lead. So we try to do that here too. So we have some of the world’s leaders in different fields, but they’re doing hands-on work.
Brandon [01:12:52]: Do you want to brag or just call them out? You know.
Ekin Doğuş Çubuk [01:12:54]: Oh, I would love to yeah. So, when I was doing my PhD, my favorite computational material scientist who’s kind of from my age group is Muratan Aykol. He did his PhD with one of our advisors, Christopher Wolverton. And he is kind of like our computational material science expert. He’s incredible. He’s sometimes, a really good experimentalist, even though he does simulations. He understands systems really well. Our lab lead is Joe Checkelsky, who is a professor at MIT, but he’s on leave to work with us full-time. And Joe is this incredible physicist, understands superconductivity really well, but he also was a professor in Japan for a bit, so learned synthesis and chemistry really well, especially for a physicist. We have Daniel Chica, who was a graduate student in Mercury Canasides’ group, which is maybe the best solid-state chemistry group in the world. And Daniel was one of his synthesis experts, so he’s like a magician with his fingers. There’s so many examples like this.
Liam Fedus [01:13:50]: No, I know. I think Dima Bahdanau, so he’s been leading a lot of our AI and LLM efforts. He was the inventor of neural network attention.
Brandon [01:13:59]: Oh, Bahdanau. The Bahdanau attention.
Liam Fedus [01:14:01]: Exactly. Yeah.
Brandon [01:14:03]: Yeah.
Liam Fedus [01:14:03]: Yeah. Yeah, so very hands-on in sort of getting into the traces, the training distribution, how to connect it to the lab.
Ekin Doğuş Çubuk [01:14:10]: The chemistry.
Brandon [01:14:11]: Chemistry.
Ekin Doğuş Çubuk [01:14:11]: He’ll go deep.
Liam Fedus [01:14:12]: He was very deep. He actually has solid-state synthesis books at his lab in Montreal. Ray Nakano, so he used to tech lead some of the operator work at OpenAI. He’s in the lab. So when we were doing some lab tours, we were initially, we’re like, “Oh, this new, scientist, this new technician really looks like Ray.” And no, it was literally this was Ray actually in the lab, with the scientists automating the machines. And it’s like this is sort of the DNA and the type of people we need to kind of pull this off.
Forward-Deployed Engineering for Semiconductor Partners
Swyx [01:14:43]: So, we also saw that you have forward-deployed engineering, roles. We happen to be spinning up a forward-deployed engineering podcast because there’s so many engineers that would do that role. They may not necessarily know that there even is a role for them at Periodic. What is it, and how do they work with customers?
Liam Fedus [01:15:01]: So we’ve been building our own products, our own tools, our own AI and computational
Swyx [01:15:06]: For yourself. Yeah.
Liam Fedus [01:15:06]: For ourselves. So we’ve been customer zero. Now what we’re doing is we’re taking these same tools to different industries, and we’we have a huge focus right now in the semiconductor industry. However, in order to do this these are very, private companies. We need to be able to operate within some of the most secure environments. And so our Periodic forward-deployed engineers and researchers will actually be on site working and actually integrating our AI systems, our computational systems to help our partners get to their goal states more quickly. So basically the tools, the know-how that we’ve built in doing our own materials discovery, materials engineering, we’re now bringing to industry. And we think this is the most effective way to get partners to their end states and to their goals, rather than just throwing some technology over the wall and telling them to figure it out. Then these forward-deployed engineers will also do the inference locally, but then also can, train on the data. So again, once we take our system that understands these different areas and we deploy, we can make it expert on the customers or the partners’ data so that they can own their own intelligence and have systems that understand this much more effectively than just hitting some API or some untrained model. And so this is a much expanded, forward-deployment engineering role than typical because this really requires some very deep, machine learning expertise. You have to be incredibly precise from a data training perspective, infrastructure perspective, but also a physics perspective. So we’re working with highly technical industries. They have to understand, different areas of chemistry, material science, some of the devices. And so that, those are the roles we’re building out right now.
Swyx [01:16:53]: Part of it is willingness to spend extended periods on site in Taiwan. It’s like, okay, well, I know that kind of customer.
Swyx [01:17:01]: It’s interesting that you’re monetizing. That is the primary way that you’re monetizing now, but it might change in the future. But like, that’s like a model that I think people don’t really get about APL Research Lab. It’s. And when you compare yourself to Bell Labs, that’s what people are thinking, right? Which is.
Commercializing Scientific Discovery
Liam Fedus [01:17:19]: I think a really great analog would be software engineering.
Swyx [01:17:22]: Right.
Liam Fedus [01:17:22]: So in the early days of software engineering we had GitHub Copilot, then we had early versions of ChatGPT. People were using these things as copilots to help them get to their solutions. And as the automation improved, now we have things like Codex, and very few of our engineers are writing code the way they used to. And we think a very similar thing could play out for Periodic as well, where you can have systems to accelerate the researchers, the material scientists, the materials engineers, the process engineers. But as the autonomy, intelligence, and capability grows, you can begin to price outcomes. So help me get to this kind of goal state, and I think that’s, a really interesting area for us.
Ekin Doğuş Çubuk [01:18:03]: Yeah. We kind of feel like our best contribution to solid-state physics and science could be if we made these tools and topics profitable, similar to how ChatGPT made CS and LLM majors way more popular in colleges before and after. We’d love to show that solid-state physics, material science research can make a big impact commercially, and then that will attract more attention. Young people will want to study physics, which will be, a dream. And Bell Labs did make a huge commercial impact, right? So they failed to commercialize some of their incredible advances. They, of course, did some research that just couldn’t be commercialized like the cosmic, background. But they did benefit a lot from the vacuum tube connecting, East Coast to West Coast by phone lines.
Liam Fedus [01:18:49]: Yeah. And I think it’s really also interesting to think that technology and capital are incredibly intertwined. So if you look at what was the progress on chatbots for many years imagine recounting, okay, what was the progress from chatbots from I’ll say 2010 to 2015? We’d be kind of at a loss to say over that five-year period, how much did they improve? Whereas if you look at the period from, 2021 to 2026, it’s night and day. And what happened was ChatGPT and these other systems were able to achieve a product market fit, and it changed the capital landscape entirely. This changes the landscape for hiring, compute, data, et cetera.
Swyx [01:19:31]: The virtuous cycle.
Liam Fedus [01:19:32]: Exactly. And so
Swyx [01:19:32]: Like success breeds success.
Liam Fedus [01:19:33]: Exactly. So technology, it’s incredibly coupled to the capital and those resources, and we want to achieve the same thing in the physical world.
Open-Source Contributions and Academic Research
Swyx [01:19:41]: Yeah.
Brandon [01:19:42]: So one thing I’ve been seeing is there’s this move from science being funded from government grants and so on to VCs and private funding. Do you all plan on making anything you’re doing, open source, like either releasing data sets or actual models? Is this something in the future even. You know, there are many models which may be not well, your state of the art, but which could still be you know, useful releases.
Ekin Doğuş Çubuk [01:20:07]: Absolutely. So, we have a couple of things. So we have been contributing to open source to kind of like PyTorch and the Materials Project, code base, Custodian, their DFT runner. TorchSim was actually created by one of the researchers here, Abhijit Ganggan, and he maintains it still. JAX-MD, Abhijit maintains it. So we do a lot of open source contributions, including LLMs.
Liam Fedus [01:20:31]: XLANG, Megatron.
Ekin Doğuş Çubuk [01:20:32]: Yeah.
Liam Fedus [01:20:32]: Yeah. So we’ve been very active contributors back into open source.
Ekin Doğuş Çubuk [01:20:36]: And we also have an academic grant program where we give academic gift grants to academic groups in universities that we feel like are really advancing this direction towards synthesis superintelligence. So that’s also been great to see. I think the first paper from this funding is about to come out, so it’ll be exciting to just get more papers come out like this. Yeah.
The Path to Better Superconductors
Brandon [01:20:57]: Okay. Let’s steelman this, say that you’ve achieved 100% lab automation, which does everything, characterizes everything correctly. You have all of the fun agents which can do everything. I still am somewhat unclear about how you actually get to superconductivity. There’s a lot of hard problems to solve. What is the path there when there is no sort of theory for model for most of the strongly correlated systems or high-temperature systems?
Ekin Doğuş Çubuk [01:21:23]: Yeah. So, our view is that it’s not hard to think of chemical spaces that would host superconductivity. What’s really hard is to synthesize them. That’s why we’re really emphasizing synthesis superintelligence. So maybe one example to give historically is Alex Müller, when he thought that transition metal oxides might be a good path to superconductivity, he was actually trying nickelates, the nickel, oxygen, and some, cations, and that didn’t work. And then he tried cuprates, which is copper, oxygen, and some cations, and that worked, and then he got a Nobel Prize. But then we know that later on turns out nickelates was a good idea. Nickel and copper are next to each other on the periodic table.
Brandon [01:21:59]: But only as two layers, right?
Ekin Doğuş Çubuk [01:22:00]: Yeah. Like there are only so many 3D transition metals.
Brandon [01:22:03]: Yeah.
Ekin Doğuş Çubuk [01:22:03]: And then, Harold Hwang, one of our, Stanford professors here who’s incredible, he realized he can make nickelates in thin film form and show the superconductivity. So, if we have a really good synthesis superintelligence in the lab and we can really scale up these things, I think we won’t run out of ideas, or the LLM won’t run out of ideas about what directions to try, like 3D transition metals, oxygen. There are very related ideas here. Many people have been trying cobaltates, which is like instead of copper or nickel, you try cobalt. So I think that’s super exciting. Maybe one historical example to also study is the Japanese group that discovered magnesium diboride. As you know, that’s the highest temperature, ambient pressure, conventional superconductor, and they discovered it by just trying a bunch of materials. I think they tried 30,000 different things. Thirty of them seemed to host interesting superconductivity, and MgB₂ was one of them. But magnesium diboride sat on people’s shelves as a precursor for decades before then, and BCS theory had been invented back in 1957. So the way people found these materials wasn’t projected from theory. It wasn’t because they couldn’t synthesize until then. It was just like they tried a bunch of things. So now imagine if we had this perfect automation that you’re describing, which sounds like a dream, and we tried the 30,000 the Taketo group tried in over his career in a month. We just really increased the surface area for luck.
Closing: Increasing the Surface Area for Discovery
Swyx [01:23:30]: No, but yeah, it’s very inspiring what you’ve done. Congrats on all your success. I feel, I feel like you’re creating, yeah, the modern playground. I imagine recruiting must be super easy for you, so I’m just a little bit jealous, but you’ve worked hard for it to get here, so yeah.
Ekin Doğuş Çubuk [01:23:44]: Yeah.
Swyx [01:23:44]: Yeah. Well, thank you so much. Yeah, it was great chatting with you both.
Ekin Doğuş Çubuk [01:23:46]: Yeah, super fun. Yeah.
Swyx [01:23:47]: Yeah. Thank you.
来源：Latent Space · latent.space