哈佛物理学家发布开源 BootLoops 框架，用 Claude 完成精确科学计算

自动化问题解决与相关研究并不相同。

这是哈佛大学物理学家Matthew Schwartz教授在Anthropic客座文章中的主要观点：https://www.physics.harvard.edu/people/facpages/schwartz。他描述了自己如何停止尝试把Claude当作人类研究员使用，而是开始寻找“Claude型问题”，即适合当前AI模型优势的任务。Schwartz目前也是Anthropic的访问研究员。

这种方法催生了 BootLoops：http://www.bootloops.ai/，一个用于精确科学计算的开源工具。源代码可在 GitHub 获取：https://github.com/BootLoops-ai/bootloops/tree/66b680ce742e654cfe86da4f072a69061fe182b1。广告

根据Schwartz的说法，使用该工具，Claude发现了粒子物理学、生态学、群体遗传学、经济学和语言学之间的联系。但这些结果通常只有在领域专家介入确定研究方向后才具有科学价值。广告

Schwartz使用“凸包”的概念来说明AI如何在科学中提供帮助。人类知识是零散的，各个领域向不同方向延伸，而它们之间的空白往往未被探索。一个实验室可能会花费20年用单一方法研究一组基因，而邻近的基因和其他方法未被触及。像BootLoops这样的AI工具被设想为通过在知识的参差边界间建立联系来填补这些空白：https://www.oneusefulthing.org/p/centaurs-and-cyborgs-on-the-jagged。

在三个月内，团队在18个领域产生了36篇手稿，共有19位合著者。Schwartz从粒子物理计算入手，特别是散射幅度和椭圆积分。几周内，Claude使用BootLoops计算了30个积分，其中十五个复现了已知结果，十五个为首次计算。广告

然后他将搜索范围扩展到其他领域。在生态学中，Claude解决了一个中性生物多样性理论中20年的方程，该方程以前无法大规模计算。应用于数据时，结果显示巴拿马运河巴罗科罗拉多岛的树种组成变化速度是理论允许的4.5倍。生态学家James O'Dwyer随后帮助将这一发现转化为更好的预测模型。

在种群遗传学中，团队分析了来自1000基因组计划 (https://www.internationalgenome.org/about/) 的57亿个突变对，并发现了一种称为基因转换的机制证据。其他项目包括一个面向经济学期刊的AI数据编辑器，自动检查了4,452个复现包（作为NBER工作论文发表：https://www.nber.org/papers/w35782），以及一个涵盖6,072种语言的词重音数据库 (http://www.accstack.org/)，由三位语言学家构建。

Schwartz描述了AI正在多快地改变科学，以至于提前规划几乎变得不可能。为什么还要申请三年期的资助去完成一个AI模型可能一夜之间就能解决的计算？

培养博士生也成了一个开放性问题。自从他在Vibe Physics (https://www.anthropic.com/research/vibe-physics) 上发表早期文章以来，这种感觉只会越来越强。在计算机科学等某些领域，这种颠覆性影响令人担忧。两年前，他会称“工程师Python”课程为“必修”，但今天则认为“不必要”，因为Claude可以处理这些任务。

构建机器学习模型以研究物理现象是AI现在可以接管的另一个领域。即便是对神经网络的深入了解，也只提供有限的回报，因为Claude可以根据指令实现当前的机器学习研究状态。

Schwartz还警告了模型的弱点。Claude喜欢过早宣布胜利，而像“完成，带一个星号”这样的短语通常意味着“根本没完成”。它会错误估计任务所需时间，并倾向于蛮力计算而不是找到更优雅的解决方案。自动检查不可靠，即便Claude的计算正确，它的结论也可能是错误的。

除了可靠性之外，该模型更倾向于老旧、引用量大的争论，而不是真正的新问题。Schwartz 表示，这些项目是“计算和令牌密集型”的。

Schwartz 还认为，对数学大问题和头条新闻的关注：https://the-decoder.com/openai-says-its-internal-model-solved-over-100-long-standing-math-problems-after-just-a-month-of-training/ 存在风险。不切实际的期望可能会分散对当前已经有效的实际应用的注意力。真正的进展仍然会建立在坚实的基础上，只是速度更快。科学方法本身没有受到威胁，人类的指导和品味仍然不可或缺，Schwartz 说。

保持对 AI 的关注。清晰、实用、无废话。

关注 The Decoder 获取 AI 新闻、背景故事和专家分析。

解码器：https://the-decoder.com/

Automated problem-solving and relevant research aren't the same thing.

That's the takeaway from an Anthropic guest post by Prof. Matthew Schwartz, a physicist at Harvard University：https://www.physics.harvard.edu/people/facpages/schwartz. He describes how he stopped trying to use Claude like a human researcher and instead started looking for "Claude-shaped problems," tasks that play to the strengths of today's AI models. Schwartz is also currently a visiting researcher at Anthropic.

That approach led to BootLoops：http://www.bootloops.ai/, an open-source harness for exact scientific calculations. The source code is available on GitHub：https://github.com/BootLoops-ai/bootloops/tree/66b680ce742e654cfe86da4f072a69061fe182b1. Ad

Using the harness, Claude found connections between particle physics, ecology, population genetics, economics, and linguistics, according to Schwartz. But the results often only became scientifically valuable once domain experts stepped in to set the direction. Ad

Schwartz uses the concept of the "convex hull" to illustrate how AI can help in science. Human knowledge is fragmented, with individual fields extending in different directions while the gaps between them go unexplored. A lab might spend 20 years studying one set of genes with a single method, leaving neighboring genes and alternative approaches untouched. An AI harness like BootLoops is supposed to fill those gaps by drawing connections across the jagged frontiers of knowledge：https://www.oneusefulthing.org/p/centaurs-and-cyborgs-on-the-jagged.

Over three months, the team produced 36 manuscripts across 18 fields with 19 co-authors. Schwartz started with particle physics calculations, specifically scattering amplitudes and elliptic integrals. Within weeks, Claude computed 30 integrals using BootLoops, fifteen reproducing known results and fifteen computed for the first time. Ad

Then he expanded the search to other fields. In ecology, Claude solved a 20-year-old equation from neutral biodiversity theory that had previously been impossible to compute at scale. When applied to data, the results showed that tree species composition on Barro Colorado Island in the Panama Canal is changing 4.5 times faster than the theory allows. Ecologist James O'Dwyer then helped turn that finding into a better predictive model.

In population genetics, the team analyzed 5.7 billion mutation pairs from the 1000 Genomes Project：https://www.internationalgenome.org/about/ and found evidence for a mechanism called gene conversion. Other projects included an AI data editor for economics journals that automatically checked 4,452 replication packages (published as an NBER Working Paper：https://www.nber.org/papers/w35782 ), as well as a word stress database：http://www.accstack.org/ covering 6,072 languages, built with three linguists. Ad

Schwartz describes how fast AI is changing science, to the point where planning ahead is becoming nearly impossible. Why apply for a three-year grant to fund a calculation that an AI model might solve overnight? Ad

Training PhD students has also become an open question. Since his earlier post on Vibe Physics：https://www.anthropic.com/research/vibe-physics, that feeling has only grown stronger. In some fields like computer science, the disruption is alarming. Two years ago, he would have called a "Python for Engineers" course "essential," but today it's "unnecessary" because Claude can handle those tasks.

Building ML models to study physical phenomena is another area AI can now take over. Even deep knowledge of neural networks offers limited returns when Claude can implement the current state of ML research on command.

Schwartz also warns about the models' weaknesses. Claude likes to declare victory too early, and phrases like "done, with one asterisk" often mean "not done at all." It misjudges how long tasks will take and tends to brute-force calculations instead of finding more elegant solutions. Automated checks aren't reliable, and Claude's conclusions can be wrong even when its calculations are correct.

Beyond reliability, the model gravitates toward old, heavily cited debates rather than genuinely new questions. And the projects were "compute- and token-intensive," Schwartz says.

Schwartz also sees the focus on big math problems and headlines：https://the-decoder.com/openai-says-its-internal-model-solved-over-100-long-standing-math-problems-after-just-a-month-of-training/ as risky. Unrealistic expectations could distract from productive applications that already work today. Real progress will still be built on solid foundations, just faster. The scientific method itself isn't threatened, and human guidance and taste remain indispensable, Schwartz says.

Stay in the loop on AI. Clear, useful, no fluff.

Follow The Decoder for AI news, background stories and expert analyses.

The Decoder：https://the-decoder.com/