语言模型能推导广义相对论却无法发明它？
Could a language model derive General Relativity yet still be unable to invent i…
Could a language model derive General Relativity yet still be unable to invent it?
语言模型能否推导出广义相对论，却仍然无法发明它？
This Google DeepMind paper argues yes by separating discovery into induction, deduction, and abduction.
这篇Google DeepMind论文认为可以，通过将科学发现分为归纳、演绎和溯因来论证。
Induction extracts rules from observations; deduction derives consequences once axioms are supplied.
归纳从观察中提取规则；演绎在给定公理后推导出结论。
The paper argues that induction and deduction still leave a missing operation in scientific discovery.
论文认为，归纳和演绎仍然遗漏了科学发现中的一个关键操作。
The missing operation is the abductive jump from experience to a new explanatory premise.
这个缺失的操作是从经验到新的解释性前提的溯因跳跃。
Einstein is the case study because Newtonian gravity offered almost no empirical error signal: inertial and gravitational mass agreed to 10⁻⁹, while Mercury’s perihelion anomaly was patched with the hypothetical planet Vulcan rather than treated as a reason to rebuild spacetime.
爱因斯坦是案例研究，因为牛顿引力几乎没有提供经验误差信号：惯性质量和引力质量一致到10⁻⁹，而水星近日点异常被用假设的行星“火神星”来修补，而不是作为重建时空的理由。
A compression-driven system would therefore have little gradient toward General Relativity.
因此，一个由压缩驱动的系统几乎没有朝向广义相对论的梯度。
Einstein developed General Relativity despite the data strongly favouring the old theory. His motivation came largely from conceptual conflicts and thought experiments, not from simply fitting a better model to a large set of observations.
尽管数据强烈支持旧理论，爱因斯坦还是发展了广义相对论。他的动机主要来自概念冲突和思想实验，而不是简单地将更好的模型拟合到大量观测数据上。
The paper’s proposed direction is action-controllable world models that let agents intervene in physically consistent simulations and translate simulated experience into candidate axioms.
论文提出的方向是行动可控的世界模型，让智能体在物理一致的模拟中进行干预，并将模拟经验转化为候选公理。
This is still a position paper, not an experiment proving that language models cannot perform abduction.
这仍然是一篇立场论文，而不是证明语言模型无法进行溯因推理的实验。
tomzahavy. com/files/llms-cant-jump.pdf
tomzahavy.com/files/llms-cant-jump.pdf
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