Gary Marcus:Riemann 证明结果离不开 Lean 与人类数学家协作
Gary Marcus 拆解了 Claude 参与 Riemann 证明的真实流程:Lean 负责验证证明片段,数学家 Alpoge 提供判断,不是一句话让模型独立解题。想搞懂这类成果到底怎么达成的,这篇值得细看,别被"纯 LLM 证出来"的说法带偏。
Gary Marcus 在 X 上质疑外界对 Claude 参与 Riemann 相关证明工作的解读,指出一份报告提到该工作采用 agentic loop:把证明片段输入 Lean 验证后迭代修正。这意味着 Lean 形式化验证工具是完成证明的认知系统组成部分。
他强调数学家 Alpoge 也贡献了数学判断,并非某个 LLM 收到"solve riemann"提示词后独立得出答案。Marcus 还指出 Claude Code 本身已是神经符号混合系统(neurosymbolic hybrid),推测 Claude 亦然。
这条评论的核心是把功劳从"纯 LLM"重新分配到"LLM + 形式化验证 + 人类数学家"的组合上,提示读者在评估模型数学能力时要先问清 methodology。
why the methods in the Riemann result matter, and why most people are probably misenvisioning what happened. Gary Marcus @GaryMarcus As noted elsewhere I would really like to know what the methodology was, but e.g. one report sggests that agentic loops in which bits of proof were fed through lean and then iterated on, in which case lean was an essential part of the cognitive system that did the proof. so too was Alpoge, who also add his mathematical judgement so it’s not like a pure LLM came out with the answer from a prompt “solve riemann” in the way that some people are erroneously envisioning. of course i doubt that Claude is a pure LLM at this point; we know for a fact that Claude code is a neurosymbolic hybrid (what I long favored) and i imagine Claude is. 🔗 View Quoted Tweet📊 2 ⚡
本条情报归属 AITOP「行业」核心栏目,重点聚焦北美前沿实验室与硅谷科技巨头。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:Claude、Lean、Gary Marcus、神经符号,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「Gary Marcus」(@GaryMarcus),经 AITOP 多维评分管线结构化提炼。