Claude-linked GRH proof claim remains unvalidated after Stanford rival draft
cnBeta reports that Anthropic mathematician Levent Alpöge claimed he and Claude found a proof of the generalized Riemann hypothesis, publishing a paper, supporting code, and a Lean formalization split into 14 data packages. Hours later, Stanford University researchers released a competing GRH proof draft made without AI, but both efforts remain unvalidated, and multiple people reportedly found possible problems in the human-authored draft. cnBeta cautions that neither result can yet be treated as a final solution to the Riemann hypothesis.
The reports present a notable attempt to apply Claude to a major open problem in mathematics, but they also show why publication, Lean code, and a competing draft do not establish mathematical acceptance without independent review.
据科技媒体The Decoder报道,围绕数学界最著名的未解难题之一——黎曼猜想,一场人工智能与人类数学家的“证明竞赛”正在上演。Anthropic数学家莱文特·阿尔佩格(Levent Alpöge)宣布,他与Claude合作找到了广义黎曼猜想的证明。仅数小时后,斯坦福大学研究人员也公开了一份完全由人类独立完成的竞争性证明草稿。 不过,两项成果目前均有待数学界进一步验证。

阿尔佩格于当地时间周五宣布,他在Claude的协助下完成了相关研究,并在个人网站公开论文及配套计算机代码。他表示,研究成果不仅涉及1859年提出的黎曼猜想,还涵盖范围更广的广义黎曼猜想(GRH)。
据报道,阿尔佩格过去数月一直与Claude合作研究这一难题。起初,他负责确定研究方向,并在简化数学模型上整合不同方法。但在进入实际问题后,Claude在部分关键推导中取得了超出其本人理解进度的突破。
#Anthropic#Claude#Levent Alpöge#Stanford University#Ravi Vakil#Alessandro Masullo#Brian Conrad#Lean