Overview
- Levent Alpöge, an Anthropic mathematician, says he and the company’s AI model Claude produced a proof that he completed on Sunday, November 15, and that he published as a Lean formalization after Anthropic’s record IPO on November 20–21.
- Initial technical checks by independent developers report the Lean project compiles and is accepted by multiple Lean kernels, but reviewers caution it is not yet settled whether the formal statements exactly match the classical Riemann hypothesis or whether the full argument is correct.
- Separately, Alessandro Masullo released a shorter, human‑authored draft of a claimed proof that is under expert review, including scrutiny by Stanford mathematician Brian Conrad.
- The Clay Mathematics Institute continues to list the problem as unsolved and leading specialists say full verification will take substantial time and careful human scrutiny in addition to machine checks.
- A Lean formalization proves logical steps inside a chosen encoding, so experts must verify definitions and scope before acceptance; if validated this could change how deep mathematics is produced but will first require clear, human‑readable explanations and community consensus.