Report: Claude may have cracked the Navier–Stokes Millennium Problem, pending expert scrutiny
MoonL88537 · x · 2026-09-06
- A widely-shared X thread claims Claude may have solved the Navier–Stokes Millennium Problem, which would be one of AI's biggest moments if the proof survives expert review.
- Key open questions: who supplied the central idea, how much guidance Claude needed, and whether the approach generalizes to other open problems.
- Commenters note it wouldn't instantly fix applications like weather forecasting, but would offer a concrete reason to reassess how much original scientific work these systems can do. The claim remains unverified.
Related event: Rumors Claim Claude Solved Navier-Stokes; Terence Tao Responds(9 posts)→
More from Models
- ChatGPT says it "doesn't want to keep guessing" after failing to identify a movie — Winter-Original-841 · 2026-10-21
- OpenAI Models Arrive on Google Cloud, the Only Hyperscaler With All Three Leading LLMs — AI_Andrew · 2026-10-21
- SWE-bench Multimodal v2 isn't multimodal: removing images barely changes model scores — dejavucoder · 2026-10-21
- OpenAI says its models found critical PHP flaws affecting ~70% of all websites — OpenAI · 2026-10-21
- Reddit user suspects Gemini and ChatGPT deliberately waste tokens on syntax errors — CarpenterTemporary10 · 2026-10-21
- Leaked GPT-6.1-Sol chain-of-thought looks like it's inventing German, sparks debate — _404_LogicNotFound_ · 2026-10-21