OpenAI's Navier-Stokes proof isn't what everyone thought: a different method OpenAI itself didn't grasp
量子位 · wechat · 2026-10-15
A Scientific American investigation reveals a twist in OpenAI's claimed Navier-Stokes breakthrough: mathematicians studying the 166-page proof found its forcing construction differs entirely from the presumed 'Spanish route' by Córdoba and Martínez-Zoroa — and OpenAI itself didn't understand what method its model used when announcing it. An early small agent-team result may have followed the Spanish route, but the 10,000-agent system took a different path. The proof is Lean-formalized, yet formal validity isn't human understanding; experts still use other LLMs to parse it. The episode validates Tao's warning that AI now produces proofs faster than humans can digest them — OpenAI has since published 722 manuscripts covering 372 results.
More from Models
- OpenAI blocks manual reset with 13% usage left: 'no reset needed right now' — AccomplishedSugar490 · 2026-10-23
- Meta's 816B Muse Spark Open scores 39, losing to 180B Qwen3.8-Flash-Next at 40 — BTA_Labs · 2026-10-23
- Gemini refuses to generate image in Docs, blames document size limit — viksit · 2026-10-23
- Dev jokes his true calling is telling everyone to try the latest DeepSeek — intellectronica · 2026-10-23
- SubQ Snap 1.1 ties for first on Artificial Analysis Decision Index with 384K context — ArtificialAnlys · 2026-10-23
- PolyAI unveils Dialog-RSN-1: audio-native voice model answering in under 300ms — matthen2 · 2026-10-23