Tao on AI math repo at ~42% formalized: AI less helpful for fuzzy math tasks than hoped
iskander · x · 2026-10-08
danintheory's team updated their GitHub math repo with 6 new Lean formalizations, 19 modifications, and 3 withdrawals; roughly 42% of top-line results are now formalized, with ongoing updates and errata.
Terence Tao commented that one withdrawn manuscript, Algebraicity of Kuga–Satake Correspondences for K3 Surfaces, was a question he likes with seemingly exciting ideas. He notes that wrong human-authored arguments often still contain salvageable intuitions, but it's unclear whether that's possible here — and his experience is that AI is less helpful for this kind of fuzzy task than one might hope.
More from Models
- Cyber Index Alliance launches with IBM, NVIDIA and Vercel as founding members — ArtificialAnlys · 2026-10-09
- GPT-6 Sol Daybreak Blue tops Artificial Analysis Cyber Index at $1.77 vs Grok's $11.67 per task — ArtificialAnlys · 2026-10-09
- LightOnOCR-3 draws buzz: "insane" says early watcher — IgorCarron · 2026-10-09
- Unverified screenshot shows "Gemini 4 Argon" appearing in Google Antigravity CLI — gaganghotra_ · 2026-10-09
- Unverified: OpenAI delayed GPT-6.1 Astra's October release over authorization and reporting concerns — johnkwaters · 2026-10-09
- Dev reports GLM-5.3 tool calls appear completely broken in Cursor CLI — DanielLockyer · 2026-10-09