Ranking three years of math discoveries: 87% AI-made, 59% released by OpenAI today
willdepue · x · 2026-10-07
willdepue asked GPT 6 Pro and Fable 5.1 to rank all math discoveries from the last three years, tagging human vs AI contributions:
- 13% human (e.g. geometric Langlands, 2024), 87% AI-discovered, mostly from OpenAI's math repo
- 59% of the list was released "today," matching OpenAI's dump of results from an internal frontier model
- Items include Lean-formalized finite-time blowup for 3D Navier-Stokes, a rational Hodge conjecture variant, a quasi-Riemann hypothesis, Unique Games, and undecidability of Hilbert's tenth problem over Q
- "81% of them have been released today. wtf"
More from Models
- 'Compression of the RH proof has begun': AI circles buzz over unverified Riemann claim — teortaxesTex · 2026-11-22
- GLM 5.3 and DeepSeek V4.1 Flash resold at 60% off: the economics of open-source giveaways — lxfater · 2026-11-22
- Baseline model test: no persona, no forced formatting, no safety bypasses observed — Sauers_ · 2026-11-22
- 2-day Gemini stress test: $4.99 Google AI Pro nearly maxed, orchestrator UI revealed — lipeng0820 · 2026-11-22
- Timed eval task asks users to deliberately make Claude fail at real work in 12 minutes — Icy-Measurement10 · 2026-11-22
- Tesla FSD 15's real-time view isn't 3D rendering — it's the AI model's own synthetic perception — mark_k · 2026-11-22