OpenAI denies user data influenced Millennium Prize solution; researcher says "I just don't believe them"
QuintinPope5 · x · 2026-09-15
Mathematician Tristan Buckmaster publicly questioned whether OpenAI's landmark Millennium Prize problem solution was influenced by his user data from the past two months. OpenAI says it's "categorically impossible."
Buckmaster's response: "I just don't believe them."
The exchange has sparked debate, with some defending OpenAI and others treating it as a data-contamination red flag—highlighting how hard it is for labs to prove their results are isolated from user data.
More from Fun
- Beff Jezos mocks EAs: "ready to ruin the world for crumbs of status in their Berkeley polycule" — beffjezos · 2026-09-15
- The entire Renaissance table setting is edible — and the internet can't get over it — Substantial_Depth744 · 2026-09-15
- User sarcastically salutes Claude after another botched task — ccerrato147 · 2026-09-15
- GPT 6 Astra is halfway through Fallout 3 after 12 hours of play — imjustnewatai · 2026-09-15
- Self-proclaimed OpenAI "GPT-6 Astra" halfway through beating Fallout 3 after 12 hours — imjustnewatai · 2026-09-15
- Intel: AI doomer Coxon has been pitching himself to news outlets and podcasts on doom — RileyRalmuto · 2026-09-15