Mathematicians slam Meta's AI math announcement: ellipsoid-fitting problem already had established results
qberthet · x · 2026-10-03
Meta AI announced a collaboration where mathematicians, working with Muse Spark, proved a sharp cutoff for exactly fitting random points onto a stretched ellipsoid — framing it as exploring whether AI can help on genuinely open problems with no existing solution path.
Physicist Florent Krzakala pushed back hard, calling the framing outrageous, especially amid ongoing tensions over AI and scientific credit (following the recent Buckmaster/OpenAI dispute):
- The ellipsoid-fitting problem had an established research programme with substantial rigorous progress, not an open problem without a solution path;
- In 2023, Antoine Maillard and Dmitriy Kunisky gave a statistical-physics derivation of the conjectured transition at n/d² = 1/4 using the replica method;
- Publicly available rigorous proofs of the exact threshold existed by the time of Meta's announcement.
Krzzakala argues the promotional language blurs the history of the research problem and is misleading and insulting to prior contributors.
Related event: Meta's Claim of AI Solving Six Open Math Problems Disputed by Scholars(4 posts)→
More from Companies & People
- Palo Alto Networks buys OpenAI-alumni startup Applied Compute for $4.5B — KateClarkTweets · 2026-11-19
- FT: US tech outages blamed on rush to fix bugs found by AI — nordicinst · 2026-11-19
- Stanford professor: AI takes students from problem to working prototype in 80 minutes — HariSubramonyam · 2026-11-19
- marclou launches $100K hackathon: build a profitable startup in 30 days — marclou · 2026-11-19
- Activists to rally outside Anthropic HQ demanding a ban on recursive self-improvement — DavidSKrueger · 2026-11-19
- Reported: Musk offers NVIDIA engineers a path to SpaceX's Terafab if Huang exits — inductionheads · 2026-11-19