AI-generated, AI-verified proof no human can read sparks unfalsifiable review debate
cloneofsimo · x · 2026-11-09
cloneofsimo highlights an unprecedented situation around a math proof (defended by Elliot Glazer, who says the repo's accumulated evidence goes far beyond "trust me bro"): one side demands verification of its correctness, the other demands finding its flaws — but no human understands the proof, since it was fully AI-generated and AI-verified, making either task impossible without AI.
He likens it to "discussing a movie without watching it" and invites correction if experts have since understood it. Glazer pushes back on vague critics, insisting they address the accumulated evidence.
More from AGI Musings
- OpenAI co-founder Greg Brockman: frontier models with test-time compute are becoming the engine of scientific discovery — gdb · 2026-11-09
- Long-term memory vs. search is like L1 cache vs. RAM, argues zetalyrae — zetalyrae · 2026-11-09
- AI is eroding traditional moats: years of advantage now replicated in days — DeryaTR_ · 2026-11-09
- Epoch AI: 2027 chip supply could run 30–170M frontier AI agents, or 1.9B cheaper ones — Dr_Singularity · 2026-11-09
- White paper: moving beyond the AI compute arms race as big tech faces historic spending losses — ReportProfessional57 · 2026-11-09
- When machines do everything else, relationships are what's left for humans — ZeroStateReflex · 2026-11-09