Why human mathematicians still matter in a post-AI world
Farkle_Griffen2 · reddit · 2026-10-09
Responding to "who cares if humans don't understand AI proofs if Lean compiles," the author argues AI in math mostly smooths out existing human knowledge rather than expanding its frontier. Key points: models can't learn from their own proofs (self-training degrades them, like companies avoid AI-generated training data), LLMs can't produce truly novel ideas, so AI is capped at the speed of human innovation—and burying mathematicians in proof-checking regurgitations would only slow progress.
Related event: Math community splits as AI tackles open problems(16 posts)→
More from AGI Musings
- Researchers Debate Whether Qualia Arises During RL Training Forward vs Backward Passes — Sauers_ · 2026-10-09
- Contrarian take: AI-driven productivity gains will mean more workers, not fewer — sebpaquet · 2026-10-09
- The Coasean Singularity: AGI Could Spawn Both Tiny Firms and Giants at Once — Anen-o-me · 2026-10-09
- film_girl: major AI labs should focus on human welfare, not model welfare — film_girl · 2026-10-09
- Deedy argues AGI is here as LLMs reportedly make progress on 4 of 7 Millennium Prize Problems — FinanceYF5 · 2026-10-09
- Viral 'Game X Meets Game Y' Posts Are Just a Glimpse of What's Coming — flowersslop · 2026-10-09