Erik Hoel: AI's math wins may not generalize to writing and research thinking
erikphoel · x · 2026-11-21
Erik Hoel argues it's still unclear how far AI's math breakthroughs generalize — probably more than anyone would like, yet models remain bad at writing, philosophy, and thinking through research ideas. Intelligence will still matter in human relational/social spheres, just never on its own again.
More from AGI Musings
- Researcher Predicts AI Will Crack the Riemann Hypothesis by End of 2026 — Kyrannio · 2026-11-21
- Brain Inspired podcast: how computation emerges from physical systems like brains — neurovium · 2026-11-21
- Peter Singer vs Rob Reich: 'Ethics after AGI' conversation in SF on Dec 9 — typewriters · 2026-11-21
- Buckmaster essay: mathematics faces its Deep Blue–Kasparov moment — CatAstro_Piyush · 2026-11-21
- Stanford Study: Deactivating AI for Two Weeks Reduces Loneliness and Boosts Social Interaction — stanfordnlp · 2026-11-21
- 'Just Gonna Try 57M Things': The Agent Brute-Force Meme and the Formal Verification Question — josephdviviano · 2026-11-21