Why AI writing won't beat humans: writing is anti-inductive and shared models wear out
herbiebradley · x · 2026-09-25
herbiebradley argues "AI will be better than humans at writing" won't come true while everyone uses the same model:
- Writing is an anti-inductive domain: as we collectively absorb a new model's "better" style, it wears on us and stops being a quality marker—just as if all AI wrote like V.S. Naipaul
- Even as AI wins competitions judged by unfamiliar literati, heavy readers of AI text still consider it slop
- Root cause: on-policy RL seeks the mode of the distribution and neglects the tails. His challenge to skeptics: come back when you've solved in-weight continual learning.
More from AGI Musings
- Should leading AI labs face radical transparency at the cost of trade secrets? — dhadfieldmenell · 2026-09-25
- The New Yorker: There is no AI "race" — first to invent rarely wins — jjding99 · 2026-09-25
- Same AI tools are making every app look identical, just like algorithmic content — vhanagwal · 2026-09-25
- AI Benchmarks Now Test Real Lab Science Skills, Not Just QA — 141_1337 · 2026-09-25
- Study: top AI experts badly underestimated how fast the field is moving — The Decoder · 2026-09-25
- Researcher Warns AI Is Breeding Nihilism Among Scientists; Jack Clark Responds — doomslide · 2026-09-25