Economist Paul Novosad: if AI can't find real flaws, your paper is probably fine
paulnovosad · x · 2026-10-29
Paul Novosad shares a useful heuristic: ask an AI for the three biggest errors in a paper. Since AI is poor at judging error severity, it often labels trivial nitpicks — like interval 40–100 vs 56–95 — as 'major.' If the best it can come up with is 'this paper depends on assumptions,' the paper might actually be decent.
Related event: Economists Push Back on AI Audit of Saez-Zucman Wealth Tax Paper(5 posts)→
More from AGI Musings
- 「望远镜发明让天文学家失业了吗」：AI 冲击下人类智识退路的争论 — zetalyrae · 2026-10-29
- KOL doubles down: a billion people have no idea what AI will do to jobs — A_K_Nain · 2026-10-29
- Gabriel: AI's Unbounded Creative Possibilities Are Painful — This Alpha Will Never Happen Again — gabriel1 · 2026-10-29
- Open models with Astra-level cyber skills: defenders-first access debate splits security researchers — rez0__ · 2026-10-29
- zetalyrae essay argues the post-singularity future may be a lonely place — zetalyrae · 2026-10-29
- "AI could give humanity its childhood back" sentiment gains traction — AIandDesign · 2026-10-29