ML researchers debate AI-written papers: correct proofs, unreadable prose
kfountou · x · 2026-09-13
Csaba Szepesvari complains AI-assisted academic writing is nearly unreadable and says he's sitting on cool results rather than littering the world with crappy papers. kfountou responds that ML conferences reward what is proved, not how — with proofs relegated to appendices — so poorly written appendices are inevitable; he prioritizes verifying correctness first, since a correct translation of a proof means a correct result.
Related event: RL Scholar Withholds Long Protests Over Poor AI Writing Quality(3 posts)→
More from Research
- Ji, Lei, Zrnic revisit surrogate outcomes in the age of AI in Biometrika — lihua_lei_stat · 2026-09-13
- Multimodal Agentic Frameworks Survey Hits arXiv With Open-Source Tracking Repo — MikeShou1 · 2026-09-13
- Complexity theorists: P vs NP out of reach for AI, but L/NP and BPP/NEXP may fall soon — _onionesque · 2026-09-13
- SEED-UMI uses matching human-robot exoskeletons for reliable dexterous imitation learning — jeasinema · 2026-09-13
- xeophon commits to supporting independent evals with fully open-source tooling — xeophon · 2026-09-13
- Training away its own bad reasoning lifts Qwen3-4B math scores by 7.5 points — imjustnewatai · 2026-09-13