CNRS Mathematician: 'I Trust LLM Audits More Than Human Ones' and Will Stop Refereeing
IgorCarron · x · 2026-11-19
Michael Goldman, Research Director at CNRS, argues in a Proofs and Prompts essay that mathematical journals have lost their purpose — his personal answer to "Do we still need mathematical journals?" is no.
- He now trusts LLM audits more than human ones, including his own, and will no longer spend hours tediously checking papers, especially for journals that deny referees LLM assistance
- He contends AI has exposed the only remaining function of journals: producing a metric for hiring and promotions
- The piece reviews journals' five classical functions (dissemination, registration, validation...) and earlier reform efforts like Gowers' 2011 Cost of Knowledge
- Part of a series on AI-era publishing, authorship, and disclosure; the author disclosed using AI for brainstorming and literature exploration
Steve Hsu shared the essay commenting the trend is clear.
Related event: CNRS Mathematician Says He Trusts LLM Review Over Human Referees(2 posts)→
More from AGI Musings
- Read More Facts and Fewer Theories: Why Real Experts Memorize Specifics — panickssery · 2026-11-19
- Dario Amodei calls open-source model path dangerous; critics slam it as a control grab — adamamcbride · 2026-11-19
- Benchtop synthesis device releases have 3x'd; half of active manufacturers are in China — JacquesThibs · 2026-11-19
- iamtrask: humans will keep paying a premium for human entertainers, Mozart-style — iamtrask · 2026-11-19
- Researcher: Real scientific progress isn't rewarded, expect decades of incrementalism — gerardsans · 2026-11-19
- Kevin Roose: ~50 people made nearly all key AI decisions over the past decade — trevposts · 2026-11-19