SciMuse in PNAS: 100 Max Planck leaders grade 4,500 AI research ideas
MarioKrenn6240 · x · 2026-11-12
A Max Planck team spearheaded by Xuemei Gu published SciMuse in PNAS: LLMs combined with a knowledge graph of 58 million papers generate personalized research ideas and suggest unexpected collaborations.
- Evaluation: 100 research group leaders from 50 Max Planck Institutes, spanning physics, chemistry, biology, CS and humanities, assessed 4,500 AI-generated ideas
- The authors claim this is by far the largest interdisciplinary human evaluation of AI-generated research ideas involving senior scientists
- Early findings tease that high impact appears anti-correlated with something unexpected; full results in the paper
More from AGI Musings
- Cells rejuvenated to under 20: the case for Longevity Escape Velocity — Dr_Singularity · 2026-11-12
- Tom Goodwin on AI panic: the internet promised radical change too, and arrived painfully slowly — DanGrover · 2026-11-12
- FOOM Debate Revisited: Panel of AI Timelines Forecasters Hosted 15 Years Later — Jsevillamol · 2026-11-12
- In 2026 an obviously Claude-written cover letter makes recruiters doubt your whole CV — Afinetheorem · 2026-11-12
- Dev complains security features are gated behind expensive enterprise plans — AccBalanced · 2026-11-12
- The AI trap: doing unneeded work faster—prioritize before you scale — jiayuan_jy · 2026-11-12