Mathematicians clash over posting AI-derived results humans don't yet understand

littmath · x · 2026-09-16

Mathematician Jabaluck pushes back on the growing norm that mathematicians shouldn't post results they don't understand (e.g., AI-generated ones). He argues that norm arose when verifiability was scarce; machine-verifiable important results are a contribution to human understanding and should be shareable, even if arXiv stays reserved for human-comprehensible work.

Related event: Mathematicians Debate Publishing, Credit and Priority in the Age of AI Proofs(4 posts)→

Original post →

More from AGI Musings

AGI Musings channel →