Yale Psychologist Puts Claude Opus On Peer Review Duty, Catches Errors Human Reviewers Miss

lu_sichu · x · 2026-09-18

Yale psychology professor Paul Bloom says he's been having Claude Opus review published papers and his own unpublished manuscripts (no confidential uploads). The reviews are excellent: the model catches statistical errors, internal contradictions, and conceptual confusions human reviewers miss, offers useful suggestions, has good judgment on what's genuinely novel, and avoids human "myside bias" — no reject-the-paper-because-it-didn't-cite-me, no insults. Reposter lusichu goes further: most reviewers aren't real specialists anyway, and he won't read papers that haven't been stress-tested by top models. Try it on the classics — the newest models can improve even very good papers.

Original post →

More from AGI Musings

AGI Musings channel →