Researcher: most research output is terrible — and that's exactly why LLMs work
RexDouglass · x · 2026-09-23
Rex Douglass argues that academics overestimate the authenticity of research artifacts because almost no one does metascience work — most research output is derivative, cargo-cult, vague, and rushed. That, he says, is a "great baseline." His follow-up makes the deeper point: this is precisely why LLMs work at all. Human text is embarrassingly predictable; everyone lightly rewording others' sentences to avoid "plagiarism" produces many ngrams but few ideas, making language statistically learnable.
More from AGI Musings
- Rishi Bommasani co-founds nonprofit STABL with SayashK and Arvind Narayanan — RishiBommasani · 2026-11-12
- Investor: AI diffusion will be slowest in consumer software — Oura, Strava, Labcorp still have no real AI strategy — MartinGTobias · 2026-11-12
- On AI Progress: When Models Lag on Your Load-Bearing Skill, You Can't See Them Improving — dfrsrchtwts · 2026-11-12
- Slack's David Hoang: agents beat people on coordination, proposes 'boundary agents' — davidhoang · 2026-11-12
- CS student mourns coding flow states he'll never experience in the AI era — chrissyykat · 2026-11-12
- Ford's 370% Turnover Led to $5 Days — But ASI Never Quits — krishnan · 2026-11-12