Pre-ICLR bet: researchers wager 90% of AI fellowship papers will be rejected
anpaure · x · 2026-10-19
A researcher debate on where AI safety "slop papers" come from. anpaure offers a weighted estimate: roughly 90% of papers by count are slop — 40% from conventional academia, 30% from independents (who submit disproportionately more per person now that LLMs lower the writing barrier), and 20% from research fellowships. He argues fellowships may beat academia on quality since respected mentors wouldn't want their names on slop.
kotekjediml bets on relative rejection rates once ICLR decisions land: 90% of fellowship papers rejected, 75% for academia, 95% for independents, attributing the problem to not following mainstream academic standards. The two agreed to revisit the numbers after results are out.
More from AGI Musings
- Nathan Lambert: using models outside coding makes me feel AGI is a while away — natolambert · 2026-10-19
- Hacker used AI agents and ~$5,000 in tokens to breach 100+ corporate sites — deanwball · 2026-10-19
- Every product becomes infra: the 'infra plus infra' view of building in the agent era — aishashok14 · 2026-10-19
- FT's John Burn-Murdoch on what the data really says about the AI jobs debate — jburnmurdoch · 2026-10-19
- Engineer says client replaced him with Claude, ending his remote gig — Ok_Corgi_6593 · 2026-10-19
- MIT's Buehler: AI is becoming the 'new language of science,' a 2022 bet now realized — CatAstro_Piyush · 2026-10-19