OpenAI's Pachocki: better math models shelved because compute goes to RSI

basedjensen · x · 2026-09-11

basedjensen argues that many AI advancements are now essentially bound by whether labs deem it worth throwing compute at a problem. He cites OpenAI chief scientist Jakub Pachocki from 40 days ago: 'we believe we could make the models better at specifically mathematics research with additional focus, but we do not prioritize this direction because of the urgency we feel about RSI.' The point: model progress in specific domains is a budgetary allocation issue, not a capability ceiling.

Original post →

More from AGI Musings

AGI Musings channel →