# @repligate — 2024-09-19

♥182 ↻9 · https://x.com/repligate/status/1836907275585474655

Seems like O1 is good at math/coding/etc because they spent some effort teaching it to simulate legit cognitive work in those domains. But they didn't teach it how to do cognitive work in general. The chains of thought currently make it worse at most other things. https://t.co/91bZ2Omiwg https://t.co/UMoKRjxpex

tags: author:repligate, kind:tweet, model:o1, on:o1, year:2024
cited on: _dossiers/o1.md, o1