Hacker News new | ask | show | jobs
by valine 498 days ago
The RL is done on problems with verifiable answers. I’m not sure how o1 slop would be at all useful in that respect.