Hacker News new | ask | show | jobs
by m00x 391 days ago
Many are speculating it was trained by o1/o3 for some of the initial reasoning.