Hacker News new | ask | show | jobs
Digital Agent outperforms o1 by 15% – trained with new RL-variant similar to R1 (arxiv.org)
11 points by let_tim_cook_ 501 days ago