Y Hacker News new | ask | show | jobs

user: Danau5tin created: 2020-06-18 karma: 107 www.danaustin.ai submissions: 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments 0 points | 0 comments Show HN: I RL-trained an agent that trains models with RL (for ~$1.3k) 107 points | 49 comments 0 points | 0 comments Scaling Coding-Agent RL to 32x H100s. 160% Improvement on Stanford's TBench 2 points | 1 comments Show HN: Multi-Agent-Coder Is #12 on Stanford's TBench. Beats Claude Code 5 points | 1 comments 0 points | 0 comments My weekend project accidentally beat Claude Code – #12 on Stanford's TBench 2 points | 2 comments 0 points | 0 comments Show HN: Terminal-Bench-RL: Training long-horizon terminal agents with RL 125 points | 12 comments