from Hacker News

kcorbitt

joined 3/6/13, 5:30 PM has 6252 karma

• Show HN: RULER – Easily apply RL to any agent

  by kcorbitt on 7/11/25, 5:47 PM, with 11 comments

• Everything I know about reward hacking

  by kcorbitt on 6/12/25, 5:14 PM, with 0 comments

• Show HN: ART – a new open-source RL framework for training agents

  by kcorbitt on 4/30/25, 3:35 PM, with 12 comments

• ART·E: how we built an email research agent that beats o3

  by kcorbitt on 4/29/25, 5:38 PM, with 2 comments

• Using GRPO to Beat o1, o3-mini and R1 at “Temporal Clue”

  by kcorbitt on 3/6/25, 7:51 PM, with 55 comments

• Analyzing OpenAI's Reinforcement Fine-Tuning: Less Data, Better Results

  by kcorbitt on 12/30/24, 10:25 PM, with 0 comments

• Using reinforcement learning and $4.80 of GPU time to find the best HN post

  by kcorbitt on 10/28/24, 5:17 PM, with 95 comments

• Show HN: Agent.exe, a cross-platform app to let 3.5 Sonnet control your machine

  by kcorbitt on 10/23/24, 4:44 PM, with 232 comments

• DPO fine-tuning outperforms SFT

  by kcorbitt on 10/2/24, 5:46 PM, with 0 comments