Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

How does this compare to something like Cursor Cloud Agents with a solid set of skills and tools?


Similar but reusing lab-native CLIs like Claude Code or Codex, which they perform RL on. And so in the long-run, we believe this approach wins over custom harnesses.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: