Hacker News new | ask | show | jobs
Show HN: I pre-registered the rules,then measured AI agents shipping wrong fixes (github.com)
2 points by kolesnikov-arch 26 days ago
A controlled comparison of ungated vs gated AI patching on SWE-bench Lite. Full methodology, results, and reproducible scripts in the repo. Breakthrough for flash models.
1 comments
kolesnikov-arch 26 days ago
Author here.
link