Curious how it handles the non-technical rejections. A lot of kernel patches get turned down for reasons that have nothing to do with the code itself.

53% catch rate on bugs that already cleared human review, with the main bottleneck being false positives... Sounds about right for where LLMs are useful: they don't get tired on patch 47 of a series

Previous discussion: https://news.ycombinator.com/item?id=47427647