Look, I’m not an AI hater, but AI is… not great at multi-threading code. And having it analyse multi-threaded code proves nothing because… it’s not good at multi-threaded code. This isn’t entirely shocking because I’m not good at it either and need to write in some very particular ways to have even a hope of being correct. But basically, unless it was written by a genuine expert, I wouldn’t want to even glance at this PR. And it wasn’t.
I don't think AI is particularly worse at multi-threading code than humans are: humans are notoriously bad at it. I've had reasonably good success with telling Claude and Gemini to go into a codebase, insert clang -Wthread-safety annotations, fix any issues it finds, and refactor code that isn't amenable to the annotations to make it possible.