[nostrebored]

what types of problems are you seeing this for? i'm wondering if it's just a difference in domains, but this is not our experience at all. llms will write something, but the odds that it is efficient for any problem with deep complexity has been basically zero.

if it really were fuzzing and finding different candidate spaces, then I'd expect it to be good at things like dynamic programming, where I've only seen it fail. usually i see it get stuck in a bad solution and just thrashes around in that minima. these are problems where we can construct a verifiable test space, and it will eventually wind up with a solution, but one that is thousands of lines long and uses no structure of the problem space