Slightly offtopic, but…

This should be pretty trival to show if all these AI assisted researchers are finding human coded bugs or recent-ish bugs contributed by AI…

surely that would be solid proof that LLMs are somewhat useful