[AIpol]
a number of people have expressed a very strong distaste for building systems that automatically generate exploits. I think there's a bit of nuance here that isn't immediately obvious
if you are going to build _any_ kind of LLM-driven (more generally, fuzzer-driven) system that […]
