Comment by reassess_blind

I really doubt that. If these models continue to improve, they will be producing higher quality code than, maybe not the best human developers, but certainly the average human developer who would have typically been writing code on these systems.

Pre-2024 code is full of human-written vulnerabilities.