Comment by simonw

I don't think we even have a year. The current batch of LLMs are ferociously good at identifying vulnerabilities.