No, the AI writes far less secure code than I do to start, even with the SotA models and careful prompting/detailed plans.

You’ve moved the goalposts so far that you’re now talking about a different game altogether.