That may well be so, I’m not saying we should refactor these systems. I’m in no position to recommend such things.

My point is that these are the kind of systems we should be talking about, when discussing the safety of our civilization… Not some imagined “AI threat” that everyone can bikeshed about.