That seems like a reasonable take, I also see that being radical anti-AI doesn't really work in practice. That's probably true for most domains btw.