Geoffrey Miller comments on Safety-concerned EAs should prioritize AI governance over alignment