Dagon comments on Best arguments against worrying about AI risk?