The claim that we can easily turn off a superintelligent AI is naive—a superintelligent system would have considered this possibility and likely taken steps to prevent it, making the argument that we need not worry because 'we can just switch it off' analogous to claiming we can easily beat Deep Blue at chess by making the right moves.