Alan Chan is skeptical of "super persuasion" as a viable AI takeover path, arguing that humans can choose to ignore AIs