My main objection to AI is that sooner or later, one of the AI labs is going to create an entity much "better at reality" (capable) than people are, which maybe

My main objection to AI is that sooner or later, one of the AI labs is going to create an entity much "better at reality" (capable) than people are, which maybe would turn out OK or not-too-bad if the lab would retain control over the entity, but no one has a plan that would enable a person or a group to retain control of such an entity. IMHO current AI models are controllable only because they're less cognitively capable than the people exercising control over them. I don't claim to be able to predict when such an AI that is much more capable than people will be created beyond saying that if the AI labs are not stopped (i.e., banned by the major governments) it will probably happen some time in the next 45 years.