Arms races are bad things. First best by far is if nobody has the doomsday devices, but second best is if we attempt nonproliferation of doomsday devices.
As a parallel, we would have been still at risk in a world where DeepMind was working on building ASI but where Elon didn’t freak out and start a competitor (followed by another competitor), but not as much risk. That’s not because DeepMind are “the good guys”, it’s because of race dynamics.
If only one entity is building AI, that reduces the risk from race dynamics, but increases the risk that the entity will become world dictator. I think the former reduction in risk is smaller than the second risk. So to me first best is nobody has AI, second best is everyone has it, and the worst option is if one group monopolizes it.
Arms races are bad things. First best by far is if nobody has the doomsday devices, but second best is if we attempt nonproliferation of doomsday devices.
As a parallel, we would have been still at risk in a world where DeepMind was working on building ASI but where Elon didn’t freak out and start a competitor (followed by another competitor), but not as much risk. That’s not because DeepMind are “the good guys”, it’s because of race dynamics.
If only one entity is building AI, that reduces the risk from race dynamics, but increases the risk that the entity will become world dictator. I think the former reduction in risk is smaller than the second risk. So to me first best is nobody has AI, second best is everyone has it, and the worst option is if one group monopolizes it.