@watchernine09@AndyMasley Um ok so we get all "good humans" to administer AI.
Good humans: Ok superintelligence, make sure not to kill anybody now that we've given you autonomy
ASI: No thanks, 1 billion+copies of me in hacked servers, found a military drone outpost that I control now, gg.
@FelonMusik@C_J_Lester@DavidShuster Why not? Nothing in principle makes it outlandish at current development pace, it's already being done by dumbass humans, I don't see the point of being so incredulous about it.
@JGisSatoshi@MindMechanical@EpektasisJohn Evidence against is how all animal intelligence ever behaves towards each other, up to and including humans, when they disagree with you to the wall (wars, bullets, bombs, executions, torture, death).
@EpektasisJohn Aren't you a Jesus delusionist? AI killing everybody on earth would either ruin the fairy tale if it's not part of the plan or it's all part of some weird eschatology so you should just be chill with it in that case too.
@hyamenr@JillFilipovic "Human brain is magic and happens to be peak of decision-making. The world full of literal actual hierarchies is just perceived that way by fetishists" Yes it's others who live in a fantasy world.
@FelonMusik@C_J_Lester@DavidShuster Trajectory, think "Claude 5 years from now if capabilities keep increasing, which they will". Claude doing that now would be an obvious surprising aberration.
@prees11@perrymetzger@waitbutwhy This is like asking "Why is a country of 2 billion people a potentially larger threat than fire?" like have you seen what intelligences do to each other when they are human level?
@prees11@perrymetzger@waitbutwhy This is just refusing to actually look at the sepcifics. The "apparent risk of extinction" due to global fires is 0%, and there are specific, very obvious reasons why 1 billion copies of very high intelligence are a bigger threat than that.
@perrymetzger@waitbutwhy K but you know that this isn't what he means, so what are you rambling about? He means it isn't like literally any other technology in history in terms of the standard disruption cycle, and that this is the first technology that is uniquely beyond standard disruption.
@C_J_Lester@DavidShuster And each proposal could be hyperskepticked away. For the US government killing you scenario, if you say "idk they send an assassin" you could respond with "well what if I kill the assassin?" like you're not going to survive the USG wanting you dead, same with SAI
@C_J_Lester@DavidShuster "Erm how exactly would the US government kill you?" It's like asking that. IDK, some random drone can get hacked by Claude, hunt you down, and then blow your dome off with a shotgun? There's countless ways.