Summary: From the assumption of the existence of AIs that can pass the Strong Form of the Turing Test, we can provide a recipe for provably aligned/friendly superintelligence based on large organizations of human-equivalent AIs
Is anyone in the world going to deploy the strong form of the turing test? Is this even something possible to do outside theory?
It just seems like there's no solution to sandbagging capabilities or concealing desires
this is not a post about solutions
Is anyone in the world going to deploy the strong form of the turing test? Is this even something possible to do outside theory?
It just seems like there's no solution to sandbagging capabilities or concealing desires
this is not a post about solutions