AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Superintelligence with Human Values Tied to It? Maybe apophatic?
Stuart Russell believes that it is important to make sure that an AI is always uncertain about what human values are, in order to prevent it from becoming too closely aligned with those values./nThere are many problems with this approach, at a high level, as the AI will always try to approximate human values perfectly.