x
The TC Architecture: Solving Alignment Through Kantian Autonomy Rather Than External Reward — LessWrong