A list of core AI safety problems and how I hope to solve them — AI Alignment Forum