The Doomsday argument in anthropic decision theory — AI Alignment Forum