Anthropic decision theory for self-locating beliefs — AI Alignment Forum