MIRI/OP exchange about decision theory — AI Alignment Forum