John_Maxwell comments on The problem/​solution matrix: Calculating the probability of AI safety “on the back of an envelope”