The process of reasoning in AI involves trillions of motor control actions, even though they are not planned in detail. Progress in AI involves removing assumptions which make problems easier to solve, such as the assumption of complete observability of the situation.