My thought would be that "choosing to accelerate [more/less]" and "choosing to brake [more/less]" are strategic-level decisions; and, if you have a self-driving

My thought would be that "choosing to accelerate [more/less]" and "choosing to brake [more/less]" are strategic-level decisions; and, if you have a self-driving car, the tactical-level fine input of how much more/less can be handled for you. Kind of like what's going on with mouse-"controlled" movement in this AI gameplay here: https://www.youtube.com/watch?v=qYluZRwrw9w https://www.youtube.com/watch?v=qYluZRwrw9w Or, to put that another way: the car already has a (low-confidence) idea that there's an obstacle in the path, and so it already has a "possible plan" to brake hard to stop in time to not collide with the obstacle. You just have to choose that plan — declare your strategy — by pulling back on the yoke. It then executes on the plan.