Stuart_Armstrong comments on A toy model of the control problem