Having this kind of anxiety about what can go wrong is extremely useful. It is what drives one to develop systems and metrics for monitoring the outcomes your machine is producing and motivates those who manage well to constantly taste-test the outputs of the system and to look for problems in its nooks and crannies.
