Alarms with incessant false positives are inherently dangerous. Sure, there's some threshold of false positives, under which we should still expect people to investigate all alarms. But above that threshold, how can we continue to blame the people involved? The hardware is at fault.