Engineers make many assumptions about human error and about their ability to design against it. This chapter tries to unpack some of those assumptions. Does human error exist as a uniquely sub-standard category of human performance? Are humans the most unreliable components in an engineered human machine assembly? Once we embrace the idea that errors are consequences, not causes, can we still distinguish between mechanical failure and human error? In fact, engineers themselves are prone to err too not only with respect to the assumptions they make about operators, but because the very activity of engineering is about reconciling irreconcilable constraints. The optimal, perfect engineered solution does not exist because by then it has already violated one or more of the original requirements. The chapter also discusses two popular ways of restraining human unreliability: procedures and automation. It tries to shed some light on why these solutions to human error do not always work the way engineers thought they would.