In this article we outline a new technology that could allow control engineers to get the best of both worlds: state-of-the-art performance on cheap hardware. But first it is useful to take a step back. The bread and butter of control engineers is the proportional-integral-derivative (PID) controller introduced by the Taylor Instrument Company (now part of ABB) in 1940, this remarkably simple control law still does a good job for many systems today (if well tuned of course). But as systems become more integrated, with increased interactions between control inputs, it becomes increasingly difficult to design (tune) PID controllers. If you add constraints the problem becomes even trickier. Where do these constraints come from? The short answer is that most systems are constrained in some way. Valves, for example, do not go on opening forever. There may also be safety limits or product quality/performance requirements. The fact is that as you push a product design closer to its economic optimum, you invariably push your system closer to the constraints that are active at that optimum. Because of disturbances, you need to back away from these constraints. The better your controller, the less you need to back off.