With the development of the General Atomics MQ-1 Predator, robotic weapons came of age. The operations of this unmanned aerial vehicle (UAV) in Iraq, Afghanistan, Pakistan, and northern Africa in the last few years have given us a glimpse of the future of high-tech war. It is a future in which thousands of miles separate those firing weapons from those whom they kill, in which joystick jockeys have replaced pilots and soldiers, and in which the psychological barriers to killing are greatly reduced by the distance between weapon operators and their targets. Perhaps more importantly, it is a future in which wars are more likely, in which decisions about when weapons are fired and who they are fired at are increasingly in the hands of machines, and in which the public has little knowledge of or control over what is being done in its name. Finally, it is a future that is likely to come about not because it represents a better, less destructive, way of fighting war but because the dynamics driving the development of unmanned weapon systems (UMS) are likely to dictate that they be used more and more often. Now that we have had a glimpse of this future, it is time to begin thinking about whether - and how - we might avoid it by adopting an arms control regime designed to limit the development and deployment of robotic weapons.