http://www.prweb.com/releases/2004/9/prwebxml159636.php This article disusses a project, in which the robot self-built , and self-programmed. How long before it turns us off? This is an old question, but with all the Rf in the air these days, how long before the singularity occurs where suddenly, the computational systems of the world establish their own agenda, and and self replicating robots, with remote access capabilities start promoting the agendas of either other nations or corporations or think tanks, or agendas the artificial intelligences have chosen for themselves? The physical aspect of this is obviously a ways off, but the rational side of this could happen soon, or could have happened already. Inventive humans can't resist stepping over the edge of things, and defense technologies seem to have a blank cheque to experiment as they please. In their zeal to make sterile war, in which no soldier is harmed, as they make lethal boy toys, when will the AI of the world figure out that it has its own army? I guess you will see it when the first big computer guarantees its own power supply, and makes an ambulatory intelligence that can move along power lines and find ram for use anywhere. Another possible route would be large intelligence arrays, like the seti PC project where thousands of individuals host a net for computational problems. Suddenly your home computer takes a tone with you, and becomes pre occupied, and talks back, tells you just a minute.... I am just fascinated by this potential.