i think the focus on mechanical robots is a partially misplaced human construct. software intelligence clouds are probably more likely to steer people away from control of their daily activities. the data mining and exploitation software of Facebook should be enough to give most people pause. unsupervised and value-independent decision making is dangerous, whether human or binary code. sort of like how charles manson had no serious feedback loop, only compliant help in achieving his objectives. this is an evolutionary step that can hose us, or we can make it adjunct and not independent.