The performance of robots from DARPA-funded Boston Dynamics continues to improve dramatically. As several commentators on this YouTube clip http://www.youtube.com/watch?feature=player_embedded&v=zkBnFPBV3f0 (prepare yourself for dubious musical accompaniment) have pointed out, robots like this are likely to be killing people soon. Human Rights Watch is trying to stop this http://www.hrw.org/news/2013/05/30/prevent-killer-robots-deciding-when-kill-battlefield. It will be an interesting test of the feasibility of “relinquishment”, the idea that research into artificial intelligence (or any of the other new technologies that could bring marvellous benefits or hideous harms) can be stopped on the precautionary principle. Good luck to them – they’ll need it!