I thoroughly enjoyed it and think it was really well done. I can't perfectly judge how accessible it would be to those unfamiliar with x-risk mitigation and AI, but I think it was pretty good in that respect and did a good job of justifying the value alignment problem without seeming threatening.
I like how he made sure to position the people working on the value alignment problem as separate from those actually developing the potentially-awesome-but-potentially-world-ending AI so that the audience won't have any reason to not support what he's doing. I just hope the implicit framing of superintelligent AI as an inevitability, not a possibility, isn't so much of an inferential leap that it takes people out of reality-mode and into fantasy-mode.
http://www.ted.com/talks/nick_bostrom_what_happens_when_our_computers_get_smarter_than_we_are
I realize this might go into a post in a media thread, rather than its own topic, but it seems big enough, and likely-to-prompt-discussion enough, to have its own thread.
I liked the talk, although it was less polished than TED talks often are. What was missing I think was any indication of how to solve the problem. He could be seen as just an ivory tower philosopher speculating on something that might be a problem one day, because apart from mentioning in the beginning that he works with mathematicians and IT guys, he really does not give an impression that this problem is already being actively worked on.