A vague discussion of AI risks has just broken out at http://marginalrevolution.com/marginalrevolution/2012/03/amazing-bezos.html#comments Marginal Revolution gets a lot of readers who are roughly in the target demographic for LW - anyone fancy having a go at making a sensible comment in that thread that points people in the right direction?
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.