As linked from Ken McLeod, eSkeptic Website rains on the Strong AI hopefuls' parade.
I'm always suspicious about any AI skeptics who don't at least mention the explorations of self-referential semi-random systems by Douglas Hofstadter. His approach seems most likely to avoid the Scylla of Minsky's reasoning-machine approach and the Charybdis of overly imitative neural models that simulate the brain without thinking about what thought is. But I haven't heard of anything new coming out of the Fluid Concepts and Creative Analogies team, so that might be a dead end for all I know.
It's just as well. I'd like to start collecting social security before the robot takeover.