The latest episode of Talking Robots is an interview with Inman Harvey of the University of Sussex.  He observes that when researchers attempt to build autonomous robots, their approach is strongly influenced by their philosophy of mind, even if that philosophy is only implicit.  He also points out that what he calls "good old-fashioned AI" fails to represent how brains really work.

This is a point I have emphasized repeatedly.  Inman observes that approaches liked embodied artificial intelligence (which we discussed with Rolf Pfeifer in Episode 25) are really based on a different philosophy of mind that "good old-fashioned AI."

His paper,  Philosophy of Mind Using a Screwdriver, is available as a PDF.