The Cambrian House blog has a great piece on Harvard's take on prediction markets like Cambrian House's "IdeaWarz" that leaves product promotion and development to a competitive process. Popular ideas get center stage; unpopular ideas get nada (that's where my ideas end up :( ).
I like the idea of prediction markets and crowdsourcing because it smacks of natural select and psychohistory: that crowd opinions will wash out peaks of dissension and what I called "Digg pinheadism."
Anyways, I urge you give this Cambrian House piece a read. It's interesting regardless of whether you like or hate these concepts.