You know the idea of the homunculus? He is a little man looking at a screen inside your head, who sees pictures for you. He is invoked by people trying to explain how human vision works. It is an example of bad realistic philosophy and you should see the idea is circular and totally stupid. Anyway, the story is that an MIT professor (could have been Minsky) assigned computer vision as a student summer exercise. How, you might ask, could anyone be so naive?
Not only does the "homunculus" fail to explain anything, you have to acknowledge that people who slip in to that way of thinking really don't understand the problem that they are trying to solve. The problem is not how do we recognize objects in the world? The problem is how do we turn the world into objects? Until we get started on that, the Deep AI stuff (which thinks objects already exist and is a correlation engine) will be dumb no matter how wealthy the "intellectual" who hypes it.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment