Jeff Hawkins is interested in building a theory of the brain. Apparently, the various initiatives approaching this, like AI, neuroscience, and psychology, have certain incorrect assumptions preventing them from building a theory. Jeff Hawkins explains what he thinks that incorrect assumption is and what he’s doing to try and get the ball rolling towards a theory of the brain and creating real intelligent systems.
The thing I like about Jeff’s approach is that he’s integrating these separate fields and then building stuff. I consider that an ideal recipe for progress. I’ll have to look into what he’s done since 2007, but I found this talk pretty enjoyable.
Here is another great talk by Chris Hecker on Structure vs Style. It’s an exploration of a pattern in problem solving, particularly for “hard interactive problems” common in games and other interactive systems, of separating structure and style. The primary example used to convey this decomposition is the texture mapped triangle. He goes on to suggest that AI is likely to have a decomposition based on this structure/style paradigm, and that this means we may someday have a “Photoshop of AI.”
