I just finished reading this book by Jeff Hawkins. It basically explains his theory on how the brain works and how it creates intelligence.

It’s an interesting theory. The basis of it is that the brain is an associative memory machine, and most importantly it uses those memories to make predictions about its future inputs. Whenever these inputs deviate from the brain circuit’s prediction, a new lesson is learned. Other notable claims he makes are that there should be a simple algorithm all brain circuits use, that when scaled up makes intelligence. There’s no need for a complex algorithm with completely different behavior in different parts of the brain, etc.

My main complaint is that he doesn’t provide enough details to implement his ideas in a computer, or for a programmer to even get a feel for how the algorithm might work in low level detail.

It’s an inexpensive book, so if you have any interest in the workings of the brain, or intelligence in general, you ought to pick up a copy.