BRAINS

No, this isn't a post about zombies. :) I'm a sort-of armchair neuroscientist. I love studying the brain. Because of that, I've been looking forward to Jeff Hawkins's new book 'On Intelligence' about his theories on the larger overall workings of the brain. I was surprised in reading the book that his theories seem to mesh with what I've always thought about the brain works. The big addition he made to my personal theories (along with adding a huge amount of technical details) is that the patterns are also used to predict the future. When you're walking, your brain is not only telling your muscles how to move based on previous walking patterns, it's predicting how your body and the world is going to react based on those same patterns. If you've ever missed a step on the stairs, you've experienced this. Your brain predicted one thing and prepared you for that but you experienced something different.

Like I said, it meshed really well with how I've always looked at things and has caused me to reexamine a lot of things I previously explored such as AI. As a computer scientist, Mr. Hawkins's natural focus is on how we can make machines that use this pattern-prediction framework to think the way we do. He goes beyond the old ideas of having a human-like AI and moves to real applications like a weather predicting brain. His idea is to build a pattern-prediction system and then feed it data from say, weather satellites, instead of eyes and ears and the like. Then the machine would learn to make patterns and predict the weather the way we predict what we're going to see in a room even before we turn on the lights. This would allow the machine to see deeper patterns in the weather without having to be a human-like being. A fascinating idea and one I'm hoping to learn more about in the coming years. Hawkins has formed a company called Numenta to build products from these ideas and I'm going to be watching them closely. This is the guy who built the original PalmPilot and my beloved Visor handheld so he's not to be ignored. If I knew enough to work with him and his partners, I would jump at the chance. I really believe it'll turn out to be ground-breaking work.