Jeff Hawkins talks to Wired Magazine about his company Numenta. I want to believe what he is saying, but I find myself very skeptical about it all. Many people have promised breakthroughs like this before.
I always find it interesting the way philosophy plays into AI. One of the reasons that AI is so fascinating to me is that it finds itself at the intersection of computer science, philosophy, religion, psychology, neurology, and other disciplines. Hawkins says,
When you are born, you know nothing. ... You don’t know anything about tables and language and buildings and cars and computers. The brain has to, on its own, discover that these things are out there. To me that’s a fascinating idea.
This betrays his philosophy of life which is along the lines of Empiricism and John Locke’s idea of the tabula rasa. The problem with this kind of blank slate philosophy is that it really takes knowledge to get knowledge. You can’t understand a single fact in a vacuum devoid of reference to other ideas or to other people.
Some of the strongest critics of empiricism are the empiricists themselves like David Hume. They realized that to learn empirically you need induction, and to have induction you have to have an assumption that the future will be like the past, or that there is some sort of uniformity to nature and our experience of it. You cannot justify the uniformity of nature in an empiricist’s universe, because there is no logical necessity to the uniformity of nature, and to say that the uniformity of nature can be proved by the fact that things have always been uniform in the past, is to reason in a circle.
Furthermore, there are some rather large problems with identity from the empiricist’s perspective. How is it that we define a persons “identity” in a world that is constantly changing? Our bodies are constantly changing as our skin cells die and are replaced, and we age. As we learn facts our minds (if there is such a thing in a completely materialistic world) change and we grow wiser. The only thing that remains constant about me is that I am the one observing my experiences, but again that argues in a circle, and if we are honest, then we get to the same point as Hume where he famously said that we are just a bundle of perceptions.
Well…I could go into a long critique of empiricism, but honestly I don’t believe I am enough of an expert to do so, but even if I was successful at it one could still take a pragmatic approach and say, “This is the way the human brain works, so it must work, so let’s just do it.” I think on the face of it this is what Hawkins is purporting to be doing. He’s spend lots of time and money studying the way the human brain does it, and he’s replicating what he’s found in software. However, even that makes philosophical assumptions about the nature of the human brain (i.e. that it is nothing more than a machine and the sum of its parts, that it can adequately be simulated on current processors with the Von Neumann architecture, etc.).
My point is that however you approach AI there is a system of philosophy that undergirds your efforts. I’d like to see someone formulate a theory of artificial intelligence from the Reformed Christian perspective. I may take a stab at it some day.
Originally published on March 23, 2007 at 4:17 pm