Chris Hecker's talk, Structure vs. Style, was much better regarding this subject. His argument is that we don't have a Structure vs. Style decomposition for AI. We have it on the web => HTML(structure) and CSS(style). We have it in computer graphics => triangle(structure) and texture map(style).
Of course, this got the panties of all of the LISPers in a bunch.
If Kurzweil is right about even a fraction of the stuff he says, I'm fairly certain he'll prove to be an exceptionally important thinker. Unfortunately, I have serious doubts about almost everything the man says. He often seems to me to be taking the canned shot approach to writing: predict as many big things as possible and support them all as best he can, and see what hits.
"As the American economist Tyler Cowen points out, loony extropian and transhuman nanotechnoartificiallinteligenceoboosterist Ray Kurzweil is, because of his wild technoutopian prophesies, "the most important thinker today, if only in expected value terms".
Or at least he was, until this morning when while having my breakfast, I made the following prediction:
"Everything that Ray Kurzweil says will come true, and not only that but we will all get a pony"."
Hype aside, he has an interesting approach to invention. He planned on launching a tool to help blind people read. It is a digital camera with a tiny computer. During development, people worked on algorithms on PCs knowing they would be deployed on a portable device not yet available. There is some use to plotting technological growth trends.
The talk was similar to his TED talk but an hour long rather than 20 minutes. http://www.ted.com/index.php/talks/view/id/38