Monday, May 28, 2007

Go Jeff, Go!

[Originally posted on Friendster Blog Mentation Station on 3.24.05 at 8:24PM, prolly from VC]

It seems that Jeff Hawkins' book, On Intelligence, and the goals he set out within it are finally making a splash.  Posts on Slashdot and BoingBoing, as well as stories in the NYT (you know, subscribe) and other outlets are reporting that Hawkins and fellow Palm entrepreneurs Donna Dubinsky and Dileep George have finally founded Numenta, which is going to attempt to implement Hawkins' theories from On Intelligence.

Last semester, I was fortunate enough to read this book with a group of Cognitive Science majors, at the behest of the all-knowing Prof. Broude, for the Cog Sci book club.  Alongside some more hearty works, such as William Uttal's The New Phrenology (recommended to destroy your hopes of localizing brain processes) and Lakoff & Johnson's Philosophy in the Flesh (recommended to destroy your love affair with just about all philosophy post-Aristotle), Damasio's The Feeling of What Happens (just plain recommended), On Intelligence was a bit of a breeze to read (thanks in part, no doubt, to bestseller-maker Sandra Blakeslee).  But it did set out a fairly radical approach to understanding cortical function and how it underlies intelligent behavior.  Hawkins' theories, I believe, fit in as the next step from the work of Joaquin Fuster, Gerald Edelman, Vernon Mountcastle, and some others.  They all emphasize the distributed, network nature of cortical processing, the role of feedback, and the importance of memory as a key cortical process.  In my humble opinion, the work of these guys pretty much represents the future of brain research, and, as Hawkins is showing, perhaps the future of AI as well.

No comments: