by Chris Furmanski

Even though we find ourselves at the start of the 21st century, computers that can make decisions with breadth and creativity that humans do are still the subject of science fiction rather than the reality of modern science. In On Intelligence, a recently released book from the founder of Palm Computing, Jeff Hawkins lays out an impressive blueprint for making intelligent machines a reality.

With resume bullets like inventor of the PalmPilot and CTO of PalmOne, a popular-science book about the future of computing certainly seems like an obvious choice for Hawkins. But as soon as you open up the sharp, electric-blue dust-cover, you’ll realize On Intelligence was probably the last thing you’d expect from a Silicon-Valley techie. Missing are detailed technology roadmaps and ethereal speculations about fantastical improbable futures. Instead, On Intelligence adeptly intertwines lay-English summaries of decades of research from neurophysiology, computer science, cognitive psychology, and even includes some well-placed philosophical sidebars that mesh into an approachable and well-written narrative addressing the plausible future of computing.

In a nutshell, On Intelligence is Hawkins’ popular-science proposal for building intelligent machines. The basic premises go something like this: (1) to build intelligent machines, we must understand human intelligence; and (2) to understand human intelligence, we must understand how the brain works.

From these intuitive assumptions, On Intelligence launches into an in-depth description of brain function, physiology, and cognitive psychology, which ultimately provides quite an education. Hawkins doesn’t shy away from the daunting task of defining intelligence, imagination, consciousness, perception, and reality using a single general theory. Dare I say, that is not an easy thing to do. Now I won’t spoil Hawkins broad-brush-stroked Memory-Prediction Framework, which is the core of the book, but his theory is an interesting, almost Ockham’s-Razor synthesis which certainly fits nicely in the pursuit of next-generation intelligent machines.

Hawkins’ concludes the book up with a brief overview of how his theoretical framework could be applied to a range of intelligent machines in the near future. Thankfully, Hawkins’ future is not deluded with fantastic visions of Schwarzenegger-esque Terminators; instead, it intelligently concentrates on the numerous pragmatic applications (e.g., smart cars or machines that think about theoretical physics) that are realizable in the next 10 or so years. I was a little surprised to find that only the last chapter of the book was been dedicated to Hawkins’ application of his theory. It felt a little like the end of Tarantino’s Kill Bill Vol 1; boom, it’s over, and the audience wants more. Hawkins probably could have gotten away with making all of the same arguments just as effectively without quite so much emphasis on human cortical physiology.

In the end, if you have a basic science or technical background and are interested in an explanation of how the human brain works, then On Intelligence is certainly for you. And even if you don’t know your cerebellum from your cerebral cortex, don’t be intimidated; one the book’s strong suits is the clarity of its writing, and Hawkins on-going use of good metaphors to simplify otherwise complex theories is fantastic.

And don’t think On Intelligence is just a great neurophysiology lesson. Hawkins mixes in well-timed (auto)biographical material (the book is co-written by science writer Sandra Blakeslee) to make it readable by readers with a wide range of interests and backgrounds. In retrospect, I think it was the biographical snippets of how Hawkins’ quest to understand the human brain molded his entrepreneurial successes that I found most interesting. For example, On Intelligence give us glimpses of Hawkins’ invention of Graffiti (Palm’s writing ‘language’) during his graduate research at Berkeley, and he shares some of his pursuits of artificial intelligence while at Intel working for the illustrious Gordon Moore. And as more of these historical vignettes come to light, Hawkins convinces us that he was not just a successful technologist, but is also someone that holds a deep-seeded, life-long passion for neuroscience--- so maybe On Intelligence’s heavy focus on brain physiology starts to seem less out of place.

Speaking personally, having spent almost 15 years in the fields of cognitive psychology and neuroscience, as well as having the opportunity to work on bleeding-edge artificial intelligence, I have to say Hawkins’ philosophy and general approach are right on the mark. That said, veterans of many of these multidisciplinary fields might find Hawkins’ theory not quite as ground breaking as advertised--- though the list of testable predictions in the appendix made the empiricist in me smile.

To make sure I wasn’t letting my closeness to field blind my assessment of the book, I polled a few of my friends and colleagues that don’t share my affinity for cognitively-inspired computing. As I expected, the general consensus was very positive and everyone I talked to would recommend On Intelligence to their friends. One psychologist I know remarked that Hawkins too often paints the world with generalities and falls into portraying the progress of psychology as more bleak than it really is. An executive I know thought On Intelligence has probably thrust Hawkins to forefront of the artificial intelligence community in the popular press’s eyes-- if CNN is going to do a story on the super intelligent computers, they probably call Hawkins first. A marketing acquaintance I know would have liked to see more illustrations in the context of the physiology descriptions, but all and all, everyone that it was thumbs up.

So, on one hand, Hawkins does a fantastic job of explaining how the human brain works in a way that is clear, well organized, and not jargon laden. The real question is, will the general public find On Intelligence engaging cover to cover? Maybe more to the point, will popular-science readers be scared away by Hawkins’ rather lengthy neurophysiology lesson when they were probably expecting something completely different from the inventor of the original PalmPilot?

It may just be that On Intelligence’s is a “betweener” --- it may go into too much detail about neurophysiology for the average reader and yet may still lack sufficient detail to be awe-inspiring for experts in the field. In the grand scheme of popular-science, On Intelligence may not have the controversial impact of Hernstein and Murray’s The Bell Curve, and probably won’t have the mass appeal of Sacks’ The Man Who Mistook His Wife for a Hat.

But as the once disparate fields of computer science and cognitive neuroscience become more aligned, On Intelligence could eventually replace Minsky’s The Society of Mind as the contemporary bible of modern artificial intelligence. In the final analysis, Hawkins’ On Intelligence is a clearly written book that would certainly be a worthy companion for science-savvy readers looking to learn about the human brain and the future intelligent machines. Enjoy!