On Monday, I spent several hours at a Los Angeles meeting interviewing Jeff Hawkins in front of an audience of about several hundred business people. I would not ordinarily agree to take on this kind of duty, but I admire Jeff, regard him as a friend — and he asked me to do it.
One of Jeff’s main claim to fame is the invention of the hand-held computer. As a young engineer, he led a team that created one of the first “tablet computers” (with touch screens that you could write on in cursive script), and came to understand that the greatest practical use for such a computer would result from loading this capability onto a hand-held device. With the development of more-reliable handwriting recognition software assisted by teaching its users how to write less confusably (“Graffitti”, which Hawkins invented), and with the development of software that allowed you to exchange important data files with your desktop computer (“HotSync”, ditto), his “Palm Pilot” and its successor (the “Treo) were great commercial successes on the path to our PDA-rich modern lives.
Throughout this journey, Jeff Hawkins was very aware that his own brain had pattern recognition capabilities that were impossible to match with his computer software. This led him to pursue models of the real human brain, as a strategy for creating a new generation of pattern recognizers that could have the pattern recognizing power of REAL brains. To recruit the best kind of help to achieve that goal, he created the non-profit Redwood Neurosciences Institute, where he began to develop his own special theoretical perspective guiding him to the creation of a new generation of real-brain-like computers. The overall perspective that evolved from this research is documented in an important book written with New York Times science writer Sandra Blakeslee, <em>On Intelligence</em> (Time Books:New York). It’s highly recommended reading, for anyone interested in how their brain actually operates. While the Hawkins-Blakeslee treatment is heavy on theory and light on the elucidation of specific, underlying brain mechanisms and processes, a great mass of behavioral and neurological evidence supports the main premises and conclusions of the Hawkins view, to whit:
1) The brain records all information in the context in which occurs, in serial time.
2) That empowers it to grow the ability, through brain plasticity, to make ongoing serial predictions (“What goes with what?” “What comes next?”).
3) Through these brain plasticity-based serial memory/predictive processes, the machinery of your brain creates an increasingly powerful and reliable and complete model of your external and internal world, constructed from massively recorded-and-stored serial memories.
4) The recording of hundreds of millions of serial events provides the neurological bases of your incredibly powerful “pattern recognition” capability.
5) This powerful capacity for pattern recognition (reconstruction of an orderly world that the brain now knows about, and now understands) results in
a. reliable information FLOW guiding our ongoing interpretation of “What’s happening.” Ongoing prediction is the “engine” that drives your stream of consciousness (mental — and physical — ‘actions’) in controlled (useful, meaningful) directions.
b. a syntax (associative memory-based predictions of next-expected and later-expected events) that contributes greatly to representational completion and reliability
c. our capacity for “invention”, which stems from innumerable predictions, near and remote, arising from many sources and from many directions, which we can peruse and recover from our incredibly elaborate stores of complex-serially-recorded information.
I know that this superficial explanation may a little bit difficult for many of you blogsters to follow. It would help to read the Hawkins-Blakeslee book, which is not difficult to follow. ‘Tis highly recommended reading.
Would computers that operate more like human brains be useful for us? You bet! At the same time, don’t worry about being replaced by one just yet, as Numenta (Jeff Hawkins’ small brain-like-computer startup company that is building this new class of machine) is still operating with early-stage models. Moreover, YOU have been very connected to YOUR world for 20 or 30 or 40 or 50 or 60 or 70 or 80 or more years, now. You’ve probably recorded BILLIONS of serial memories/associations. YOUR predicting/pattern recognizing machine is loaded up to the gills with information that can reliably give you the mostly-right answer, and that empowers your capacity for considerable wisdom and invention from a very neatly constructed model of YOUR external AND internal world. Loading Jeff Hawkins’ Numenta machines with the information provided you about the real outside world by your eyes, ears, nose, balance organ, body, etc., and about your real inside world from your brain operations themselves is no walk in the park!
On the other hand, you can make book that it shall soon be possible to load a Numenta machine with almost anything ever recorded relating to a specific goal or task – then to put that fully-loaded machine onto a creative task. Or as in the old science fiction movies, you’ll be able to ask your Numenta-type computer what IT thinks about ITS model of the world. Before too long, this kind of brain-like machine can potentially be a VERY good source of advice! Moreover, it can gather information (live its life gathering info about what goes with what and what comes next) at very high speed — can have a head size (associative memory capacity) that dwarfs even YOUR processing capabilities – will NEVER get tired or nervous or distracted — and shall NEVER grow old, or (with suitable backup strategies) die.
Which raises the question: “Where is this all headed, Alice?!” Onward, perilously, to a Brave New World in which machines have powers of “understanding” and “invention” that shall almost certainly dwarf our individual capabilities. This represents another of a long series of evolving technologies that are being delivered into our culture with little guidance, and with no real understanding of its down-stream consequences, for modern cultures and societies. Jeff Hawkins is a wonderful, caring, do-gooder human being. He is unleashing a powerful genie from a bottle, I think with his fingers crossed.
Which is all another way of saying that you should prepare yourself for still another ton of bricks landing on your head in the form of a completely new class of computers that cn go to school and learn to be damn smart just like you did, via their own ‘brain plasticity’ mechanisms – as one of the next great ‘gifts’ contributed to humankind via the collaboration of brain science and technology! So “Hang on, Alice! The ride may be a bumpy one, and we may be driving over a cliff…..”