After a solid intro from Hawkins, Stefan Rudnicki takes over the narrating reins. The effect is an audio program with a compelling ability to anticipate the question taking form in your own brain as you listen, then answer it with clarity and sincerity. That's a feat worthy of admiration.
Hawkins develops a powerful theory of how the human brain works, explaining why computers are not intelligent and how, based on this new theory, we can finally build intelligent machines.
The brain is not a computer, but a memory system that stores experiences in a way that reflects the true structure of the world, remembering sequences of events and their nested relationships and making predictions based on those memories. It is this memory-prediction system that forms the basis of intelligence, perception, creativity, and even consciousness.
In an engaging style that will captivate audiences from the merely curious to the professional scientist, Hawkins shows how a clear understanding of how the brain works will make it possible for us to build intelligent machines, in silicon, that will exceed our human ability in surprising ways.
Written with acclaimed science writer Sandra Blakeslee, On Intelligence promises to completely transfigure the possibilities of the technology age. It is a landmark book in its scope and clarity.
©2004 Jeff Hawkins and Sandra Blakeslee; (P)2005 Audible, Inc.
"[Hawkins's] argument is complex but comprehensible, and his curiosity will intrigue anyone interested in the lessons neurobiology may hold for AI." (Booklist)
"[Hawkins] fully anticipates, even welcomes, the controversy he may provoke within the scientific community and admits that he might be wrong, even as he offers a checklist of potential discoveries that could prove him right. His engaging speculations are sure to win fans." (Publishers Weekly)
1. A sudden manifestation of the essence or meaning of something.
2. A comprehension or perception of reality by means of a sudden intuitive realization: ?I experienced an epiphany, a spiritual flash that would change the way I viewed myself? (Frank Maier).
Jeff Hawkins proposes a theory for how the mind works that rings true. I found this book to be a profound mind-expanding experience, allowing my mind to finally understand how my mind works. I am a 45-year old Software Engineer strongly tempted to take up Hawkin's call to young scientists to join the new revolution that will take place as we learn how to apply this model of mind to computers.
This book will probably be of most interest to people who are interested in both computers and psychology, but I think many people who are not interested in computers will still find this book fascinating. Anyone who is interested in the miracle of consciousness and sentience should read this book.
Another reviewer complained that the "ghost narrator" is a poor reader. This narrator is Stephan Rudnicki, a professional voice actor with a deep, resonant voice. He is one of the reader's on the Ender's Game series. There is something about his voice that is a bit unsettling, but I would never call it simpering. Overall, I think he was an excellent choice for narrator for this book.
Note: there are some figures/diagrams in the printed book that can be downloaded from audible.com. The link is on the main page for this book.
This is, all and all, a pretty solid book, particularly if you already have some interest in the field. It explains both current state of modern neuroscience as well as the author's theory of the basis for intelligence. If you are interested in this topic it's probably worth the download. You will have to put up with a few minor irritations, such as (1) the author's ego occasionally bleeds through with little snotty asides. He's fairly arrogant and quickly dismissive of alternative views, (2) his theory seems internally consistent but is ultimately a little reductionist. The brain almost certainly does work, at some level, just as he describes it. But when he pushes the model into describing the creative process, the basis of consciousness and some other areas, it feels a little thin. It's an engineer's view of the essence of these aspects of human behavior, overly wedded to a simplistic structure, ignoring nuances which are clearly important. (3) There is one chapter in the book which doesn't translate into a book tape very well, because it's fairly technical and relies on diagrams. But even if you don't entirely drink his Kool-Aid, this is still an interesting, thought-provoking way to spend a few hours.
I focus on fiction, sci-fi, fantasy, science, history, politics and read a lot. I try to review everything I read.
An excellent book well worth the listen. Not very technical, and a lot is left out. The most frustrating is the almost total lack of discussion of feedback, neuron death, and connectivity changes. I never found this tedious, indeed I would love two more volumes. Even if the authors theories are utterly wrong, anyone interested in intelligence or AI would benefit from reading this book. The author is a bit of a salesperson, a bit conceited, but may very well be correct, and is, in any case, quite interesting.
I don't normally talk to friends about what I'm reading. But for the past two weeks I've been obsessively telling friends - even my 78-year-old mother - about this book. Maybe it's just me, but I find hearing On Intelligence has changed the way I think about computing and human intelligence. I suddenly feel I understand a range of phenomena which have intrigued or baffled me for years.
I can't do justice to Hawkins's thesis, but I'll take a stab. He claims to have figured out how the cortex - the part of the brain that was "new" to mammals, and whose size seems to be the greatest difference between humans and other primates - works. In other words, he asserts a general theory of what makes humans intelligent, and more generally, of what we're doing when we think.
Based on several lines of evidence, he rejects the idea that the cortex is primarily sifting through input data, looking for general patterns, with each stage summarizing and passing information up to more abstract levels. Instead, Hawkins asserts that the cortex is mainly MODELING the world we sense, and spends the great bulk of its effort actually passing predictive data DOWN to lower levels, including sensory areas.
These predictions are broken into an unbeliveably detailed representation of the world, modeling at the level of individual sensory neuron, what we will see, hear, touch, in the next second or two. Wherever this prediction is more-or-less right, it is treated as "confirmed" and the world we experience is mainly THAT PREDICTION, not a summation of this instant's actual nerve receptors' sensations. Where the prediction is NOT confirmed, an "exception" is generated, which either causes minor adjustments in the predicted scenario, or draws our conscious attention to the unexpected event.
If that doesn't make sense, listen to the book. There's a lot more to it.
This book has changed the way I see the world. I think it's a really big deal.
Old & fat, but strong; American, Chinese, & Indian (sort of); Ph.D. in C.S.; strategy, economics & stability theory; trees & machining.
I'm not qualified to judge how well this book stands up to peer-review within the biological sciences. I'd guess that it not fully accepted by high-end biological researchers, but I'd also guess that this has more to do with the interdisciplinary nature of the work than with any fundamental disagreement about the big concepts. However, I am qualified to say that as a continuation of the cognitive science tradition within computer science this book is simply masterful.
The overall thesis of the book is that the human brain is architecturally not a computer, but rather a smart memory. Of course, this is an architectural distinction rather than a functional distinction. But the distinction yields important clues about how to build smart machines.
The book goes on to propose that memories are recalled using a mixture of forward error correction and search that is biased on invariants. The algorithmic details of this recall mechanism are not fully understood, but can be imitated, at least clumsily with various electronic circuits. The book further suggests that learning is the combination of laying down memories and finding useful patterns in these memories.
My husband and I have altered our language. "That didn't make it to the fourth level," we say to each other when there's an automatic reaction to an event. Or "Maybe you need to engage the sixth level..." when facing a problem. And: "Intelligence is the ability to predict..." when my husband's keys are still in his pocket as he approaches the door of the car...
This book was sometimes redundant-- but repetition leads to retention and I needed it sometimes.
I have a Ph.D. in applied behavior analysis because learning theory has always intrigued me. This was the closest I've ever come to feeling like someone has united the nature-nurture issue. I often think of the question he asks: What would a brain do with an eternity of learning?
What would happen if each generation did not have to re-learn the same things?
There were some areas I felt were incompletely developed. I want to know how emotions interact with the learning process. I want to know to enhance the generalizations which lead to creative thinking. How do we Brain Train in a general fashion?? And how do we untrain the brain?
A book that leaves me with the desire to know more is definitely a book I'd recommend to friends.
Gen-Xer, software engineer, and lifelong avid reader. Soft spots for sci-fi, fantasy, and history, but I'll read anything good.
If, like me, you're a software developer with an interest in true artificial intelligence, this is a very stimulating book. Hawkins applies his own engineer's mind to an effort to discern and describe the human brain's underlying "cortical algorithm", the means by which intelligence "works". As Hawkins sees it, the neuroscience community has been too focused on the minutae of how neurons function, without giving adequate consideration to the brain's overall learning and decision-making architecture, while the computer science community has been too absorbed in traditional symbolic and procedural computation methods, ignoring insights that might be gleaned from studying the most powerful problem-solving system in nature. Of course, it's untrue that neuroscientists and comp-sci academics aren't interested in each other's disciplines, but the crossover is still a long way from mainstream. For coders working in industry (like me), Hawkin's thoughts may be revelatory.
The author focuses most of his attention on the cortex, the most recently evolved part of the human brain, and the one responsible for many functions of higher intelligence. His speculation is that this system uses the same generalized learning/prediction algorithm throughout, with little difference in how input from vision, hearing, touch, and other senses are processed. All this data is just sequences of patterns that the cortex filters through its multilayered hierarchy, each layer discerning trends in the input from lower layers, and forming models of the world.
This may sound like the traditional AI concept of "neural networks", but Hawkins breaks from that model with his view that the cortex uses massive amounts of feedback from higher, more time-invariant layers (which view the world more abstractly) to lower, more time-variant layers (which deal with more concrete experience), activating many context switches. He sees the cortex as a blank slate upon birth, which follows relatively simple programming to accumulate and categorizes knowledge. As our minds form, we find ourselves experiencing the world less through our sensory input, and more through our pre-formed models. Only when there is conflict between those models and our input sequences, is our conscious attention drawn to our senses.
In terms of biological neuroscience, this is all probably overly simplistic and not completely accurate (Hawkins doesn't give a lot of attention to the older, more instinctive parts of the brain), but if he's even partly right, his ideas have huge implications for artificial intelligence. If much our human intelligence really does boil down to a generalized memory-prediction algorithm -- one that may be complex, but not beyond our understanding -- the effects on the future will be astounding. Even if Hawkins wasn't able to prove his claims, they're fascinating to contemplate, and the next few decades will certainly shed a lot of light on their truth.
If this book speaks to you, consider also reading Marvin Minsky's A Society of Minds, which contains a lot of complementary ideas.
Having listened to many non-fiction Audible books, some good, some mediocre, this was my first listening experience that literally captured my imagination. As a result I found it rather difficult to stop listening and attend to my daily chores. I've struggled for many years to understand the mechanics of what brains do with very little success. As Jeff began to lay out his framework my own brain was firing patterns of YES, YES, YES! I began to experience the very rare and pleasant sensations that occur when getting a first glimpse of a puzzle solution. There's nothing like it in the world! This will be the first audible book that I feel compelled to purchase in hard copy. I'm glad I listened first; it's the perfect format for experiencing the marvelous fun in thinking about thinking and the profound joy in gaining understanding about understanding.
Of course, I can't guarantee that you'll experience anything similar to what I've described. I can guarantee that Jeff will provide you with a fascinating and powerful framework by which some surprising clarity can replace the murky waters of notions like "intelligence", "understanding", and "reality".
This is one of the worst books I have heard.
Yes, there were some interesting ideas.
But there were so many errors and faulty assumptions, that it made it hard to accept even his simplistic and obvious claims (and he made many of these).
I resented his assertions that everyone but him, all scientists and AI (Artificial Intelligence) researchers in particular, just don't get it. I assume that some of this attitude comes from the fact that MIT rejected him. But many of his claims regarding what certain fields think about the brain are just wrong.
For example, he mentions several times that AI researchers believe that a computer is just like the human brain. While some may talk metaphoricaly of the mind as computer, I have never heard the claim that the human brain works just like the hardware of a computer.
(disclosure: My PhD is in AI)
I greatly enjoyed On Intelligence. Not only is it well written and easy to get into (at least in most cases), but it introduces a new paradigm for neurological functioning that will probably replace the models most commonly used by engineers, doctors and academics. This new paradigm has the potential to lead to enormous breakthroughs in artificial intelligence and other areas that stand to greatly benefit mankind in general and individuals in particular. This is a truly multi-disciplineary book that I found both entertaining and fascinating.
Report Inappropriate Content