This fascinating book is really three books in one. It is well written and fairly easy to read.
Part 1: A New Understanding of the Brain
This is the section of the book that is most exciting from the context of the theory that has been developed and the quality of the exposition. It reflects the lifetime that the author has devoted to explaining how the brain works.
Hawkins provides a theory of the functioning of the neocortex which is dramatically different from other theories. He asserts that his theory covers the phenomena that we can observe and have tested over time. It is at once simple and complex, providing a compelling description of how it is we are able to function in our world and in particular how we are able to learn and extend knowledge.
The core of the theory is centered on the cortical columns in the neocortex. These structures are essentially the same throughout the neocortex, even though various areas of the neocortex seemingly handle different function. Each of these columns is capable of controlling movement, of learning and of creating predictive models. These models are called reference frames and through these we sense our world, create our thoughts and remember things.
While the total description of the theory is long, Hawkins leads the reader to an understanding through a well organized and articulated explanation of his thoughts in a manner that provides comprehension and awe.
Part2: Machine Intelligence
In this part of the book, the author tackles the issue of machine intelligence. He first dispels the notion that what passes for “artificial intelligence” in modern parlance is intelligence. He also rejects the notion that smart machines that can defeat a human in a particular field (say at playing chess or go) have intelligence.
He provides a list of four criteria that a machine would have to fulfill so as to be considered an intelligent machine:
• The ability to learn continuously
• The ability to learn via movement
• The capacity to create many models of objects
• The ability to use reference frames to store knowledge
The arguments behind this list are compelling and clearly indicate how far we have to go to create machines with intelligence, while holding out belief that it will come to pass.
Part 3: Human Intelligence
In this section of the book the author diverges from his scientific areas of expertise to ponder on the future of intelligence in the collective sense of the accumulated knowledge of mankind.
He spends considerable effort on discussing how people can have different beliefs on things – referring to some of these as false beliefs. Yet, he takes as given issues that are much argued from credible points of view – seemingly ignoring his concept of reference frames. At the same time he gives very good descriptions of how our models of events and objects, stored in our personal reference frames, are almost always going to be approximations of realty if for not other reason than that we do not have the full ranges of inputs by which to create a perfect model.
He ponders on the risks of our intelligence to our survival as a species and ways in which we can extend our species survival through interplanetary diaspora as well as through preserving our accumulated knowledge through “estate planning for humanity”