Dear Commons Community,

I have just finished reading A Thousand Brains:  A New Theory of Intelligence by Jeff Hawkins, an author and neuroscientist who founded Numenta, a company that conducts research in the field of brain science. The book is divided into three parts:   

1. A new understanding of the brain
2. Machine intelligence
3. Human intelligence.

Part I was most informative and focuses on a new theory of intelligence called The Thousand Brains Theory wherein Hawkins proposes that the brain uses maplike structures  called reference frames to build models of the world.  He takes a deep but lucid dive into the old brain v. new brain, neocortex, columns, neurons, axons, dendrites, and synapses.  A basic element of his theory is that reference frames work simultaneously and reach a consensus of what is being perceived through the senses.  The theory gets a little fuzzy when dealing with higher-order brain functions such as mathematics and language.  Hawkins admits that his theory is speculative at this time and not proven.

Part II examines the state of machine learning as used in artificial intelligence (AI).  He questions whether AI can ever be considered “intelligent” without knowing more precisely how the brain works.

Part III considers the human condition and where it might be heading given issues related to evolution, genes, and our limited understanding of intelligence. 

While all three parts are interesting, I found Part I to be an education on brain function.  

Below is a brief review of this book that appeared in The New York Times.   

Try it if you have any interest in this subject matter especially if you have little training in it.

Tony

————————————————————————————-

The New York Times

A Review of A Thousand Brains by Jeff Hawkins

By Tali Sharot

April 16, 2021

A Thousand Brains takes the reader on a journey from the evolution of our brain to the extinction of our species. Along the way Hawkins beautifully describes neuroanatomy and landmark discoveries in neuroscience, including the existence of cells that signal our location in space and populations of neurons that process information by “voting” to reach a group decision. The book is framed around Hawkins’s theory of intelligence, according to which columns in the neocortex encode thousands of “reference frames.” However, the theory has yet to be empirically tested, and he does not spend much time detailing it or how it could account for high-level functions such as language and thought.

Hawkins, the inventor of the PalmPilot and a neuroscience researcher, aims to crack human intelligence in order to develop artificial intelligence. The problem with current A.I. systems, he explains, is that they can solve only a limited set of predefined problems. They do not possess general intelligence as humans do. According to Hawkins this is because they are unable to represent knowledge. Part of Hawkins’s motivation for developing “true” A.I. is to prepare for human extinction.

Although not predicting when or how we will meet this fate, Hawkins advises readers to craft an “estate plan for humanity” now. Homo sapiens may try to dodge extinction by habitation of other planets, but Hawkins is not optimistic. Instead, he suggests, we would be wise to use machines to preserve human knowledge for the benefit of other beings, even if we are unable to sustain mankind. With this and other ideas Hawkins keeps the reader constantly engaged.

This entry is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license.