Context: Watched a YouTube TED talk a while ago. Interesting ideas. Hoffman is a cognitive scientist at University of California, Irvine.
Hoffman’s main argument is that our senses are tuned by evolution for ‘fitness’. They are NOT tuned to perceive the ‘true’ nature of the world. The analogy is that what we perceive is similar to a computer desktop. The icons on the desktop and the actions we take on them are not a true reflection of objective reality.
One implication of this is that our model of the world as a 4D space-time with ‘objects’ that persist when not observed by us is not necessarily ‘true’.
Hoffman argues this from a physics perspective: quantum mechanics and the collapse of the wave function; attempts to develop a quantum version of general relativity (and a TOE); the holographic universe theory; etc.
He also argues this using examples drawn from visual perception and commonly described visual illusions/mis-perceptions.
A key problem in cognitive science (and philosophy and artificial intelligence) is the nature of consciousness. Hoffman argues the the bottom-up model starting with basic physics to neurons to computation has not been successful in expelling consciousness. He proposes a top-down approach instead.
The top-down approach he proposes is that reality can be modelled as a collection of agents. The agent perceives, decides, and acts. The act changes the state of the world, and the agent then repeats the cycle. A Markov model can thus be used to describe the agent. The ‘world’ is just a (large) collection of agents. The ‘agent’ may be thought of as ‘conscious’.
Factoids of interest:
- Human vision is a 130 megapixel sensor (65M rods/cones per eye).
- The high-resolution part of the eye contains most of the cones.
- This high-resolution part of the eye view a path of the world the size of the thumbnail held at arms length.
- One third of the brain is devoted to visual processing.
- Visual processing collapses to 130 megapixel raw input to something like 40 bits of information.
An interesting book, but left me unsatisfied. I found the arguments from physics unconvincing and not fundamental to his premise about evolution optimizing for fitness rather than truth. The perceive-decide-act-world cycle sounds interesting, but the book does not provide much in the way of implications.