Intelligence And The Machines That Have It

Rating:
Summary: finite refinability = no intelligence
Review: This book introduces the concepts of infinite refinability, and finite refinability and splits the world apart into those two categories. Nearly everything falls into the category of infinitely refinable, except for digital computers that are by definition finitely refinable. Then it examines the consequences of that, and what it means for artificial intelligence.

The argument is made that intelligence is infinitely refinable, computers are not, therefore computers do not even posses the ability to have intelligence.

Or to put it more concisely, a digital machine at any point in time can be represented by an integer--it's Godel number. (An extremely large, but finite integer.) Since intelligence is not an integer, no digital machine at no point in time can be intelligent.

Then this analysis is applied to the Turing Test, and Searle's example of translating Chinese to Spanish languages, and explains how computers will not accomplish it. (They will attain any arbitrary level set by finite requirements.)

And it analyses different myths about AI. Is playing chess intelligence? Is anthropomorphizing machines intelligence in those machines? Is there some critical-mass of programming that can be achieved that will somehow magically make the machine intelligent beyond that? That's what this book is all about.