Brainmakers

by David Freedman
New York: Simon & Schuster, 1994
214 pages, includes Index and Bibliography.
Reviewed by Sam Werberg
November 1996

Creating intelligence is proving to be not as easy a task as some of the early AI pioneers had believed it to be. Any effort must start first with a solid theory of how the mind works, how observations are made, and how choices are made and actions taken. Then that theory must be applied using the appropriate technology. And behind it all we have the reasons for desiring the technology in the first place. David Freedman uses this book to tell the tale of some of the leaders in the AI field and the paths, technologies, and theories they are following. Each chapter showcases a different school of thought in somewhat of a chronological progression. In the end, Freedman wants to show us that a nature-based approach to AI will eventually have to win out over conventional AI if the field is to continue to grow and break new ground.

Conventional AI has been concerned with using computer hardware and software to attempt to accomplish tasks that humans currently undertake. The conventional approach, as Freedman characterizes it, is based upon rules and facts, 1's and 0's. This is in contrast to the nature-based approach which says that we should model artificial intelligence directly from how nature actually accomplishes tasks. This may seem to go without saying, but as Freedman demonstrates in the first half of the book, conventional AI thought that it could replicate nature while not considering how molecules, neurons and DNA actually work.

As a documentary of the current and past groundbreaking work in AI, this book is excellent. The author conducted personal interviews with most of the featured researchers, and it comes through in the candid remarks made by researchers about other researchers' work. Included are most of the pivotal people in the field, from early pioneers like Marvin Minsky to cutting edge techno-wizards like Pattie Maes. Freedman does a good job of stepping back and letting the researchers and their work defend themselves for the reader. At the same time, he keeps us tuned in to his framework concerning the evolution of "brainmaking".