Mark already posted a glowing review of this book. I was a little
disappointed in the book, so decided to write my own review. Mark, I don't
disagree with your review -- there were many interesting ideas. But perhaps
my expectations, and what I wanted to get out of the book, were out of line
with what the book provided. I wanted good analysis and strong conclusions,
and instead I got:
- History of the development of genetic theory, including interesting
little anecdotes about the personal lives of some of the people involved
- History of the development of computers, including interesting little
anecdotes about the personal lives of some of the people involved
- History of the development of networks, including interesting little
anecdotes about the personal lives of some of the people involved
- Speculative discussion of ideas involving evolution, computers and
networks
The history and little anecdotes were pretty interesting, and it was a lot
like "Connections", linking together ideas and people and new ideas.
However, I knew most of the information, except for the juicy irrelevant
details e.g. personal relationship between von Neumann and Mauchly/Eckert.
Eventually Dyson moves on to make arguments for why intelligence might well
arise, due to selection pressure, among computers or computer networks.
Related topics include the nature of group intelligences (like an ant farm).
I started to bring up some of the ideas to discuss among my critical
friends, and they immediately came up with several objections and issues and
problems with the ideas, which Dyson never dealt with.
The problem of the major premise (evolution causing artificial
intelligence), as I see it, is that selection pressure doesn't necessarily
mean selection for intelligence. We like to believe, as humans, that
intelligence is the pinnacle of evolution and therefore the primary goal,
but it doesn't seem to be true: "survival" can be due to many other factors
besides intelligence. Selective pressure is applied to computers and
networks to make them more useful to us, not more intelligent.
That said, the book is an easy read with good explanations of technical
subjects. It is well written, and the anecdotes are intriguing and
entertaining. The book could be sub-titled "A lot of intriguing history and
some unanswered questions relating to evolution of artificial intelligence"
-- but then again, I think most books on artificial intelligence raise many
more questions than they answer. This book is particularly speculative, and
I feel that speculations should be more clearly labelled so.
One thing I did get out of the book was a desire to look up some of the
lesser known authers Dyson quotes -- such as Olaf Stapledon and Samuel
Butler. It is clear Dyson is very widely read and knows a lot about the
fields involved.
If you approach the book expecting a poetic romp through ideas, from
scientists and science fiction writers, about computers and networks and
evolution, you may enjoy it much more than I did. But if you expect
rigorous science, the book will be disappointing.
Then again, a lot of people have loved this book. The recommendation is
already in the archives, but here it is again:
http://www.amazon.com/exec/obidos/ISBN=0201406497/forkrecommendedrA/
You might want to borrow it from a library since it's more of an interesting
read than a reference to keep around.
Lisa