In the late 70s, it seemed that every s.f. writer had to write an Artificial
Intelligence story. For the most part, these stories described humanlike
intelligence being somehow instilled into conventional systems not very far
removed from the big mainframes of the day--usually by a smart, young, maverick
programmer--or arising spontaneously in a sufficiently large network.
I never found either of these scenarios convincing. Coming anywhere close
to mimicking the stupefying complexities of thinking processes was way beyond
any plausibly achievable technology of the times--or in sight today, for that
matter. And without any selective improvement or refining factor in operation,
extending a network merely results in a bigger network, not a smarter one.
It seemed to me that a more realistic setting for true AI could come about would
require a technology more advanced than anything we can visualize today already
in place, being directed toward that end by some of the best expertise in the
I had been thinking about a book along these lines for some time when I arrived
in the U.S. with my second wife, Lyn, in 1977. What I needed before getting
seriously down to writing was to talk with somebody who knew something about
AI and associative learning processes, and could comment on what did and
didn't sound plausible. One morning over breakfast I said, "Who
do we know who's into machine intelligence? I need to bounce my thoughts
off somebody." Lyn replied, "We've only been here a couple of months.
We don't really know anybody." Literally the next day, the phone
rang and a voice said, "Hi. We haven't met, but my name is Marvin
Minsky. I run the AI lab at MIT. Nobody's produced a good book on
AI yet. Judy-Lynn Del Rey told me to read Inherit the Stars.
So I did, and I think you could write one. How would you like to
come to the lab and see what we're doing, and let's talk about it?"
Which was how Marvin and I got to know one another. You couldn't ask for
better Divine Intervention than that.