This was my first venture into a full book-length, nonfiction project.
One of the great things about life is that education never really has to end
if you keep looking hard enough.
Well, actually, I didn't have to look. One day in September 1996 the phone rang,
and my agent, Eleanor Wood, said, "Owen Lock [of Ballantine] wants you
to do a book on AI for them." Apparently there had been a management meeting
there, and somebody had proposed bringing out a popular-level coverage of the
subject to coincide with the Kasparov-Deep Blue replay scheduled for May, 1997.
So I flew up to New York to talk about it. The rub was that they wanted it done
by January-February to meet the scheduling. It was a subject that I'd touched
on a couple of times and thought I could get up to speed on by filling in details,
so in a classic example of confidence being what you feel when you don't fully
understand the situation, I said yes, I thought I could manage that.
In the weeks following, I traveled around talking to people like Marvin
Minsky at MIT (an old friend), John McCarthy at Stanford, Herbert Simon and
Hans Moravec at CMU, Doug Lenat at Cycorp, Texas . . . and gradually the size
of what I had taken on unfolded in all its ghastly enormity. The result
was a marathon of 10-12-hour-day, 7-day weeks through to the day I FedExed off
the final package. No Thanksgiving, No Christmas, No New Year's.
Mind Matters isn't intended as a textbook for Ph.D.s to add to
their shelves--scores of excellent ones have been written already, and I'm not
the person to add to them. Neither is it a history of the field-- which
has also been well covered by other writers in ways that need no improvement.
Rather, it's a mixture of background and techniques, with an emphasis of understanding
why, but also with a historical thread--as the subtitle says, more of an amble
around the world of AI, stopping to have a look at assorted things that I, personally,
find of interest.