[ Jocelyn Ireson-Paine's Home Page | Publications | Dobbs Code Talk Index | Dobbs Blog Version ]

Heroes and Differences

Here is a grook — a short rhyming epigram — by Piet Hein:

No cow's like a horse,
and no horse like a cow.
That's one similarity
But which are more important? Differences? Or similarities? In Stewart Brand's The Media Lab: Inventing the Future at M.I.T., continuing the interview that I quoted in Consciousness is not a Window, Marvin Minsky answers.

He [Minsky] said he was trying to attract people to the Lab interested in working on constraint languages, or at least attract hackers who might grow into it. "But I don't know whether it's possible anymore, because good hackers are very quickly aware of their hundred-thousand-dollar value making products for people. I look for selfish people who don't give a damn what happens in the outer world for five years. At some point you need a hero who will actually work for himself rather than make it easier for others to work. All the people who have short-range goals will be forgotten."

All of Minsky's examples of hard problems seemed to circle back on semantic questions, problems of meaning and cognition rather than just signal processing. "As far as I'm concerned," he said, "the heuristic for making discoveries is start with a distinction that people make and argue that there are three ways rather than two. Probably all good ideas start by making a distinction, and then they usually die by stopping there and dividing everything up into those two. Information theory is interested in signal and noise. Maybe we should make a tri-stinction — signal, noise, and meaning."

After a dinner of take-out dim sum, Minsky, who had been reading the Koran with some dismay at its violent inquiry-blunting formulae, sermonized, "Religion is a teaching machine — a little deadly loop for putting itself in your mind and keeping it there. The main concern of a religion is to stop thinking, to suppress doubt. It's interested in solving deep problems, not in understanding them. And it's correct in a sense, because the problems it deals with don't have solutions, because they're loops. 'Who made the world?' 'God.' You're not allowed to ask, 'Who made God?' "

I said, "Science feels and acts like a kind of religion a lot of the time." Minsky had heard that one before: "Everything is similar if you're willing to look that far out of focus. I'd watch that. Then you'll find that black is white. Look for differences! You're looking for similarities again. That way lies mind rot." That lively loop has been cycling in my mind ever since.

If the heroes we need should be focussing on differences, and on hard problems, and not on things made for other people, perhaps they need to pay attention to another of Piet Hein's grooks:

I concentrate on
   the concentric rings
produced by my pen
   in the ink.
The thing that distinguishes
   thoughts from things
Is that thoughts are harder
   to think.