NPR had a piece yesterday on the idea that machines may one day be so intelligent as to become capable of designing and building even more intellingent machines, triggering a exponential growth of intelligence, an intelligence sigularity, such that human intelligence becomes negligible by comparison. Some consider this intelligence sigularity a existential threat.
Here is a link to the NPR Story:
In this story we learn of The Singularity Institute, dedicated to "Ensuring humanity's future in a world with advanced artificial intelligence."
I'm writing this short note simply because the man who introduced the idea in 1965, Dr. I. J. Good, was a friend of mine. Here is a citation to his seminal paper on the idea:
Good, I. J. 1965. Speculations Concerning the First Ultraintelligent Machine. In Advances in Computers, vol. 6, ed. F. Alt and M. Rubinoff, pp. 31-88. Academic Press.
This article is available online.
Dr. Good was a remarkable fellow. I knew him as a statistics professor at Virginia Tech in the mid 1970s. I only vaugely understood his very human intelligence, except to percieve it as much greater than my own.
I asked him once how he came to Virginia Tech. "Well, I said if they'd double the salary, and remove the teaching assignments, I'd come. And they did, so here I am."
He was also the most prolific human being I believe I have ever personally met. He had an outer and inner office, and the walls of both were covered with bookcases, overburdened, and spilling back into the rooms. I was astonished and asked
"You've read all this?"
"Oh no. I wrote "all this'"
Good was a very strong chess player and I asked him to sponsor our chess club. He asked what he would have to do, and I assured him "nothing, really," to which he replied simply "alright then."
One evening after chess club, a very faceitous friend of mine invited Dr. Good to join us for a late show at the movies. The three of us watched Jack Nicholson in "Five Easy Pieces," and after had fun analysing it into more than five easy pieces.
There was only one other faculty member on campus who expressed an interest in chess, a computer science type who was, as I recall, a chess Expert named Kalan, one notch below Master. Kalan agreed to do one of those simultaneous exhibitions to help promote the chess club. I don't now recall how many games he played simultaneously, but I would guess twenty or thirty.
I asked Dr. Good if he would participate and play against Kalan. His reply was pure game theory:
"If I should win, people will say it is only because the other fellow was playing so many games simultaneously. And if I should lose, well, in that case I lost."
I knew only a small slice of the man, but I'm better for having that. Only after he died in 2009, at age 92, did I learn he had been a consultant on supercomputers to Stanley Kubrik during the filming of "2001: A Space Odyssey".