In memoriam of John McCarthy

[melajara]

This gentleman was one of the few giants from Computer Science still among us. He just passed away at 84.

Yes, this is the man who pioneered AI in 1956 at the famous Dartmouth conference.
In 1958 he devised LISP, still one of the most intriguing and potent computer language, especially as LISP has never been superseded in meta-programming abilities.
Then, most logically, McCarthy won a Turing award for his achievements.

McCarthy was also involved with computer chess, as one of the ways of exploring computer decision-making
(remember Nobel prize winner Simon and his General Problem Solver).

The following is a quote from this source: http://hightechhistory.com/tag/john-mccarthy/

In 1962, the first chess program at the Massachusetts Institute of Technology was written. It was the first chess program that played chess credibly.
It was chiefly written by Alan Kotok (1942-2006), assisted by John McCarthy (father of artificial intelligence) of MIT. The program ran on an IBM 7090, and was able to beat chess beginners. Kotok went on to become one of DEC’s leading computer designers (chief architect of the PDP-10), and created the first video game (Spacewar!) and the gaming joystick.

In 1965, McCarthy, who had been at Stanford University since 1962, visited the Soviet Union. There, a group at the Moscow Institute for Theoretical and Experimental Physics (ITEP), led by Alexander Kronrod, challenged his chess program to a match with their own, later called KAISSA. A match was held over nine months in 1966-67. The Kotok-McCarthy program lost the match 3-1. The Soviet chess program ran on an M-20 computer.



Now, this obituary http://www.guardian.co.uk/technology/20 ... n-mccarthy adds


McCarthy, (already in the seventies) became disillusioned about the direction taken by computer chess.
He said "Unfortunately, the competitive and commercial aspects of making computers play chess have taken precedence over using chess as a scientific domain...
It is as if the geneticists after 1910 had organized fruit fly races and concentrated their efforts on breeding fruit flies that could win these races."


Such a witty and relevant remark. I couldn't resist to post it here, for all the people of the Talkchess boards, nowadays so obsessed by the so called "clones" going commercial.
You see, those were already legitimate concerns 40 years ago.


After Ritchie last week, now McCarthy.
They were two giants and two gentlemen we'll sorely miss.


Requiescat in pace.

[rbarreira]

McCarthy, (already in the seventies) became disillusioned about the direction taken by computer chess.
He said "Unfortunately, the competitive and commercial aspects of making computers play chess have taken precedence over using chess as a scientific domain...
It is as if the geneticists after 1910 had organised fruit fly races and concentrated their efforts on breeding fruit flies that could win these races."


Such a witty and relevant remark. I couldn't resist to post it here, for all the people of the Talkchess boards, nowadays so obsessed by the so called "clones" going commercial.
You see, those were already legitimate concerns 40 years ago.

I don't think the comment is so witty, especially for someone as intelligent as McCarthy... That is unless bad analogies are supposed to be witty.

Chess is a game, how can one be disappointed at a game bringing out the "competitive" spirit.

[melajara]

For quite a long time Chess has been to AI what Drosophila melanogaster has been to biology, so I'm finding McCarthy's remark both witty and pertinent.

[rbarreira]

For quite a long time Chess has been to AI what Drosophila melanogaster has been to biology, so I'm finding McCarthy's remark both witty and pertinent.

But the competitive aspect in Chess AI (i.e. people trying to one up each other with better and better engines) did lead to quick improvements, and people did share ideas and data. The commercial aspect may not have done so, but then again why would one be surprised that things can have separate scientific and commercial aspects?

Of course one can complain that Chess AI did not lead to "true AI", but that's more a failure of over-ambitious expectations than a failure of Chess AI itself...

[melajara]

I think part of McCarthy grievance could have been to witness the preponderance of "brute force" over the Shannon-B-Strategie advocated in KAISSA.
When it was clear that computer chess could achieve GM level without resorting to an emulation of the thought process of a human player at chess
(pattern matching, chunking, long range plans, etc) Chess lost most of its appeal to AI.

[marcelk]

Today's programs are of type B.

[melajara]

Well, this is debatable.

IMHO, current programs are progressively shifting from A to B but without enough selectivity as for now to be true B.

[rbarreira]

I think part of McCarthy grievance could have been to witness the preponderance of "brute force" over the Shannon-B-Strategie advocated in KAISSA.
When it was clear that computer chess could achieve GM level without resorting to an emulation of the thought process of a human player at chess
(pattern matching, chunking, long range plans, etc) Chess lost most of its appeal to AI.

No matter how much pattern matching, selectivity and long range plans computers use to play chess, a vast number of people will always say computers are using "brute force" just because they can calculate millions of positions per second (never mind the fact that with a branching factor of 2 per ply it's certainly not possible for brute force to be taking place)...

The image of chess programs as "brute force" is permanently burned onto the public's brains, and apparently also on many programmers' (who of course should know better).

[marcelk]

I think part of McCarthy grievance could have been to witness the preponderance of "brute force" over the Shannon-B-Strategie advocated in KAISSA.
When it was clear that computer chess could achieve GM level without resorting to an emulation of the thought process of a human player at chess
(pattern matching, chunking, long range plans, etc) Chess lost most of its appeal to AI.

No matter how much pattern matching, selectivity and long range plans computers use to play chess, a vast number of people will always say computers are using "brute force" just because they can calculate millions of positions per second (never mind the fact that with a branching factor of 2 per ply it's certainly not possible for brute force to be taking place)...

The image of chess programs as "brute force" is permanently burned onto the public's brains, and apparently also on many programmers' (who of course should know better).

Humans use brute force. Computers have comparatively very limited capabilities and must therefore resort to other techniques.

[michiguel]

I think part of McCarthy grievance could have been to witness the preponderance of "brute force" over the Shannon-B-Strategie advocated in KAISSA.
When it was clear that computer chess could achieve GM level without resorting to an emulation of the thought process of a human player at chess
(pattern matching, chunking, long range plans, etc) Chess lost most of its appeal to AI.

No matter how much pattern matching, selectivity and long range plans computers use to play chess, a vast number of people will always say computers are using "brute force" just because they can calculate millions of positions per second (never mind the fact that with a branching factor of 2 per ply it's certainly not possible for brute force to be taking place)...

The image of chess programs as "brute force" is permanently burned onto the public's brains, and apparently also on many programmers' (who of course should know better).

Humans use brute force. Computers have comparatively very limited capabilities and must therefore resort to other techniques.

That is what Kotov wanted us to believe... but we don't.

Miguel