← All games

The people

Who proved a game, who built the machine, and who was the champion

Machines do not decide to play. Somebody chose the game, somebody wrote the rule the program follows, and somebody sat on the other side of the board and lost — or did not. This page is the people the rest of the site keeps naming.

They fall into three rough kinds, and mixing them up misreads the history: the ones who proved something about a game with a pencil, the ones who built a machine to play it, and the ones who were beaten by that machine at the top of their own field. The third kind is the smallest and the one the story usually flattens.

Alan Turing
1912–1954. He defined what computability means and worked on ciphers at Bletchley Park. He treated chess as a convenient test for machine thinking and wrote a chess program years before any machine could run it. Convicted for homosexuality, he died at forty-one.
Hodges, Alan Turing: The Enigma, 1983
Arthur Samuel
1901–1990, an IBM engineer. He wrote a checkers program that corrected its own positional evaluation from the games it played, and put the phrase "machine learning" into circulation. He worked on it for over twenty years and regarded checkers as merely a convenient excuse.
Samuel, IBM Journal of Research and Development, 1959
Charles Bouton
An American mathematician who published the complete theory of Nim in 1901. You win by leaving your opponent a position whose heap sizes have a bitwise XOR of zero. No search and no evaluation: one formula closes the whole game, any number of heaps, any size. The first solved game in this list.
Bouton, Annals of Mathematics, 1901
Claude Shannon
1916–2001. He founded information theory and, in passing, wrote the first paper on how a computer ought to play chess and built a machine at Bell Labs that guessed human choices — a simplified version of Hagelbarger's. He treated games as a convenient testbed: clear rules, non-trivial problem.
Shannon, Bell System Technical Journal, 1948; Shannon, Philosophical Magazine, 1950
David Silver
Born 1976, he led the AlphaGo and AlphaZero line at DeepMind. He argued that human games do not help but hinder: the program trained purely on itself overtook the one fed borrowed experience. The claim was tested on three different games in succession.
Silver et al., Nature, 2017; Silver et al., Science, 2018
Demis Hassabis
Born 1976. A chess master as a teenager, then a games designer, then a neuroscientist with a doctorate on memory. He founded DeepMind and chose games as problems where success is measurable and cannot be argued about.
Silver et al., Nature, 2016
Feng-hsiung Hsu
Born 1959 in Taiwan. He began building chess chips as a graduate student at Carnegie Mellon, carried the line from Deep Thought to Deep Blue and remained its chief designer. He believed specialised hardware would win rather than general-purpose, and in 1997 he was right.
Hsu, Behind Deep Blue, 2002
Garry Kasparov
Born 1963, world chess champion from 1985 to 2000. In 1996 he won a match against Deep Blue and in 1997 he lost to it — the first world champion beaten by a machine in a match. He spent the years after arguing that a human paired with a program beats either side alone.
Kasparov, Deep Thinking, 2017
Gerald Tesauro
Author of TD-Gammon, a program that learned backgammon almost entirely from games against itself, years before that approach became routine. It changed human theory: several opening plays long considered weak entered the standard repertoire after its games were studied.
Tesauro, Communications of the ACM, 1995
Hans Berliner
1929–2017. World correspondence chess champion, later the author of the programs BKG and HiTech. In 1979 his BKG 9.8 beat backgammon world champion Luigi Villa 7-1, the first reigning champion to lose to a machine. Berliner said himself that the dice ran the program's way.
Berliner, Artificial Intelligence, 1980
Jonathan Schaeffer
A Canadian computer scientist who began the Chinook project in 1989 and closed it in 2007. Between those dates: a world title handed to a program by an unfinished match, Tinsley's death, and eighteen years of uninterrupted computation for one line of result.
Schaeffer et al., Science, 2007
Ken Thompson
Born 1943, the author of Unix and the B language. In chess he did two things: with Joe Condon he built Belle, the first machine to reach a master rating, and computed the first endgame tablebases backwards from mate. Some of the wins he found could not be reproduced by grandmasters.
Condon & Thompson, Belle Chess Hardware, 1982
Marion Tinsley
1927–1995, mathematician and the strongest checkers player who ever lived. In forty-five years of tournament play he lost about seven games, two of them to Chinook. In 1994 he withdrew from the match on health grounds and died months later. The match was never finished.
Schaeffer, One Jump Ahead, 1997
Murray Campbell
Born 1957, a Canadian IBM researcher and a strong chess player. On the Deep Blue team he owned the evaluation function — the part where the machine decides whether a position is good. He was also the one who later explained that the strange move in game one had been a bug.
Campbell, Hoane & Hsu, Artificial Intelligence, 2002; Silver, The Signal and the Noise, 2012

the same names in date order →

Where to go from here
To cite this page: “The people”, You vs Computer, https://youvscomputer.com/games/people.html