It is a kind of Turing machine issue, whether Man can beat AI by simply appearing immune to stupidity rays, the fate of humankind hanging in the balance.

Martin Gardner’s fascinating article on Michie’s invention appears in “The Colossal Book of Mathematics: Classic Puzzles, Paradoxes, and Problems,” and is called “A Matchbox Game-Learning Machine.”

It may be worth noting that, with programmable computers being widely available, the “matchbox” model described in Gardner’s article can be readily expanded to more complicated games. In other chapters of Gardner’s “Colossal Book,” he discusses several other non-trivial games that may lend themselves to the “matchbox” model.

Michie accidently took credit for a Human versus Machine chess theorem I’d invented in grad school (early 1970s) and published in SIGART. Accidently, because my theorem was transmitted to him via an International master with whom I played in Amherst, who then went to U Edinburgh as Chess advisor to Michie, but failed to pass on my contact information. Basically, I proved that if you KNOW that you’re playing against a machine that makes errors, your BEST move is not necessarily the optimal move that you would play against a machine that played “perfect” chess (which would take more matchboxes than the cosmos can probably hold).

To fill in some details: In Chess, however, I stated a
theorem, and gave examples, early in my CS grad
student days at UMass/Amherst, 1973-1977. I showed it to our resident International Master, Danny Kopec, who then went to be Chess Advisor to the AI programme at University of Edinburgh. SIGART had the Edinburgh professor.

I gave examples (horizon effect was shown in SIGART)
and classified situations with 1 error per game
players, 2 errors per game players, and suggested a
generalization with distributions.

Michie, sad to say, recently died in a tragic way. “Professor Donald Michie, 84, and his ex-wife, Dame Anne McLaren, 80, were in a car which left the motorway as they travelled from Cambridge to London. Prof Michie was a researcher in artificial intelligence who worked as part of the British code-breaking group at Bletchley Park during World War II.”

I iterate that I am not accusing the late prof. Michie of plagiarism; it was probably accidental on his part, because perhaps Danny Kopec did not have the right contact information for me. At the time, I was writing the world’s first PhD dissertation on Nanotechnology and Artificial Life, neither of which fields were yet
named.

When I was in elementary school, some children’s science magazine had a simplified version of this computer. They started the game a bit down the tree. I remember making up little boxes, each with a tictactoe configuration on it, and. as directed, putting in the appropriate colored M&Ms each indicating a possible move. Training the machine was great. When you won, you got to eat the M&Ms. Eventually, the machine learned how to beat you, and that was for M&Ms.

I had forgotten this gadget until recently reading Michie’s obituary. That may be most complex algorithm for eating M&Ms that I’ve come across, and as a side effect a game playing algorithm was optimized!

These rituals would seem entirely abstract, but if they were truly not invested with personal meaning, why bother?

An action can be deeply satisfying without having meaning. Making patterns with my M&Ms gives me that satisfaction, but not because it interacts with any higher intellectual process. I just like patterns.

All I do is eat groups of two to three with all different colors in the group – usually this requires including one of the most plentiful colors in all servings.

Unfortunately, I can’t do this while driving…

as for chocolates like kit-kats and peanut butter cups, does anyone else carefully nibble off the chocolate edges and tops before consuming the rest?

Ditto on the tan. I swear they tasted better. Kind of a mocha thing. I always saved them for last.

Elliot