Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Rixie comments on Explaining vs. Explaining Away - Less Wrong

46 Post author: Eliezer_Yudkowsky 17 March 2008 01:59AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (98)

Sort By: Old

You are viewing a single comment's thread. Show more comments above.

Comment author: Rixie 02 April 2013 05:34:03PM -2 points [-]

Your chess playing software must make the decision that is most likely to win the game, wheras humans don't have anything to stop us making the bad decision.

Comment author: NancyLebovitz 02 April 2013 07:28:40PM 4 points [-]

Your chess playing software must make the decision which is most likely to win the game according to some algorithm (and assuming no computer glitches). Humans have plentiful reasons to make mistakes of kinds that computers don't, but that doesn't mean computers making the best possible moves.

Comment author: Qiaochu_Yuan 02 April 2013 08:09:13PM *  6 points [-]

Chess playing software runs an algorithm designed to play chess. It may be good at playing chess, but it probably isn't optimal: remember that until fairly recently top grandmasters could still beat top chess software. Humans run another algorithm designed to pass on genes. It may be good at passing on genes, but it probably isn't optimal; remember that evolutions are stupid.

Moreover, the algorithm that governs humans behavior is no longer working in the environment in which it evolved, whereas chess playing software has the benefit of only needing to work in the environment for which it was designed. Ask chess playing software to play checkers and you'll get nonsense.

Comment author: OrphanWilde 02 April 2013 08:44:28PM -1 points [-]

I'd be surprised if a chess program weren't easily re-adapted to playing Checkers just by adding rules for the pieces; checkers even has a similar "transformation" rule as pawns in Chess, whereby pieces which reach the opposing side of the board can turn into pieces with different abilities. Backgammon, on the other hand...

Comment author: JGWeissman 02 April 2013 09:02:09PM 3 points [-]

I'd be surprised if a chess program weren't easily re-adapted to playing Checkers

The hard parts of make chess and checkers AI would not translate well, like evaluating the strength of a position, and strategies for pruning the search tree.