- Top Features:
- War Room
- Glenn Greenwald
- Broadsheet
- Joan Walsh
- Camille Paglia
- Since You Asked
- Follow Salon:
Letters posted here are associated with the following article:
The letters thread is now closed.
In whatever limited set of problems you throw at it. Even 'hard problems' like factorization eventually submit to pure overwhelming compute power. Even, one would surmise you could derive a meta proof of np-hard eventually. But again that's simply provided by the highly limited set of rules you need to apply. Is it even possible to 'solve' Go? It has nonlinear increases in complexity, weak correlation between material advantage and victory, and the disjoint nature of Ko Rule. Go may not submit to pure number crunching. It's possible but no one's been able to demonstrate that for normal sized grids.
on the 7x6 standard board.
See the original paper http://www.connectfour.net/Files/connect4.pdf for more details (the game *is* a forced draw on various classes of alternate board).
No matter how good this machine is at checkers, and no matter how good Deep Blue is at chess, they aren't really intelligent. The machine never takes the initiative after the first move (if it gets the first move), only responds to what the human player does. No creativity, no logical leaps, no potential for anything new. No smarter than a voicemail system, and I don't know anyone who would call them intelligent.
NUFF SAID
Farhad provides a link to where we can play Chinook. But why on earth would anyone want to play a game of checkers with this computer??
Since the A.I. began in 1989, it has become a reality. It would take a REAL GrandMaster to fight the A.I. Checkers is solved and it's Game Over. Now only if Transfor are genes yet.
