AI is a rather new science. The current models for doing anything (chess included) aren't very good. Sure, we can beat Kasparov (sp?) but what it comes down to is enumerating every possible move available and the responses to those moves, and so on. At a game like go, even a novice human player can beat the most advanced computer. A step in any direction (even a poor step) may open up new possibilities that are unavailable with the way we use/make AI now.
Slashdot Mirror
AI is a rather new science. The current models for doing anything (chess included) aren't very good. Sure, we can beat Kasparov (sp?) but what it comes down to is enumerating every possible move available and the responses to those moves, and so on. At a game like go, even a novice human player can beat the most advanced computer. A step in any direction (even a poor step) may open up new possibilities that are unavailable with the way we use/make AI now.