He boxes himself into one theory. And by not keeping his options open shows he's not intelligent. He does touch on a number of interesting topics. But I'm not sure I'd agree with any single point of his.
I'm more in line with the complexity theory people who say you want to be at the edge of chaos to bring about self-organizing principles. In a completely chaotic system, there just isn't going to be any intelligence. In a completely stable system, nothing will evolve.
In go or chess, with every move you give up some possibilities for others. And that's how I see a lot of his arguments. He's arguing for more possibilities but that's not really what you're doing. You're trading different possibilities with every move. In chess as white playing the Ruy Lopez, you'd welcome locking up the pawn structure on the queen's side reducing the possibilities there. In go, there is the concept of aji, taste or potential, and you don't want to needlessly eliminate the possibilities too early. Part of that is not making decisions on a small scale where a different decision might be needed for the larger picture later. But there's not some overriding keeping the possibilities open principle. If you can kill your opponent's large group immediately and win the game, you do it.
In living creatures, there are numerous necessary processes going on before what we'd consider intelligence takes place. An amoeba recognizes self vs other. In the brain, reward systems, the determination of what seems right or true takes place in the limbic system -- before the frontal cortex. Similarly, I'm sure his programs have a huge number of assumptions/preconditions built in.
Maybe more later. But time to go play racquetball now.
George |
