>2) Don't try to implement full self-awareness. The various domdules need
>be able to interface with each other, but we don't need to create one for
>'thinking about thought'.
I think you have very good ideas. Do you assume that self-awareness does not spontaneously emerge from a sufficiently complex adaptive neuro-system? If we *can* implement full self-awareness in an artificial intelligence, why not implement it in semi-conscious fundamentalists as well? But if fundies don't have sufficient intelligence to respond to our implementation, and if the SI does have adequate rationality to do so, how can we avoid implementing full self-awareness in the SI? I mean, if the SI can outperform the brightest red neck intellectually, how can it not attain self-awareness (autonomously) equal to or greater than that of the red neck? Conversely, if we can implement full self-awareness in the SI, why not do so for the red necks of the world first? -zen