I haven't blogged for a while, I've just got back from holidays.
I have been reading HR Ekbias Artificial Dreams.
Its been a thought provoking book. Pointing out the problems with old Boolean AI, and also the lower end approaches like subsumption architecture and neural nets. A very good survey of AI over the years, as good as Daniel Creviers AI from '93.
You know I should add one of the things I got out of this book was the concept of a free ride. For any data structure, pattern matching method or AI technique to be successful out of specific boundaries it is clearly going to need a structure that will give it a "free ride" so it can generalize.