Generally I agree with this and your other post. AI has all the hallmarks of a degenerate research programme. But I did want to raise one point... ;-)

Eh, "Artificial Intelligence"? ;-) Making complex decisions as fast and in a similar way as humans can?

This combines what I would consider the two primary objectives of AI in a way that makes them sound like one.

1. AI as an investigative tool for a plausable model of cognition. Much work was invested into AI as a method to explain how we think we think. This research tended to be very domain specific and scaled poorly. It also had fundamental philosophical problems that put to lie many of the grandisoe claims that were made by it.
2. AI as an approach to solve problems that computers are typically bad at and humans are typically good at. This was a slighly more fruitful avenue. Usually the models bore little if any relationship to how we think we think. Examples of this are the AL work done in emergent phenomenon, GA's, neural networks.

The former appears to be almost totally abandonded. (Although high quality chess programs are perhaps one of its successes.) The latter is being used in a small number of unlikely areas with considerable success. (Chip design is a notable example)

Yves / DeMerphq
---
Writing a good benchmark isnt as easy as it might look.