@clacke @drwho When I was doing my undergraduate degree a decade ago, “Artificial Intelligence” was one of the courses I took. Among the things taught were:
- Pathfinding Algorithms, A*
- Bayesian Statistics
- Markov Chains
That was years before Deep Dream was even a thing.
Artificial Intelligence is something that behaves as if it were intelligent, but that intelligence is itself artificial. Not as in man-made, but as in the behavior was the intended result - even if the intelligence bit has always been smoke and mirrors.
A machine doesn’t *need* to be actually intelligent to solve a maze. It just needs to behave in a way where the side effect is that it appears to have agency. Nobody thought we were building sentient software, just algorithms and techniques that created desired behavior.
(Sorry to get on a soapbox here, I just intend to agree with you with this post.)