Okay.. It’s been a bit of a while.. which happens when a semester gears up and I get everything organized for it..
However, this morning a friend of mine from way back when forwarded me this xkcd comic and linked it to the previous NOM’s I have up here about Artificial Intelligence, Meat, Machines, etc…
The cleverbots dialogue that they are talking about, I think, is this one… which is totally hilarious (and also intriguing how it gets into religious talk so fast.. but I digress and that is a discussion for another time..)
Significantly, my mind related this comic almost instantly to a previous one by xkcd a couple of months ago that was about Watson that I have up on the door to my office…
Now.. what is interesting to me about both of these is the way in which humans capacities and machine capacities are laid pretty bare in these. While it is true that the first comic took the whole idea in a different direction (namely, that there is no difference between a robot and an attendee to burning man, since they can both play chess, drive cars across the desert and cannot hold normal conversations for 5 minutes..), what struck me is the the place where the machine fails–namely in holding a simple conversation for 5 minutes–seems so much easier than either playing chess or driving a car–and yet it is something that machines regularly fail at. If you go to Cleverbot.com and just talk to it as if it were a person–it fails markedly–sometimes in the first response–quite quickly..
Why I think this is important is that it highlights a huge difference between how our biological systems work vs. how even the more complicated AI systems are set up. Specifically, our systems are open analog systems that are constantly integrating new data as we process it–and are rewiring ourselves constantly to deal with a continually changing environment. In comparison–the model for computers is much more of a closed system that does not rewire itself but merely keep accumulating a greater and greater amount of info–and tries to process it quickly–but doesn’t really adapt to changing circumstances that well. Really–computers are meant to compute and the act of computing assumes that you have clear entities to operate upon that fit into the specified criteria. In math, for example, you can add numbers together–but not a number and the color orange.. because orange isn’t a number.
Reality is much more like adding a number and a color together, however, because it is not a closed system in nearly the way that mathematics is… and this is one of the reasons why these computers are still so easy to pick apart and thwart.. because they don’t really think.. they don’t really “play” games–all they do is follow rules… (wow, doesn’t that sound like the Matrix..)–but they cannot make up the rules themselves and/or create rules and/or just ignore them when it seems appropriate..
Anyway.. just a follow-up as I noted.. Nothing earth-shattering.. just a thought to get down..