Okay folks, I know I started the whole debate that seems to have gone slightly awry, but I'm back here after reading everyone's arguments..
And in the end after reviewing everyone's arguments, I still don't buy the premise that a "bottom-up AI" is 1) necessary for what they're trying to achieve and 2) would be what the military would be looking for - I remain unconvinced that stubborn military men would put their fate in potentially unreliable weapons.
In fact like in the real world we all live in, "a suitably complex top-down AI" would be much more effective, and achieving "true AI" is more of an academic exercise/desire. At the end of the day, of you're convinced, Great! -- but I feel a bit sad that Reki hasn't gotten me to suspend my disbelief, when he otherwise is a very convincing writer..
|