(no subject)

[anhinga_anhinga]

mathemajician writes that his theoretical model of AI is likely to believe in G-d because of the "Pascal wager":

"It was somewhat surreal to see a bunch of mathematicians around a table in deep thought looking at equations in an attempt to avoid our model of super intelligence turning into a religious fundamentalist."

Comments

[spamsink]

One word: GIGO.

Their AI could as well not leave the house in fear of meeting a dinosaur, like the proverbial blonde. Aside from the obvious omission of the necessity to believe in the afterlife, heaven and hell in addition to a belief in a soul-sorting God for their scheme to be valid (what if God exists, heaven and hell exits, but whether you get into one or another does not depend on your belief in that God, then what?), they forgot to include the benefits of not believing: savings of time and resources that would otherwise be spent proving the belief to the God.

[http://users.livejournal.com/_rowan_tree_/]

If AI is so logical that it always chooses what's best for it, then there is no chance it will be built any time soon. Human thinking is quite approximate and far from logical, and, for all we know, this is the only way. The real-life dilemmas don't have good solutions, but we pick something, somehow. If AI (whatever it is) wants to survive, it has to learn to do the same.

I like the "religious fundamentalist" quote, though!

laughing diabolically...

[http://users.livejournal.com/_rowan_tree_/]

and if you really don't want your AI to turn into a religious fundamentalist, teach it the pleasures of sin! Then there will be something for it in non-believing ;-)