/sci/ - Science & Math

>>9773478
Charles Babbage was asked a similar question about his Analytic Engine.
It's advantage was that it wouldn't make mistakes. You pun in the correct numbers and you get the correct answer.

Someone asked if it would still produce the correct answer if you entered the WRONG numbers.
I don't know if history recorded Babbage's reply, but I'm pretty sure I know what he THOUGHT of the question -- and the person who asked it.

What you're asking for is General Intelligence; a machine which only needs a problem explained to it and then works out the methodology on its own.
It's not "what's the square roof of 135435?" It can, presumably, find an algorithm for that. It's more like "Prove the Riemann Hypothesis!"
If WE don't know how to do something (or we do, but don't tell the computer) it's at a loss. "Creativity", "Insight", and "Understanding" haven't been reduced to procedures yet. Think of the pointy-haired boss in Dilbert as an example of someone who doesn't "get it".

Depends on whether you believe programmers are just following a rule-set and no originality is required. Plug and chug.

So far, no program really "understands" anything -- not even neural nets which WE don't always understand. They've "learned" to play Chess and Go by trial-and-error. They're just a lot faster and have better memories than we do.

Maybe human thought IS just a matter of following rules. That's the Chinese Room hypothesis.

>>9681071
Just a matter of time.

>>9649985
Enjoy the high life while you can.
Save so you don't wind up like the superstar athletes who go broke when the gravy train ends.

Comfy job?
Not for long.