I noticed much if the discussion seems to be missing that point. Some object that the machine driven Google isn't intelligent because it can't come up with novel ideas. But, why would we want it to?
There doesn't seem to be much reason to make an intelligent machine that capable of the full range of intelligences that humans display, other than as a novelty just to prove we can. The closest thing would be a personal assistant program, that would have to understand human speech, be able to manage a large set of information, and to make inferences that allow it to make timely and appropriate suggestions. Even then, suck a program wouldn't need the full range of human intelligences. It wouldn't need to worry about finding a mate or ponder the meaning of life, for instance.
Of course, the augmenting existing minds would be a much more useful application of AI than creating new autonomous AIs. My guess is that is the more likely route that things will take.
An interesting point. How often does your average human "move beyond his original programming?"
We can see the deficiencies in Data's AI because we're comparing his ability to that of a Human, and find it lacking.
We do the same to today's AI: we focus on what it can't do, even though what it can do today would be considered "intelligent" in the eyes of researchers 50 years ago.
There's no reason to think people won't continue to insist that a computer can't think until they can do everything a human can, and then perhaps the computers who tells the humans that they can't "think."
I saw a video once where someone said "Smart people are good at rationalizing things that they began to believe for not-smart reasons." As someone who was once a "smart," deeply religious person, I think this is true.
Slashdot Mirror
I noticed much if the discussion seems to be missing that point. Some object that the machine driven Google isn't intelligent because it can't come up with novel ideas. But, why would we want it to? There doesn't seem to be much reason to make an intelligent machine that capable of the full range of intelligences that humans display, other than as a novelty just to prove we can. The closest thing would be a personal assistant program, that would have to understand human speech, be able to manage a large set of information, and to make inferences that allow it to make timely and appropriate suggestions. Even then, suck a program wouldn't need the full range of human intelligences. It wouldn't need to worry about finding a mate or ponder the meaning of life, for instance. Of course, the augmenting existing minds would be a much more useful application of AI than creating new autonomous AIs. My guess is that is the more likely route that things will take.
That and the fact that humans would make a pretty crappy battery.
An interesting point. How often does your average human "move beyond his original programming?" We can see the deficiencies in Data's AI because we're comparing his ability to that of a Human, and find it lacking. We do the same to today's AI: we focus on what it can't do, even though what it can do today would be considered "intelligent" in the eyes of researchers 50 years ago. There's no reason to think people won't continue to insist that a computer can't think until they can do everything a human can, and then perhaps the computers who tells the humans that they can't "think."
I saw a video once where someone said "Smart people are good at rationalizing things that they began to believe for not-smart reasons." As someone who was once a "smart," deeply religious person, I think this is true.