DoorNumber1

Sayonara³: I'm trying to understand your stance on this issue. There was an original post that stated the opinion that human life is no different than bacteria and that kicked off some of the self-preservation argument. But does that mean you disagree with the stance that there's no real difference between human life and other forms or that you are just trying to ensure that anything said is factual?

I once was bored and grabbed a book called Quantum Evolution by a Brit named JohnJoe McFadden... I think this guy would benefit a lot from reading it. It will, by no means, answer all of this dude's questions (what the hell will!?) but it explains why there can be such things as directed evolution in a system that itself has no external direction and is playing by simple (well, comprehensible at least) rules. Sorry creationists, evolution occurs. Ever heard of penicillin resistant strains of bacteria? GRR Go back to school!

Yeah, that is a logical direction of an AI debate. But do we feel that what AI needs is more powerful computers or more powerful algorithms? I think that current AI research is in need of a jump on the algorithm side. Powerful computers can't do jack when the set of data they're supposed to handle (ie ALL data a potential intelligence comes into contact with) is so large. Take internet search engines: do you think that they work because there's a HUGE computer out there comparing your results against every page it's ever seen? Naw. Just too much data so do any such comparison. It took a really tricky, really clever algorithm that allowed altavista (and later google) to do the type of searching that's happening and handle that much info. I think we're asking a bit too much from our machines.

how did this find its way into this thread? haha (referring to in-depth discussions on quantum computing, etc)

Thank you! Finally a good point. You have to laugh when you look at how people view intelligence. Can you have an "intelligent" conversation with a baby? I have a 3 yr old niece who, I must say, is the cutest and smartest little kid I've ever seen in my entire life... but then again my opinion's biased! And you know, I can't have an "intelligent" conversation with her. I can talk about lots of things, but they're all sort of silly and superficial and they must relate, in some way, back to her. In 10 years though (probably a huge overestimate considering how intelligent her father is), when she learns more about the world, I'd be able to have that intelligent conversation people speak of. But I think I'll miss how things are now. What babies do when they come out, essentially, is spend all of their time just soaking up input from the world. They have certain needs, like eating (and the natural consequences of eating) and they take care of those first. Then they watch you and listen. They test you to see how you respond. They yell, they cry, they hit you and they bite you. They mimic you. And in that manner they learn from you. And, of course, they do this to inanimate objects and quickly learn the difference between you and, say, a wall (even though there's not much difference with some people). They speak because you're always talking to them. I'm pretty sure that if you looked at your baby every two minutes and said only "kitty" you can guess what his/her first word would be. So how do we create artificial intelligences? They'll have to be partially hardcoded. Any program is. And they need to want to learn. A baby is hardcoded, to some extent, when it pops out. It seeks human contact (hell, even dogs seek each other's warmth and touch when they're born... it's what social creatures do) and with no guarantee that the mother will provide for it. It puts itself in a position to have its basic needs taken care of and from which it can start soaking up human knowledge. A baby's cute and that makes us want to play with it. If we had that same desire to play with a young machine and we put even a fraction of the time into educating it that we do a human being I think we'd be shocked at what that little machine could do. After that it comes down to good learning heuristics and a good framework (that can update itself) into which it'll store and link it's learned knowledge. How to make that... good question! But it's a worthy challenge. I voted "I don't know" by the way. You can never be too certain about such things