Interesting Video - Page 3

Jaged · 2007-02-24, 04:41 PM

Well obviously I was not claiming artificial intelligence truly does exist. I am merely pointing out that computers have a ton of "knowledge" as well. Exactly what it is that makes humans self aware is unknown to us. I do not think it is impossible for computers to eventually gain this attribute.

Baneblade · 2007-02-25, 04:59 AM

Commander Data seemed to be pretty accurately portrayed at what artificial intelligence is probably going to be like capability and limitation wise. But I do suppose knowing the definition of a word and knowing what it means are not exactly the same thing.

Rbstr · 2007-02-25, 02:23 PM

Originally Posted by Jaged

Well obviously I was not claiming artificial intelligence truly does exist. I am merely pointing out that computers have a ton of "knowledge" as well.

Computers don't have knowledge, they have indexed data. To have knowledge you need to be able to apply it and think about it.

Jaged · 2007-02-25, 04:09 PM

Originally Posted by Rbstr

Computers don't have knowledge, they have indexed data. To have knowledge you need to be able to apply it and think about it.

Yes... hence the quotation marks. What you are doing is propping up my argument into something it is not so you can more easily argue against it. You believe I am claiming that AI does exist today. This is not the case and I am not trying to say it is. I am merely saying that computers are on the track to having AI in the near future.

Look at it like this. We could easily program a computer today to mimic the actions of a goldfish. Perhaps we could even program it to replicate the actions of a dog, however it may be a bit of a stretch to get it 100***37; accurate with the computers we have today. Therefore you could conclude that the mental reasoning ability of computers today is somewhere in between that of a dog and a goldfish.

Once computers are able to think faster then a human brain can, who's to say that the difference between data and knowledge wont be breached?

While humans are programed by millions of years of evolution, computers are programmed by a few decades of human engineering. If you look at simple bacteria and other primitive life forms, it is hard to believe that they evolved into human beings. It is equally hard to believe that the computers of today will evolve into true AI. However, just like bacteria became humans, computers will eventually form AI.

Peacemaker · 2007-02-25, 05:15 PM

I imagine its not impossible in the future. I can see it happening.

Sentrosi · 2007-02-25, 06:01 PM

Originally Posted by Jaged

Well obviously I was not claiming artificial intelligence truly does exist. I am merely pointing out that computers have a ton of "knowledge" as well. Exactly what it is that makes humans self aware is unknown to us. I do not think it is impossible for computers to eventually gain this attribute.

Yeah, I'm a little late to this debate, so spare me some room.

Jaged makes a point of stating that "computers have a ton of "knowledge" as well." Well, some of you may remember these books called "Encyclopedias". Before the Internet and wiki's humans had to look up info inside these books. So, if I may extrapolate from Jaged's quote above, these encyclopedias should have become self aware then.

Also, the arguement for or against artificial intelligence is mute. The statement alone artificial intelligence is false. Either a thing is intelligent or not. Once a computer walks up to me and says, "I wish I could be more" I will say that that computer has become intelligent. Bettering your own environment is how humans have become intelligent.

Caveman days: "Brr, I am cold. I think if I kill that deer over there and rip it's fur from it I can become warm."

Throwing down some hard corp knowledge on you young pups. WERD! [flashing geek gang signs]

Jaged · 2007-02-26, 12:18 AM

Originally Posted by Sentrosi

Jaged makes a point of stating that "computers have a ton of "knowledge" as well." Well, some of you may remember these books called "Encyclopedias". Before the Internet and wiki's humans had to look up info inside these books. So, if I may extrapolate from Jaged's quote above, these encyclopedias should have become self aware then.

How many compute cycles do encyclopedias have? There is no way books could ever become self aware. There is no way they can get any smarter then they already are. Computers however are getting faster and faster every year. Eventually they will become self aware.

There is a difference.

Originally Posted by Sentrosi

Also, the arguement for or against artificial intelligence is mute. The statement alone artificial intelligence is false. Either a thing is intelligent or not. Once a computer walks up to me and says, "I wish I could be more" I will say that that computer has become intelligent. Bettering your own environment is how humans have become intelligent.

Now you are just arguing over semantics. Sure what you are saying is technically correct. Artificial intelligence is really just intelligence. However artificial intelligence is the term that we have agreed upon to refer to self aware machines. The English language does not always make complete logical sense, but thats just the way it is.

Btw, great thread. I love a good debate once in a while. I am calling for people to post more provocative material like this so that we can have more debates.

Electrofreak · 2007-02-26, 03:45 AM

You noobs, stop arguing about it and read the damn article I linked earlier.

THAT's the future of AI IMO. A computer that learns to create its own associations between stimuli and stored learned data (past stimuli), and can make decisions based upon those associations.

For example... you show a modern computer an orange, it may be able to use complex image-recognition software to identify it as an orange, but it doesn't know what an orange actually IS.

True AI will be this new software (going a ways out), where the computer learns by being shown many different oranges what an orange looks like, and can be demonstrated how an orange is peeled and consumed. This computer can potentially learn the orange is a consumable item. It could learn the same thing about an apple, and eventually be incorporated into a robot that, when asked by a human for something to eat, will identify and provide that person with a consumable item.

This is just the start... that article goes into some pretty awesome possibilities. How about a computer in your car that can tell the difference between pedestrians, other vehicles, animals, etc? Such a computer could be pre-loaded with behaviors of these objects, and will know best how to avoid hitting an animal versus a pedestrian or another vehicle. It'd also be able to watch the road for potential hazards like pot holes and debris, and most importantly actually understand the sort of threats they pose the vehicle (a blowing plastic bag is probably not going to require evasion for example). Modern computers can only achieve this sort of thing through very complex and touchy software.

http://money.cnn.com/magazines/busin...ion=2007020718

Hamma · 2007-02-26, 07:53 AM

As long as computers don't launch nuclear attacks on their enemies I am good.

Giovanni · 2007-02-26, 09:42 AM

Originally Posted by Hamma

As long as computers don't launch nuclear attacks on their enemies I am good.

I can just picture it.

"Girl computer voice": SHE DOESN'T LIKE MY SHOES?!? MY SHOES!!! LAUNCH MISSILES!

Hamma · 2007-02-26, 10:32 AM

What I fear is computers picking up the mentality of all the 8 year old forum drama whores.

Could you imagine?

Sentrosi · 2007-02-26, 03:12 PM

Oh dear god....if the collective forum intelligence is what computers are going to base their 'intelligence' on I'm just going to go figure out how to play WoW/LotRO with an abacus.

Rbstr · 2007-02-26, 04:58 PM

I read your article Electro. The technology is quite niffty. But thus far it hasn't shown to be acctualy be self aware; it learns. I think it could have by my definition knowledge. Now we're down to semantices over intelligence, I'd go as far to say it is showing signs of rudementary intelligence, but no kind of human-like awareness.

To me it just seems to be an algorithm that systematicaly creates a database, an amazingly complex database, but we don't know how that relates to our own system of storing information. Untill it's developed further and acctualy studied outside of Numenta's labs we can't really know.

Is our "Self" or intelligence just the sum of all things we know? Or is it something else?

I think one key thing with computers is that It will need to have to ability to program it's self. IE, you hook up a robotic arm, and it can figure out how to use it though nothing but experimentation.

Electrofreak · 2007-02-26, 08:41 PM

Yeah, without a doubt it has a way to go, but I think they're doing the right thing by making it open source, that means experts from all over can develop their own algorithms to expand upon the "intelligence" of the AI.

Will it ever be self-aware? Who knows. The guy who is behind it all seems to think that all we really use is similar algorithms within our brains... though many neurologists argue that he's taking a very simplistic approach to how our grey matter works, which may be true.

Time will tell.

Sentrosi · 2007-02-26, 11:01 PM

90% of our brain hasn't been mapped out completely yet.

Damnit! I want my super mind bending break your will powers NOW!