Close to human AI possible?

[Carfax83]

All we've done so far is to make faster and faster calculators and come up with increasingly clever ways for them to retrieve and manipulate larger and larger amounts of hard coded data. It's so far removed from what a brain does, or seems to do, that I don't even know where to begin addressing your questions. We may _approximate_ a brain someday. We may even create a near-complete simulation of one (larger than a mouse, which has already been done I believe). But even when we do that it will still be light years from the conscious, self-aware, abstract reasoning thing that is inside our skulls. We don't even know enough to begin to guess how far off we are. We don't have the right words yet. Our memory doesn't act like storage. Our reasoning ability is not like a processor. Our senses do not act like input and output devices. It's been a mistake, since the very inception of the so-called AI movement, to think that computers were an infant stage in the development of a thinking thing. Not even close.

This kind of assumes the brain is even the source of Consciousness, something which as far as I know, is still up in the air because Consciousness has eluded the efforts of researchers to pinpoint it's origin for decades.

What gives us and other life forms self awareness and the capability to think and rationalize, is something which not even be physical in nature, but immaterial.

Humans create information on a daily basis, something which is by nature non physical, yet undoubtedly has very real effects. The sentences I am writing now has specific meaning which cannot be reduced to physical objective properties. How is it a physical organ like the brain, can create and understand abstract non physical concepts?

 

[BoberFett]

All we've done so far is to make faster and faster calculators and come up with increasingly clever ways for them to retrieve and manipulate larger and larger amounts of hard coded data. It's so far removed from what a brain does, or seems to do, that I don't even know where to begin addressing your questions. We may _approximate_ a brain someday. We may even create a near-complete simulation of one (larger than a mouse, which has already been done I believe). But even when we do that it will still be light years from the conscious, self-aware, abstract reasoning thing that is inside our skulls. We don't even know enough to begin to guess how far off we are. We don't have the right words yet. Our memory doesn't act like storage. Our reasoning ability is not like a processor. Our senses do not act like input and output devices. It's been a mistake, since the very inception of the so-called AI movement, to think that computers were an infant stage in the development of a thinking thing. Not even close.

Humans are animals, just as mice are. We may be vastly more complex than a mouse, but it's still stimulus/response. AI is far more of a question directed at biology/chemistry and understanding the inner workings of the brain than it is a computer question. Once we understand the brain, recreating one with a computer will probably be a fairly simple matter.

 

[nightspydk]

I think morale is the obstacle. Where do we get that taught? Maybe, but I think it's also in the genes. I'm a prime example of that left to my devices and I got a conscious. From where I wonder. PPl that lost the touch with other people, cannot connect, got other problems. You cannot ever program that.

So much our genes dictate not just hair color, so why not out minds. I'm positive that's where we originate. I've been told I'm like my father and he's dead since I was an infant. What does that tell you?

 

[beginner99]

Humans are ran by wanting this natural high of doing something exciting and having fun. If it doesn't actually feel a sensation it can't strive for more of this or even understand why humans like what it has programmed as happiness at all. For the purpose of this program I suppose it doesn't actually need to feel emotions and could just have emotions be another number, and the program wants to raise the number (i mean be happier or excited) but still has to take into consideration of other factors like other's safety. It does seem silly that you could program actual emotions.

define emotion?

If the robot wants to play football and is happier after it how can you say it does not feel anything?
With happier I mean smiles more, more relaxed...mimic what humans do when they are happier.

 

[kolop97]

define emotion?

If the robot wants to play football and is happier after it how can you say it does not feel anything?
With happier I mean smiles more, more relaxed...mimic what humans do when they are happier.

It may look happy and act happy, and for all purposes be happy, but it still wouldn't feel what we think of happiness, so while it would be happy, it wouldn't be the same as a human being happy. It doesn't really matter to much that it won't feel the same thing since it will still act the same. Do you see what i mean?

 

[KentState]

define emotion?

If the robot wants to play football and is happier after it how can you say it does not feel anything?
With happier I mean smiles more, more relaxed...mimic what humans do when they are happier.

Happiness is a human state used to comprehend if an activity is beneficial or not for us. For a binary machine, being happy is not necessary for AI to exists. What we may confuse as intelligence in humans is there because we are severely lacking with communication and reasoning.

For example, if using intelligence to vote on the "best" computer to lead them all, it would come down to logically picking the one with the best benefit. That's nothing more than statistics and probabilities, though it would be very complex. Humans on the other hand hold elections, use propaganda, cry, kiss babies, use humor, slander and everything else. The questions is, in an AI system, are we trying to create intelligence or have computers act as dumb as us?

 

[beginner99]

It may look happy and act happy, and for all purposes be happy, but it still wouldn't feel what we think of happiness, so while it would be happy, it wouldn't be the same as a human being happy. It doesn't really matter to much that it won't feel the same thing since it will still act the same. Do you see what i mean?

And how do you know that?

No complete topic switch:

One of the links in here made me find cleverbot.com. It seems this bot was able to fool almost 60% of the people to being a human.

I would once really like to participate in such a test. i think I had no trouble identifying a bot but maybe because I know it is one, it's always "easy". In case of cleverbot in the 4th sentence I wrote I made an accidental typo, liek instead of like and the replay from the bot was completely nonsensical. And at that point I was not even trying to trick him.

 

[colonelciller]

Not possible with current storage technology.

Just a simple image recognition, comparison, and memory recall can pull up several gigabytes of random access storage from long term memory near instantly in a brain. It may take a while to remember the primary key some times but when you finally do, all associated memory with that prompt is recalled near instantly , every vivid detail in IRL resolutions in multiple dimensions.

Do that on a hard drive and you are talking kilobytes per second...

We will be severely limited in our computer capabilities for a long time until we develop a fast high density non volatile universal main memory that is actually as fast or faster than the CPU.

Until then forget about it. We can have all the CPU power in the world but our primitive storage and network speeds are a joke. We need RAM that keeps data indefinitely until changed that can do 50+ GB/sec with nanosecond access times before actual human capable neural net AI is possible.

memories a Nowhere near equivalent to the storage of photographs. we already have plenty of capacity and waaaaaaaaaaaayyyyyy more than enough speed. SSD's are MUCH faster than data recall from the brain, and brains do not store pixel by pixel images... this is not how memory works.

the thing that is lacking is a complete understanding of how the connections between the various nuclei in the brain are interconnected on a cell layer by cell layer basis (and from there accurate models of the interconnections on a cell by cell basis within those layers), and how those interconnections interact to produce the thing wecommonly referto as consciousness. The neuroscience is still a LONG way off.

the hardware is already there.
the neuroscience is not there.
once the neuroscience is there it will become a matter of simple of software programming.

 

[colonelciller]

This kind of assumes the brain is even the source of Consciousness, something which as far as I know, is still up in the air because Consciousness has eluded the efforts of researchers to pinpoint it's origin for decades.

watch this
http://www.youtube.com/watch?v=k_ZTNmkIiBc

 

[BoberFett]

the hardware is already there.
the neuroscience is not there.
once the neuroscience is there it will become a matter of simple of software programming.

That's pretty much what I said.