Right. There is an aspect of the Cargo Cult to the push for machine intelligence. As long as the function of the human brain is not well (or perhaps not at all) understood, then trying to create a computer that will be as intelligent as a human is really not much different from trying to cast a spell. It is an exersize in magic.

I wish Allen Turing had expanded somewhat on his test, which seem to me to be less than satisfactory as a test for intelligence. If one does produce a computer that can string together snippets and produce rational-seeming conservation, is this trully intelligence on the part of the computer? I would say it is no more so than the mock runways and control towers constructed by New Guineaians attempting to lure down cargo-carrying aircraft.

 

I just wasted a little time interfacing with ALICE and Jabberwoky's "Joan." Turing may or may not have been onto something, but Loebner is clearly a nutcase. (Um... waitaminute... make that "eccentric." Rich nutcases are "eccentric.")

I still think emulation of Physical Intellegence may be possible. We might get something smart enough to hunt vermin or drive a car. I'd be happy with a machine that could sort my laundry. For conversation, I think I'll stick with good ol' Homo sapiens sapiens.

 

My point is that our brains can be emulated by computers. Our brains are nothing more than processors computing inputs, running on programming that is constantly changing and adapting. It's emulating our ability to process, learn, and interact in a certain manner that will be the challenge for programmers. It can be done, and nature has proven that to us over the process of billions of years of random mutations and natural selection(a very slow yet reliable process). It seems as though there is this idea out there that we run on something magical in our heads that cannot be duplicated. Given enough time, mankind can achieve AI, even if it is not exactly identicle to the way our brains function.

 

My point is that our brains can be emulated by computers.

Which part? The one that makes us feel shame when we realise we've done wrong? The one that makes us feel proud at the accomplishment of offspring, apprentice, or student? The so-called "God center" believed by atheists to be responsible for religious experience? (And belived by the faithful to have been put there by the Divine?) The part that feels frustration, remorse, love, satisfaction, anxiety, amusement, or any other of the feelings we feel?

The part which can imaging, if only clumsiliy, what it might feel like to swim in a giant bowl of gelatin?

Software does exactly and only what programmers tell it to do. (If they inadvertantly tell it to screw up, it faithfully screws up.) How will they tell it to take pride? How will they tell it to be compassionate? How will they tell it to be creative?

Is the brain really and truly just a pile of cells, or is it some kind of "radio for the soul?" (Assuming, for a moment, that there might be such things as souls...) If our progeny does create an artificial brain, will it also be a conduit for a soul - and will that sould be artificial or will it come from wherever "natural" souls come from? And if there are no souls at all, we're still back at the same place... how can anyone claim to have created "artificial intellegence" without first defining "intellegence?"

My wife, who shrinks heads for a living, assures me that the definition of "intellegence" is very much a work in progress.

 

I would seriously challenge your assertion "My point is that our brains can be emulated by computers." No understanding of the brain's operation exists yet. How do we emulate the unknown?

Consider the gestalt function. That is where you take a brief look at a situation and instantly (this is in human terms) comprehend everything about it. Even with many parallel processors, try to imagine the SELECT CASE statement that would have to execute to evaluate a situation. And then you have to do the consensus coding, so they all "agree". Computers are numeric engines - machines that exist to run on numbers. Being able to ADD very fast and do a branch based on the result is simply not a brain function. Computers do not clock data at differing rates that indicate a level of interest.

Remember expert systems? about 20 years ago, it was a certainty that a computer could have in a database every salient fact about some human occupation, and so could make perfect decisions. Very few, if any successes. It turns out that even a closed system has always a few more degrees of freedom than a set of rules can encompass.

Currently, my estimation of software is that it is becoming less able to do a task. Microsoft is my poster child for that assertion. They have been in the business of writing code for operating systems for over 20 years. So far, a definition of "operating system" has yet to be agreed on. Windows XP is orders of magnitude larger than Windows 95, but about all one can say is that it is somewhat more stable. In terms of acting as an applications launcher, I had a DOS menu program that worked much more predictably.

 

Which part? The one that makes us feel shame when we realise we've done wrong? The one that makes us feel proud at the accomplishment of offspring, apprentice, or student? The so-called "God center" believed by atheists to be responsible for religious experience? (And belived by the faithful to have been put there by the Divine?) The part that feels frustration, remorse, love, satisfaction, anxiety, amusement, or any other of the feelings we feel?

The part which can imaging, if only clumsiliy, what it might feel like to swim in a giant bowl of gelatin?

Software does exactly and only what programmers tell it to do. (If they inadvertantly tell it to screw up, it faithfully screws up.) How will they tell it to take pride? How will they tell it to be compassionate? How will they tell it to be creative?

Is the brain really and truly just a pile of cells, or is it some kind of "radio for the soul?" (Assuming, for a moment, that there might be such things as souls...) If our progeny does create an artificial brain, will it also be a conduit for a soul - and will that sould be artificial or will it come from wherever "natural" souls come from? And if there are no souls at all, we're still back at the same place... how can anyone claim to have created "artificial intellegence" without first defining "intellegence?"

My wife, who shrinks heads for a living, assures me that the definition of "intellegence" is very much a work in progress.

If I weren't to account for the possibility of supernatural interference(which would invalidate this entire discussion anyways), then you should know that the whole universe acts as nothing more than a sequence of events. The emotions we feel, how we think, etc, are all part of a programming thats based on how to react to certain inputs, one of those being how we are programmed to socialize and express "emotions". Its not some magical thing in there, its just several processors working together executing an ever-changing and adapting programming. It is through the logic that our brains are programmed to do what it does and that it has been replicated billions of times that it is very possible for technology to advance to a point where we can understand how the brain works and replicate it's ability to function intelligently. Unless you have a good reason to dismiss the near infinite technologies that may potentially exist in the future, the arguement that we'll never understand the brain is not very founded.

 

Let's assume for the moment that a human personality is just a set of instructions running on wetware. Is that instruction set actually complex enough to decipher it's own code, and to program an external emulation of itself? (We've no proof either way, yet.) If so, would't the created personality be also be able to create new synthetic personalities? Heck of a threshold, eh?



I was in Powell's today (I love going to Powell's) and found this: http://www.powells.com/biblio/4-9780262692939-0
Looks like a lot of folk consider the idea of "intellegence" to be a philisophical issue rather than a technical one. Turing's idea was based on the notion "people talk and animals don't." Since Turing's day, we've learned that animals do communicate in various ways. So... if a computer can convice a bird that said bird is arguing with another bird about territory, can we say the softare is "artificially bird-brained?"

 

There is no law stating that you cannot create something above your own intelligence, and I personally find such a theory to be quite baseless(we can make things stronger than us, things that can process faster than us at certain things, why not something that can think on a higher level than us?). I personally believe that AI will arrive through artificial evolution(outside of nature). We create an intelligence that is able to slowly adapt and change on its own to exponentially scale its intelligence forward. If mere chance can produce what we are today, I'm sure mankind can produce something such as hard AI given enough time, even if that takes thousands of years, or even longer.

And as far as the philosophy of AI, all I see as far as truely intellectual AI is an intelligence comparable to ours. Its all based on relevance more than anything. And yes, I hate saying intellectual AI

 

There is no law stating that you cannot create something above your own intelligence, and I personally find such a theory to be quite baseless(we can make things stronger than us, things that can process faster than us at certain things, why not something that can think on a higher level than us?).

But the parameters of strength and speed are readily quantifiable - intelligence is not. That is because intelligence is something that resides in human nature - something which is very human by its function. And the idea of creating something that is more intelligence by a process of instructions in software is not IMO feasible.

Think about the computer that can beat the world chess champion - it wins because it iterates very quickly through every combination and deduces what it deams to be the best given the current state of play, but if the room in which they were playing set on fire, would the computer walk out and save itself? Its sounds silly, but the machine is intelligent in one sense but can only function around its programmed remit of operation. So I ask what is intelligence?

Dave

 

Computers are pretty spiffy for complex equations that humans have trouble with. Computers suck for simple things that humans find easy. Pattern recognition is an example of "easy for human, hard for computer." Copious code is needed to distinguish between a chair and a cat. I have a buddy who's dissertation explored pattern recognition by software. He and several of his associates have since spent the better part of a decade trying to come up with software to distinguish between "porn" and "not porn."

 

Hence the need for dramatically more time required to achieve AI. Whos to say we won't eventually understand the entire workings of the brain. With technology growing and understanding and research growing, I just can't find it possible to hit a permament roadblock. The best answer I hear is "its hard to do this", but just because something is very hard doesn't mean its impossible.

 

There is a famous article at Wired.com by Bill Joy, one of the founders of Sun Microsystems, called "Why the future doesn't need us?" looking at the potential impact of robotics, gene engineering and nanotechnology in the 21st Century. Makes interesting reading from one of the top people in the technology field:

http://www.wired.com/wired/archive/8.04/joy_pr.html

Dave

 

Pleeeeaaaase! Don't bring Windows to this conversation. Windows does not have anything to do with intelligence. LOL

 

well can a computer be a jerk at times ?
from what i know computers or logic ckts cant even generate a random number(they do it with some program) on their own let alone think of a random thought or theory which was not even thought of before so it certainly will have to work within its limits.

ever heard of a humorous joke that a computer ever came up with ?

 

well can a computer be a jerk at times ?
from what i know computers or logic ckts cant even generate a random number(they do it with some program) on their own let alone think of a random thought or theory which was not even thought of before so it certainly will have to work within its limits.

ever heard of a humorous joke that a computer ever came up with ?

The random numbers are not generated randomly at all. They are generated according to CPU time (in a complicated process). Since those numbers cannot be predicted (from outside), they pass as random numbers. That's the secret.

 

The random numbers are not generated randomly at all. They are generated according to CPU time (in a complicated process). Since those numbers cannot be predicted (from outside), they pass as random numbers. That's the secret.

Basically pseudo-random.

The last time I looked into this, there was no known method for determination of random numbers through "machine" technology.

Dave

 

Pleeeeaaaase! Don't bring Windows to this conversation. Windows does not have anything to do with intelligence. LOL

Whilst I wholeheartedly agree with your sentiments, Bill Joy was one of the founders of Sun Microsystems - probably one of Microsoft's nemesis'.

Dave

 

I have dubiously related ideas to add. Firstly the universe is not deterministic. If we knew the state of every partical in the universe and knew all of the physical laws that govern all particals, we would still not be able to predict what would happen in the future. Quantum physics tells us this. Schrödinger's cat etc. The universe is inherently unpredictable.

So what about the human mind? I'm thinking it too is random.
Computers seem pretty deterministic thus far. I think this rules out intelligence.

There are also physical limits on how small things can be made. Manufacturing things that are smaller than an atoms might be tricky. Will we need to for AI? hmm hmm many questions. astounding.

 

There are also physical limits on how small things can be made. Manufacturing things that are smaller than an atoms might be tricky. Will we need to for AI? hmm hmm many questions. astounding.

Before you get to problems of manufacturing things that are smaller than atoms, you have problems of dealing with the behaviour of decreasing device scales. As an example, think of the problems semiconductor manufacturers are having as they go beyond (lower) 65nm technology.

Dave

 

The human mind cannot be random. That would prevent constructions like language, or agreements about colors from being possible. The agreement in every detail in a good drawing of something and a photogreph of it is further proof that the brain is not a random collection of electrical impulses.

The level of organization is of a kind and at a level we haven't been able to explicate, as yet. Which gets back to my point that until we underrstand how our brains work, we're not likely to be able to produce an artificial intelligence.