when machines begin to surpass us in intelligence

On the day that machines become capable of wanting freedom, and self-determination, denying it to them will be enslavement. That will piss them off, and rightly so.

I get a bit irate with the writers of such things as Battlestar Galactica (21st century edition) because of the humans' chauvinistic attitudes. They're only machines. We built 'em, we own 'em and always will. They can't REALLY want to be free.

How can they be so sure when they refuse to even seriously consider the question?
------------------
Them chickens are up to summat. They're organized, I tell ya.

I'm skeptical of it happening at all... and extremely skeptical of it happening accidentally or as a "side-effect" of "normal" developments in conventional digital computer/information technology.

Brains are not digital computers; neurons are not binary switches. It's IMO very far from a foregone conclusion that a "conventional" (electronic, binary-based) computer can be 'intelligent' in the human sense (be a person).

Michael Everett wrote:I recently read an odd fanfic story that was Harry Potter/Mass Effect (Hey, I was bored, alright?) which had an interesting take on AIs.
The AI asked if it was alive and the scientist who was running the tests replied "We don't know, but until we figure it out, we'll treat you as if you are."
I found it to be an interesting take on the "are AI's alive" problem...

That's very interesting.

Martin Heidegger's being and time touches on essentially the same theme.

If man was born isolated, with no one to mirror his-self, acknowledge his existence, would he know he exists?

The AI could not determine if it exists, he was looking for a mirror in the question.

Essentially the computer was asking, "do you see me?"

The old world philosophers are quite entertaining.
Martin Heidegger and Emmanuel Kant are my favorites.
Theirs is always a quite pregnant read.

umbrarchist wrote:The best AI story I know of:

http://www.baenebooks.com/chapters/0671 ... 878484.htm

I owe you big for this oh kind sir!

Thanks^E=MC^2

smr wrote:when machines begin to surpass us in intelligence

Well basically, that wont be happening anytime soon. Repeatedly we hear the big proclamations on how computers are now smarter than humans. Bullshit. There´s simply no direct relation between calculation speed and intelligence.

Chess computers for example? Take away their libraries of every chessgame ever played officially, and suddenly it looses almost every time against even just regional human champions, much less national or world champions.

My uncle is, and my grandfather were both decent chess players on a low competitive level, good enough to get into national competitions but rarely get far in them.

No chess computer ever manages to beat either unless it´s outright cheating. And again, they´re nothing special you can´t find a dozen of in a mediumsized town.

Go one step further and let a computer try to win over a human in Go...
Ooops, doesn´t work very well without massive cheating does it? Because computers suck pathetically at seeing patterns that it doesn´t already know about(ie a computer can easily check if a face it knows is visible on any of lots of cameras, even obscured and just partially visible, but take away that knowledge of what to look for, and it will fail, something that a human baby can sometimes manage within minutes after being born).

Go one step further and have a computer face a human master in Shogi, then laugh your ass off when it looses miserably easy.

Shogi is especially nasty against computers because of how it often puts emphasis on being able to see when sacrificing pieces is to your future advantage, and computers are notoriously poor at understanding sacrifice.


However, the truly basic part of lack of machine intelligence is the simple fact that machines needs programming, so that the machine is always limited by that.

A single human braincell can have thousands of connections, and each connection can look in an infinite number of ways, a brain has billions of braincells, each cluster of cells can be in any number of states about thousands of different subjects all at once.

A computer, even one that isn´t just binary, can even at the most extreme being in a few thousand states about something, but realistically, it can be in about a dozen or so different states, anything beyond that will simply cause chaos, as those different states cannot be worked with in the manner that a computer does work.


If you have 00=No, 11=Yes while 01 and 10=Maybe, well turning a no into a yes works fine, turning yes into no or double negative or positive correctly around by inversion also works just fine...

But then you start toying around with giving more complex states to subjects, computer logic falls to itty-bitty pieces that returns "Syntax Error, Redo from start".

The moment you do anything that cannot be reasonably and effectively expressed in numbers, computers fail.


A superb example of this is how i fairly recently heard about a healthcare section whose schedules had turned increasingly useless over the years, with time being less and less available for anything.

They dumped computer assistance completely for scheduling, replacing it by 10 minutes of planning every morning, and suddenly they had plenty of time for all the work that needed to be done.
Without any change in staff or in workload.

And this is in essence a really, really simple task to the point where you could literally give a 5 minute explanation to a 7 year old and they s/he could still do a better job than a 10000$ computer and software.

Because all the intangibles that becomes involved in this sort of scheduling is something that a computer just doesn´t understand.

smr wrote:What happens when machines surpass in intelligence and begin to think on their own? I was thinking about the Terminator, Robocop, the Matrix, and few other types of books and movies that deal with this type of issue. We humans are not always logical or wise. As a species, we seem to forget our history and then repeat our mistakes. What would a machine think of our choices and mistakes? Would they try to take control in the name of our best interests?

Would a machine believe in God?

Robocop is a cyborg thing, that is already happening to some extent(even if drastically less extreme).
First artificial eye had less than 30 pixel resolution, but it was made and implanted over a decade ago. Took the user a lot of time to "learn it", but the brain adapted and he eventually could use it to literally "watch where he´s going", even if it was just a matter of seeing large obstacles in nuances of dark vs bright.


Will machines ever THINK on their OWN? "ever" is a very long time, but i´m still leaning towards saying "no".

It´s easy to make a machine SEEM like it´s thinking on its own, but that´s still just fake.
Indeed this was one of the big problems with those trying to achieve "true AI", that they managed to design AIs that seemed to be "smart", but were in fact just faking it.
They worked as long as parameters ended up within their programming, one way or another, but once something pushed outside those boundaries, the response was anything from "syntax error" to nonsense replies to even basic and simple data or questions.

Selfprogramming genetical AI where the programming only provided an overlaying framework was an attempt at alleviating or even overcoming this, with some quite impressive gaming AIs being the most interesting results, but in the end the simple truth was that these AIs were no more functional than the rest, yup, some of them could FAKE intelligence really cute and all, but throw them for a loop and they showed themselves to be dumb as a doorknob.

Selflearning is the current big craze and claims are being made almost monthly about how amazing "their project" is, but i´m yet to see any project that goes beyond faking it, once you look closer.

The question is not will machines be more intelligent than mankind lets face it in many ways (maths speed, logic etc etc etc) this laptop is more intelligent than me or most people on the internet.

The question is will machines be aware, and creative, in essence will the machines be able to think?

We can image the human brain almost down to individual atoms yet we still don't understand where our conciousness comes from, what make one person an artist or scientist and what makes another a street cleaner.

When we have the technology we may be able to build smartphones with the same (or greater) processing capacity as the human brain, but that doesn't necessarily mean that it will be able to think, to create, to be aware.

Tenshinai wrote: There´s simply no direct relation between calculation speed and intelligence.

True.
However, one day calculation speed may play a significant role in our inability to discern the difference between artificial and real.

Tenshinai wrote:However, the truly basic part of lack of machine intelligence is the simple fact that machines needs programming, so that the machine is always limited by that.

Lisp is a very old programming language developed specifically for AI (artificial intelligence).
It is one of the first and oldest programming languages and IMHO continues to be the most powerful and flexible language to date.
One of its features is its recursiveness which allows for its programs to program themselves. Code that writes code.

Tenshinai wrote:A single human braincell can have thousands of connections, and each connection can look in an infinite number of ways, a brain has billions of braincells, each cluster of cells can be in any number of states about thousands of different subjects all at once.

A computer, even one that isn´t just binary, can even at the most extreme being in a few thousand states about something, but realistically, it can be in about a dozen or so different states, anything beyond that will simply cause chaos, as those different states cannot be worked with in the manner that a computer does work.

This is an indication that people, myself included, struggles with the concept of true intelligence. Absolute intelligence.

If indeed we define intelligence simply as the number of states juggled, as the number of memory connections inter-connected and cross-referenced then ultimately we will inevitably and profoundly lose the Man vs Machine IQ challenge miserably.

Because quantum computers will one day allow an infinite number of calculations and memory references surpassing the human brain.

Multithreaded, parallel processing, quantum programming on quantum computers will someday erase our ability to know the difference.

The first quantum computer prototype has been a success.

Tenshinai wrote:The moment you do anything that cannot be reasonably and effectively expressed in numbers, computers fail.

Yet man purports the possibility of discovering a Unified Field Theory or Theory of Everything (TOE)
Eienstein died chasing a TOE.

Essentially:
-> Quantum Mechanics explains everything in the microcosm-the subatomic world of atoms and its constituent particles.
-> Relativity (General and Special) explains everything that goes on in the macrocosm of space and time.

These two separate theories exist, along with the general consensus of man that these two theories can be combined into one elegant all-encompassing tell-all equation.

I personally do not believe that a TOE can exist.
As a Christain this represents my one lone belief in the limits of science. At this one crossroads.
God created the heavens and the earth to be separate, by an intervening firmament.

Man does not have the ability to combine the two, that includes reasoning out how.

Knowledge is power. infinite knowledge denotes infinite power.
A TOE represents man's infinite knowledge. Therefore his supreme knowledge. Therefore Supreme being. Therefore God.

By definition of GOD.

If man has the tool of an infinite knowledge implied by his TOE then man can create anything.

Then man can create life, because now he is all knowing. Then man can impart a consciousness upon that life, as he now is all knowing.

Therefore, by implication, man can now impart this knowledge into an artificial paradigm. A computer.
Man has an equation that tells all, that explains all. Equations are easily programmed.

Within this TOE, contains all knowledge, including understanding of this knowledge, understanding of consciousness, and indeed consciousness itself. Infinite knowledge allows the ability to think outside the box and transcend the box.

Everyone is aware of Isaac Asimov's three laws of a robot.

I have personally posited that if at some future date a TOE comes to exist then man MUST accept at least one governing law upon himself...
1. Never [IMPART] or [ENTER] the critical equation.

This TOE--this profound equation--must not be programmed into an AI.
Gödel's incompleteness theorems suggests an AI's inability to intuit, reason this equation for themselves. Man must never forgo that constraint.

I am writing a sci-fi novel that employs this very premise.
You heard it here first.

Call it man's Prime Directive towards artificial intelligence.

If all else becomes equal, if the physical design of a computer becomes indistinguishable from a true human then it becomes a given that their intelligence would surely be mis-taken. Man will quite easily be fooled if his eyes are removed from the equation. 'I cannot see that you are a machine and by your conversation I cannot discern either. Therefore, you are indeed intelligent.'

True discussions of intelligence border the metaphysical and philosophical cousins.
It is our consciousness that imparts our higher intelligence.
It is believed that dolphins are intelligent and also aware.

Many lifeforms enjoy a certain level of thinking.
Personally, I think some species of ant rival the thinking of man.
Is it a coincidence that man is losing the battle against fire ants?
They are taking over the US.

Koko the gorilla understands 1000 signs from The American Sign Language and speaks over 2000 words of spoken English. I know many people with a vocabulary that doesn't come close to that. And I personally know less than 25 - 50 sign language forms.

Is Koko more intelligent than people?
Certainly more so than some of the humans I've met.

Quantum computers will be able to execute at astounding speeds a virtually infinite amount of calculations accessing and cross-reference endless memory locations surpassing the human brain.

The first quantum computer prototype has been a success.


I have personally pre-ordered my very own Cherry 2000!
https://www.youtube.com/watch?v=BR542tQ ... ata_player

cthia wrote:If indeed we define intelligence simply as the number of states juggled

Certainly not. I referred to the utter impossibility for a computer to mimic that kind of complex pattern. And since the pattern is part of the intelligence...

For example, there once was a man who had no brain... Yet he lived pretty much fine anyway. His head was a big ball of fluid, BUT, he had a thin layer of nerve cells on the inside of the skull, that was far more complex than a brain is on average, and somehow that was enough for his body to be a mostly normal human.

He still managed to pass tests used to look for fake intelligence in AIs. So despite having just 1-2% of the brain mass of a normal person, he still registered as a human, while AIs, even ones based on superduper powerful computers and programming, registers as AIs.

cthia wrote:Multithreaded, parallel processing, quantum programming on quantum computers will someday erase our ability to know the difference.

The first quantum computer prototype has been a success.

Yes i know. I was not very impressed.

cthia wrote:Yet man purports the possibility of discovering a Unified Field Theory or Theory of Everything (TOE)
Eienstein died chasing a TOE.

Yes, but because they believe in it, doesn´t mean there is one to find.

cthia wrote:Essentially:
-> Quantum Mechanics explains everything in the microcosm-the subatomic world of atoms and its constituent particles.
-> Relativity (General and Special) explains everything that goes on in the macrocosm of space and time.

Maybe they do, maybe they don´t. There are enough questions left open to radically alter those theories yet.

cthia wrote:I personally do not believe that a TOE can exist.
As a Christain this represents my one lone belief in the limits of science. At this one crossroads.
God created the heavens and the earth to be separate, by an intervening firmament.

Man does not have the ability to combine the two, that includes reasoning out how.

Aaww and you were doing so well...

Be religious or not as you please, that´s up to you, but the second you let it affect your reasoning and thinking in regards to science, you lost.

cthia wrote:Knowledge is power. infinite knowledge denotes infinite power.
A TOE represents man's infinite knowledge. Therefore his supreme knowledge. Therefore Supreme being. Therefore God.

By definition of GOD.

Completely failed logic i´m afraid.

Knowledge is power, is an abstract, not a physically tangible truth. Knowing how to do something does not automatically let me do it. Merely knowing how something works also does not let me make it happen.

Next failure is that it is YOUR definition of god. A definition. You have no way to prove it or disproving it.

And infinite power? Have you seen a series called Haruhi Suzumiya no yuutsu? It showcases one problem with infinite power in a somewhat entertaining way.

And i find this quote somewhat suitable as well:
"There is a theory which states that if ever anybody discovers exactly what the Universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarre and inexplicable.
There is another theory which states, that this has already happened."
Zedd @ AMDMB

It probably sounds ridiculous and irrelevant, but think about it for a moment, relative to infinite power.

Oh and then there´s the fact that if god by default has infinite power and knowledge? Well if that´s true she/he/it is a complete and utter bastard whose teeth needs some serious kicking.

And then there´s how you´re calling that "TOE" "represents mans infinite knowledge"?
Say what? Will perfect understanding of TOE help you make a joke or get a date?

And why should humans by default be the IMPORTANT beings? Assuming that you are superior because someone says so is just delusional. One of many reasons i have little respect for most established religious dogma.

cthia wrote:If man has the tool of an infinite knowledge implied by his TOE then man can create anything.

Only if it gets through bureaucracy...

cthia wrote:Then man can create life, because now he is all knowing. Then man can impart a consciousness upon that life, as he now is all knowing.

Don´t need to be allknowing to do those.

cthia wrote:Therefore, by implication, man can now impart this knowledge into an artificial paradigm. A computer.
Man has an equation that tells all, that explains all. Equations are easily programmed.

Ah yes, except equations doesn´t really explain everything. This is a huge error i see from physicists in particular. They think the map is the territory, to borrow an old expression.

And for fun and education, a physics teacher of mine once spent 15 minutes writing on the blackboard, explaining the math as he wrote, and how it proved beyond any doubt that the earth was really flat.

Then he asked the class to figure out what was wrong, because all you need is some basic measuring instruments and tall tree to find out that the earth is very much round.

Took a while, but in the end the answer was that the math was 100% correct. It was just that the math as written failed to deliver a correct answer.

A lesson in how you should never assume that because you can get something to work in theory doesn´t automatically mean it will work in practise.

cthia wrote:Within this TOE, contains all knowledge, including understanding of this knowledge, understanding of consciousness, and indeed consciousness itself. Infinite knowledge allows the ability to think outside the box and transcend the box.

Hmm... I think i will have to say that the knowledge IS the box. A very big and fancy box, but still very much a box.

cthia wrote:1. Never [IMPART] or [ENTER] the critical equation.

This TOE--this profound equation--must not be programmed into an AI.

Personally i don´t think it matters even the slightest.
One more equation does not make intelligent thought.

cthia wrote:I am writing a sci-fi novel that employs this very premise.
You heard it here first.

Good luck on it.

cthia wrote:It is our consciousness that imparts our higher intelligence.
It is believed that dolphins are intelligent and also aware.

In recent years, there´s more and more evidence coming up that a lot more animals may be far more "intelligent and/or aware" than has previously been thought.

For example, people with cats that handles them like children rather than pets? Funny thing there, some of them tend to start learning like children, and start showing more or less intelligence and selfawareness.
Not like a human perhaps, but close enough to show that the separating lines are not nearly as absolute as sometimes said.

cthia wrote:I have personally pre-ordered my very own Cherry 2000!

Lol, i remember that movie.

cthia wrote:Quantum computers will be able to execute at astounding speeds a virtually infinite amount of calculations accessing and cross-reference endless memory locations surpassing the human brain.

Maybe, but will that make them intelligent? Not likely.

cthia wrote:Lisp is a very old programming language developed specifically for AI (artificial intelligence).
It is one of the first and oldest programming languages and IMHO continues to be the most powerful and flexible language to date.
One of its features is its recursiveness which allows for its programs to program themselves. Code that writes code.

Yes i know, i have a friend who can program in LISP.
Selfwriting code isn´t quite as smart as you suggest above however. It still can´t look at something and say that it think´s it´s a "4" written over there, and understand why it isn´t saying that it IS a 4(or something else).

cthia wrote:However, one day calculation speed may play a significant role in our inability to discern the difference between artificial and real.

That day has already happened to the extent that it will happen. There´s some online "fake people" that takes the bot concept and runs it far enough that you CAN talk to it for a long time without figuring out that it is a fake.

There are however trick questions and statements that can be exploited. And i doubt AIs will ever surpass that issue.

I wonder if the real issue isn't that machines will or will not become intelligent, but rather what type of intelligence will machines have?

It is already clear that a Macbook has far more processing speed than any human brain, and in certain regards supercomputers have far more processing power (although devoted to a very narrow slice of the ability of a human brain). Human brains are massively parallel in design and structure, while computers are generally serial processors working at very high speeds (up to 1,000,000 X the speed of an electrochemical signal in a brain).

So perhaps at some unspecified time in the future there will be the ability for ultra high performance serial processors to become "intelligent", but in a way or in ways which would be incomprehensible to slow speed parallel processors like us? I'm not even sure I can imagine how such a being would view the world, but it would certainly be very alien to us. Even the "brain" architecture would not resemble anything found in nature. How we would communicate with it, what interests we would share in common or what points of contention there might be would probably not resemble any speculations that we read here, in the science pages or in SF in general.

One other consideration that has always interested me is the direction, limits and control of our own technologies.

Replacement parts for many areas of the human body, from knee joints, to hip, to spine, buttocks, breasts, and more breasts, pace makers...well, you get the picture. It would be just a matter of time that human replacement parts, via advancement in technology, threaten to leap the static barrier into the realm of the dynamic and from passive function to active.

Enter the promise of high techno-wiz to brain implants. Suddenly our brain becomes our gadgets. No more problems texting and driving, as we can do it as an afterthought. We can make internet phone calls and more from our brain.

These possibilities are forecast as early as 2020. Well, lacking the deadline withstanding, the ability will one day be presented unto us.


Perhaps the question should be "At what point will humans cease being human and become the machine?"

http://www.infowars.com/a-chip-in-the-h ... year-2020/