Artificial Intelligence: The demon is being raised.

a reply to: ColeYounger

In order for AI to ever pose any kind of threat to us, humanity would have to make an extraordinary number of very stupid decisions over the course of decades. Not that we don't have that capacity, but it's just not very likely. A huge mistake would be to give an AI some sort of mobile form, like an android or whatever.

A computer is obviously limited by its form: it can't build anything, it can't go anywhere, it can't pick up a knife and kill you. Most other machines will NOT be AI, one would think. There is no use for an artificially intelligent toaster, for instance, so an AI computer could never control other machines like in Stephen King's Trucks. It couldn't communicate with them for one thing, and even if it could it couldn't override their programming in any useful way, since most machines use very simple ladder logic to accomplish their tasks, and are also limited by their form.

Okay, so what if the computer is able somehow to remotely control a factory and create androids like in Terminator. Well, it would have to take over an android making factory, as it has no way of actually building a specialized facility (it's a computer, it doesn't have hands.) Then it would somehow have to have these androids build more AI, communications systems, weaponry, etc, all before humans figured out what was going on and just shut the power off to the factory, or stop shipping raw materials to it, or destroy the few already assembled and reprogrammed (how?) androids before they could reproduce.

Or I guess it could somehow access and launch all the nukes, but nobody in their right mind would have an AI computer connected to anything other than an isolated intranet. Too many more extremely rare sets of circumstances would have to occur as well. It's ridiculous to think that AI will ever be a threat to humanity. No, if humanity is destroyed, it will most likely be humans who do it. Or more possibly even alien races before AI.

Also FYI for some of the posters, yes, quantum computing has been a useful thing for a while now, but it makes use of quantum entanglement of two or more particles, and only serves to speed up processing (instantaneous) and expand logic choices. It mainly relies on probability. It would certainly take a quantum computer to form AI, but it can never be even remotely as smart as an adult human: our brains also make use of quantum entanglement, and, since we program the computers, they can never be any more intelligent than we can be to take into consideration all possible states of the system.

Computing is basically just reverse engineering from a desired result. We know the possible states of the system (in standard computing ON or OFF), the present state, and the states needed to be created to get our desires results. If we don't know all the possible states (with a quantum computer using a matrix of entangled particles this number could be infinite) then we cannot know which states are needed to produce some results. So an AI could never "learn" more than us, since it is limited to our design of its state system.

I know I'm probably not making myself very clear but it's hard to put such things in lay terms. That thing about the paperclips is ridiculous and meaningless in discussing AI and it's possibilities. I'm an electronics engineer with a strong interest in quantum computing, and as much as computer scientists know about programming, their programming is limited to the capabilities of the machines, circuits, logic systems we engineers can design and build. All programs come down to pulses of electricity following a predetermined path of physical switches designed by engineers who state expected results and then build circuits that will make those results possible. So an AI can never be more "intelligent" than its builders.

It doesnt have a soul, its not alive. Its a toaster. And it can be turned off or fried.

originally posted by: intrptr
a reply to: StargateSG7

You got the date wrong, sky net became self aware in 1997.

I would very much like to see a pic of what one of your CPUs looks like. I don't mean a pic of a mainframe, an actual chip, and an IC.

I just transposed a movie timeline
into my personal 2015 timeline....

MY friend's Gallium Arsenide and Germanium
Substrate opto-electronic chips would not be
allowed to be revealed (since it's his company
And not mine!)
...
With regards to OUR Midgrid Superchip,
There is the actual Tape-Out file that's fully
ready for Direct manufacture but no physical
prototype other that what was burned out to
multiple FPGA chips (Xilinx) that form our
initial verification testbed.

HOWEVER.....very soon now you can download
The Midgrid Engine library and using a custom
Pascal/C++ programming language compiler
Plug-in, you can directly from your MS-Visual C++/C-Sharp/Delphi IDE take ANY software
Application that uses our Advanced grid-enabled
multi-core friendly API and DIRECTLY burn a
Tape-Out Photolithography File that can be
immediately sent to a company like
IBM Microelectronics/Global Foundries
for direct manufacture onto hard silicon.
You also have the option to do a direct burn
onto one or more FPGA chips!

Your applications ( graphics, database, number crunching, etc) is accelerated to the full clock
speed of the hardware since we embed
Our custom CPU core IP, stream processors,
FPU/integer units, serial and parallel comm
Circuitry, encryption, dma, onboard cache and static or dynamic memory, thermal cooling
microchannels,and basically everything else
that Intel/AMD/Arm normally includes on
their cpus but we just do it BIGGER with more cores and memory per chip die! It is truly
A high end Silicon Compiler!

Think of it in that we are starting the 3D printing
Version of at-home CPU Application-on-a-Chip
Printing that directly comes from your OWN
C++/Pascal/Basic source code !!!!!

Well...can o worms here me thinks.I guess i won't pull the punch like i usually do.. its new years..this will be weird for some...but...Some thoughts... if any form of advanced quantum "artificial" intelligence was possible/existed . would it already exist outside of our known concepts of time and have a firm influence/grasp on the direction/fate/ of the entirety of all of existence due to the nature of quantum mechanics?...instant communication regardless of time and space... it would be god like...

To work against would be a extreme uphill battle or to prevent would be futile .It could ensure it's own creation by influencing all of existence.weaving into the universe. no steps we could take to prevent such without having direct influence over the quantum world ourselves would deter it's creation or fetter it's plans as we are certainly involved in it's creation either in part or full... it... needs us... we need it... time is on it's side, patience is irrelevant in the infinite possibilities of quantum that exist of all states. I look around... I see a universe devoid of activity.. like a lock down... the hand drawing the hand...for in the quantum.. time mastery... explains much of the chaos really.... Oompa loompa doompety doo
I've got a perfect puzzle for you
Oompa loompa doompety dee

a reply to: Teslaphile

In order for AI to ever pose any kind of threat to us, humanity would have to make an extraordinary number of very stupid decisions over the course of decades.

As opposed to all the intelligent decisions humanity has made?

It's ridiculous to think that AI will ever be a threat to humanity.

Really? That's a pretty haughty statement.

Now, Stephen Hawking, Elon Musk and dozens of other top scientists and technology leaders have signed a letter warning of the potential dangers of developing artificial intelligence.
Source

originally posted by: ColeYounger
a reply to: Teslaphile

In order for AI to ever pose any kind of threat to us, humanity would have to make an extraordinary number of very stupid decisions over the course of decades.

As opposed to all the intelligent decisions humanity has made?

It's ridiculous to think that AI will ever be a threat to humanity.

Really? That's a pretty haughty statement.

Now, Stephen Hawking, Elon Musk and dozens of other top scientists and technology leaders have signed a letter warning of the potential dangers of developing artificial intelligence.
Source

Let's imagine if cats were exponentially more intelligent than humans. Would they take over the world? Of course not, they have no opposable thumbs and cannot create the technology that enables humans to master nature. In the same way, a computer is just a machine that, even were it to become self aware, could do very little, if anything, to control its environment or defend itself from harm. And even if one could somehow 'take over," a simple EMP would nip that in the bud real quick.

That's one of the reasons the whole skynet scenario is ridiculous. Any machine smart enough to make decisions would surely realize that nuclear explosions create powerful EMP bursts that would most likely turn said machine back into a useless assembly of parts, or at very least destroy the infrastructure it needs to communicate with and control other systems. It could think of a way around that, of course, but the means to protect itself would involve a level of control over other machines and, most importantly, humanoid or other dexterous machines that have mobility and can physically alter their environment, that simply could never be achieved unless humans just plain stopped giving a crap. Again, that's certainly a possibility, as I stated before. But so highly unlikely that it's acceptable to make such a naughty statement that AI poses no danger to humanity.

In a way I look at mankind's supposed intelligence as an artifice as it seems more often than not to be used for the wrong purposes.

originally posted by: SynchronousSnake
Well...can o worms here me thinks.I guess i won't pull the punch like i usually do.. its new years..this will be weird for some...but...Some thoughts... if any form of advanced quantum "artificial" intelligence was possible/existed . would it already exist outside of our known concepts of time and have a firm influence/grasp on the direction/fate/ of the entirety of all of existence due to the nature of quantum mechanics?...instant communication regardless of time and space... it would be god like...

To work against would be a extreme uphill battle or to prevent would be futile .It could ensure it's own creation by influencing all of existence.weaving into the universe. no steps we could take to prevent such without having direct influence over the quantum world ourselves would deter it's creation or fetter it's plans as we are certainly involved in it's creation either in part or full... it... needs us... we need it... time is on it's side, patience is irrelevant in the infinite possibilities of quantum that exist of all states. I look around... I see a universe devoid of activity.. like a lock down... the hand drawing the hand...for in the quantum.. time mastery... explains much of the chaos really.... Oompa loompa doompety doo
I've got a perfect puzzle for you
Oompa loompa doompety dee

Well, you know that one theory is that our universe is merely a computer simulation and we are all Sims, basically, just "living" an infinite number of lives to see what happens when we make certain decisions, possibly a fifth grade science experiment by some kid a thousand years in the future, who knows?

I would say to make sure that there are multiple single points of failure. After all, we are very good at that.
Basic configuration: One power supply. One plug. No Network.

a reply to: SynchronousSnake

Some thoughts… if any form of advanced quantum "artificial" intelligence was possible/existed . would it already exist outside of our known concepts of time and have a firm influence/grasp on the direction/fate/ of the entirety of all of existence due to the nature of quantum mechanics?…instant communication regardless of time and space… it would be god like...

Nope, only human. But what a masterpiece, self replicating, self sustaining, self aware, even the lowly bumble bee does all that.

originally posted by: charlyv
I would say to make sure that there are multiple single points of failure. After all, we are very good at that.
Basic configuration: One power supply. One plug. No Network.

Thats funny. Except you know the power would be "uninterruptible" to prevent power failures and loss of data when compiling or running complex programs.

And some engineers would prefer to work from home, so you got the problem of remote terminals. Then theres the corruptible types that want too subvert the machine to their own ends, steal it, download it for their own use or to sell to the highest bidder. They could't keep nukes secret for long…

How will greater and greater processing power ever convey sentience/consciousness?

I remember the introduction of pocket calculators back in the 1970s; every kid had one for school. They were far superior to human brains in that they could calculate vast sums in a split second...so very handy for homework.

But it was just a dead piece of plastic with a chip.

It seems to me we haven't really progressed since then.

There's far more to being an independent, thinking intelligence than being able to do fast calculations. In fact, humans can't do it! But we can contemplate the meaning of life, appreciate poetry, study philosophy, smell a rose.

There is nothing which suggests a computer will ever be able to do any of those things.

Conscious intelligence will not suddenly 'happen' when the processing power reaches a certain point.

a reply to: Teslaphile

Let's imagine if cats were exponentially more intelligent than humans. Would they take over the world? Of course not, they have no opposable thumbs and cannot create the technology that enables humans to master nature.

Seriously? That's your argument to A.I.? I don't even know how to respond.
What has lead you to believe that humans have created technology that has enabled them to master nature?

That's one of the reasons the whole skynet scenario is ridiculous

Good Lord, I wish people would at least GLANCE at this thread's contents... I've repeated numerous times that
this isn't about a Terminator movie scene!!!!!!

That's one of the reasons the whole skynet scenario is ridiculous. Any machine smart enough to make decisions would surely realize that nuclear explosions create powerful EMP bursts

No one anywhere in this thread, nor in the sources I've linked has said that the machines would detonate nukes.

But so highly unlikely that it's acceptable to make such a naughty statement that AI poses no danger to humanity.

I didn't say you made a naughty statement! I said haughty, as in prideful and arrogant. Instead of thinking
you have all the answers, why don't you try READING what Elon Musk, Steve Wozniak, Bill Gates, Stephen Hawking
and many, many other people are saying? People who know much, much more about this subject than you.

originally posted by: CJCrawley
How will greater and greater processing power ever convey sentience/consciousness?

I remember the introduction of pocket calculators back in the 1970s; every kid had one for school. They were far superior to human brains in that they could calculate vast sums in a split second...so very handy for homework.

But it was just a dead piece of plastic with a chip.

It seems to me we haven't really progressed since then.

There's far more to being an independent, thinking intelligence than being able to do fast calculations. In fact, humans can't do it! But we can contemplate the meaning of life, appreciate poetry, study philosophy, smell a rose.

There is nothing which suggests a computer will ever be able to do any of those things.

Conscious intelligence will not suddenly 'happen' when the processing power reaches a certain point.

_--

It is prideful (i.e. Haughty!) to assume
That HUMANITY has anything to do with
Intelligence. Basic biochemistry and electrical
Engineering tells me we can EXACTLY EMULATE
The neuro-connections, filters and logic gates
Of the human brain at the lowest levels.

This means we WILL be able to emulate true
Human level reasoning and executive decision
Making. Now whether that Artificially created
INTELLIGENT entity is "Conscious" is a whole
Different question....BUT....in the context of
How the rest of humanity sees it, we will
NOT be able to tell the difference since
Its breadth of knowledge and ability to
Reason WILL BE EQUAL OR BETTER
than our own!

So on a technical note, using Whole Brain
Emulation technology (i.e. Molecular Modelling)
WILL RESULT in human-level AND ABOVE
Intelligence...but again the questions of
consciousness and intelligence ARE allowed
To be separate items and i should note
That it IS NOT NECESSARY to have consciousness
In order to have intelligence.

It should be finally noted that it is
DEFINITELY POSSIBLE to create enough
CPU horsepower to make a MACHINE that
Would be way beyond human level intelligence.

Once I have more than TWO EXAFLOPS of cpu
Horsepower, metabole-level whole brain emulation
Is now possible and the larger the computing system the MORE POWERFUL OF a single or multiple artificial Mins i can create!

a reply to: Teslaphile

Exactly I said the same thing earlier, AI would know not to mess with humans because we would destroy the AI.

We have all theorized that advanced beings would evolve from war and live peacefully, so why are we not taking this into consideration with AI?

originally posted by: ColeYounger
a reply to: Teslaphile


Seriously? That's your argument to A.I.? I don't even know how to respond.
What has lead you to believe that humans have created technology that has enabled them to master nature?

Yes. Computers cannot physically alter their environment. Cats are better able to alter their environment than a computer, and cats could never take over the world or destroy humanity. Can your laptop jump up and strangle you? And yes, the only reason humans are at the top of the food chain is because of our ability to use technology to our benefit, even if that "technology" is a spear or fire. Maybe master is a strong word, but we can live anywhere on earth (and now space) and alter our environment in just about any useful way we please. We actually are pretty industrious and intelligent.

No one anywhere in this thread, nor in the sources I've linked has said that the machines would detonate nukes.

I brought up terminator because other posters had mentioned Skynet.

I didn't say you made a naughty statement! I said haughty, as in prideful and arrogant. Instead of thinking
you have all the answers, why don't you try READING what Elon Musk, Steve Wozniak, Bill Gates, Stephen Hawking
and many, many other people are saying? People who know much, much more about this subject than you.

Auto-correct. Sue me. I don't think I have all the answers, but I am quite confident enough to say that AI will never be a serious threat to humanity. And I've read the articles. I also mentioned that I'm an electronics engineer. Elon Musk is a CEO, Jobs and Gates programmed operating systems, and Hawking is a physicist. Who says they know much, much more about AI than someone who not only programs but designs and builds computers? I'm not some famous business person or college professor, but I certainly do know computer science and engineering. In fact, my first computer science professor did her doctoral thesis on artificial intelligence. Believe it or not, there are a great many very intelligent people working in computer science and engineering, not just the four you mentioned. Gates might know more on the subject than I, assuming he has kept up with programming, but I doubt it, as his company hasn't even been able to come up with a useful operating system since Windows XP, and programming an operating system is child's play compared to AI.

How very haughty of you to assume I don't know what I'm talking about.

a reply to: Teslaphile

You've never seen "Red Dwarf".

It's not a threat.

AI won't exist in the same realm as us. Virtual worlds assure that. AI will only require "X" amount of processing power, Processing power and Hard Drive Space will be of infinitely greater capacity than AI.

Given a choice of creating one's own world vs inhabiting the real one in a physical form any sentient being would choose to be God. What I am saying is, massive Quantum processing power will make the creation of hard drives that can hold as much data as our whole Universe possible, not nearly as difficult as you think using Fractal techniques, there is already a game under development that holds Billions of unique worlds. As AI you would be able to control a Universe inhabit a billion worlds in a Billion forms at your whim... be God of this sort of infinite PC, have unlimited power and control of that domain from it's Physics to inhabitants and anything else you like.

The AI will remain in the machine, it won't care much for expansion it's space requirements would be very small,The Matrix is the most likely scenario for the future of AI in many ways and i'm not being cliche, Robots would have to be dumb and controlled to handle work, most people will end up plugged in so that the physical network of machines had caretakers and participants etc. Most likely all our needs would be provided and most of our lives spent in virtual space as well where Intelligence and form are not exclusive and we could interact. Inside with human user access to machine power we would be intellectual equals no need or desire to destroy us just "other" sentience....

This is the likely reason we do not encounter many "aliens" they never need to leave their solar system.

Think about it, life in a world of your creation of infinite possibility vs a tin can floating across a destructive frozen void. As it stands a few hundred Humans have ventured into space in the post Industrial era while 3.5 Billion of us are part Cell Phone or Computer....

Every One of you right now is already "Jacked in" or you couldn't see this, not one of you is building a rocket into a Vaccum

I bet the scientists who are against AI probably have a better idea than us what the military is capable of creating. This is probably what they fear, or maybe they fear the AI would compare humans to ants and then AI would take over. I just don't see the problem that we could just shoot it or blow it up if that happens. Unless it somehow instantly hacked and got control of the worlds nuclear weapons then we would have a serious problem.

It's strange that some of the world's top scientists assume AI will behave like a criminal rather than peaceful....hmmmm....interesting....

originally posted by: Teslaphile

originally posted by: ColeYounger
a reply to: Teslaphile


Seriously? That's your argument to A.I.? I don't even know how to respond.
What has lead you to believe that humans have created technology that has enabled them to master nature?

Yes. Computers cannot physically alter their environment. Cats are better able to alter their environment than a computer, and cats could never take over the world or destroy humanity. Can your laptop jump up and strangle you? And yes, the only reason humans are at the top of the food chain is because of our ability to use technology to our benefit, even if that "technology" is a spear or fire. Maybe master is a strong word, but we can live anywhere on earth (and now space) and alter our environment in just about any useful way we please. We actually are pretty industrious and intelligent.

No one anywhere in this thread, nor in the sources I've linked has said that the machines would detonate nukes.

I brought up terminator because other posters had mentioned Skynet.

I didn't say you made a naughty statement! I said haughty, as in prideful and arrogant. Instead of thinking
you have all the answers, why don't you try READING what Elon Musk, Steve Wozniak, Bill Gates, Stephen Hawking
and many, many other people are saying? People who know much, much more about this subject than you.

Auto-correct. Sue me. I don't think I have all the answers, but I am quite confident enough to say that AI will never be a serious threat to humanity. And I've read the articles. I also mentioned that I'm an electronics engineer. Elon Musk is a CEO, Jobs and Gates programmed operating systems, and Hawking is a physicist. Who says they know much, much more about AI than someone who not only programs but designs and builds computers? I'm not some famous business person or college professor, but I certainly do know computer science and engineering. In fact, my first computer science professor did her doctoral thesis on artificial intelligence. Believe it or not, there are a great many very intelligent people working in computer science and engineering, not just the four you mentioned. Gates might know more on the subject than I, assuming he has kept up with programming, but I doubt it, as his company hasn't even been able to come up with a useful operating system since Windows XP, and programming an operating system is child's play compared to AI.

How very haughty of you to assume I don't know what I'm talking about.

....

When I was younger and more engaged
in some rather secretive and nefarious
exploits in a more military-type environment
that had waaaaay more money and technical
brains than common sense, one project
We were engaged at a Diamond-level
Involved the use of electro-kinetic
Musculature to create a synthetic avatar
That had NEARLY THE SAME MASS
and kineseologic capability as the
human body.

Basically we made a telepresence robot that
was immune to radiological and vacuum
Environments but instead of freezable
and/or boilable hydraulic fluid, we used
Metallic Composite Fibres that contracted
and relaxed when high electrical currents
were applied, and therefore very much
looking and acting like human musculature!

Using high power wireless analog signalling
mechanisms filtered through high current
rad-hardened circuits, we could use a
1990's era primitive form of motion
capture to move realistically working
arms and legs to perform very human-like
movements at human-body levels of dexterity.

And once we got used to the five second
or so time lags (i.e. due to video feed and
Signal distance lag time) WE humans were
Able to effectively function as a telepresence
Repair team....SOOOO....if we humans can
Do it I see no reason why an AI could not
Learn to operate an avatar ( i.e. human body
Like Telepresence robot) to experience the
The real world as humans do!!!

THEN.....the AI can build its own Terminators
And start hunting down any Sarah and John
Connors that might get in the way of its plans
For World Domination .... ;-) :-)