Will advances like AI and nanotechnology make humans obsolete?

[Genesys]

Originally posted by: LadyJessica

if youre going to do it that way, compare it piece to piece, not piece to whole.

how about neurons to the tracers on any imprinted circut board? all they can do is relay electrons from the start of that tracer to the end.
and humans are more intelligent than computers because we have the capacity for independant thought. a computer only runs a series of calculations for us and displays the result.

That doesn't work either. a better comparison is a cpu and a neuron since they're both fundamental units of a computer and brain respectively. Now you say we're more intelligent because we have the capacity ofr independant thought. Have you ever wondered fundamentally what that entails? Given, that our brains our brains are composed of neurons (and some supporting glial cells) that just pass signals here and there, the logical conclusion is that our "independant thought" is a compilation of the signals that our neurons pass around to each other. A computer runs whatever instructions it is fed. Neurons aren't much better. They just receive multiple signals and fire/not fire depending on the signal aggregate. Can't say there's much intelligence there.

lol, im going to keep breaking your argument down untill one of us gets it to a very diluted state, but oh well.

comparing a CPU to a neuron doesnt work all that well either. Youd have to be more selective in that argument by picking a certain part of the CPU, just like you have for the brain. If youre going to pick one of millions [billions?] of neurons in the brain, pick one oh hundreds of the pins on a CPU. they each have their function of relaying electrical signals to their various locations, but without the whole, neither would get much accomplished.

and thats exactly the difference, computers only run the instructions that are fed to them. they have no ability to come up with new instructions [or calculations] independantly. they still need an operator, humans do not.

also, in addition to humans being more intelligent because of independant thought, id like to add that were more intelligent because we learn. if i burn my hand on the stove, ill more than likely not make the same mistake [unless i really like pain] however, a computer has no ability to learn. it will continue to execute the same action over and over untill its operator tells it to halt.

 

[Drift3r]

Originally posted by: Ultima

Originally posted by: Genesys
i hope so. machines think logically [unless for some god forsaken reason we decide to program emotions into an AI subsystem] so any task they get set to do, they execute with utmost efficiency.

I think designing AI after humans is a big, big mistake that will only lead to our death (Darwinism anyone?) and I hope scientists and governments are smarter than to make a super-intelligent AI with emotions and the like. The best AI anyone could hope for would not have concepts like ambition, greed, survival instinct, evolutionary techniques, etc.. built into it, although if the AI is powerful enough it would be able to understand those concepts. It would be intelligent, but not the same kind of intelligence as a person.

Or then again a lack of emotion could be our biggest downfall. The AI decides that humans are obsolete and limiting it's development thus decides with no emotion involved it makes the decision to get rid of it's only limiting factor that is holding back it's devlopment, namely us ! Maybe emotions might be needed for example pity, and kindness are two emotions that could save us.

 

[LadyJessica]

comparing a CPU to a neuron doesnt work all that well either. Youd have to be more selective in that argument by picking a certain part of the CPU, just like you have for the brain. If youre going to pick one of millions [billions?] of neurons in the brain, pick one oh hundreds of the pins on a CPU. they each have their function of relaying electrical signals to their various locations, but without the whole, neither would get much accomplished.

You're right in some sense. A much better comparison would be a cpu to neuron comparision in a massive neural net cluster and intact human brain respectively. For current uniprocessor computers, a comparison of a gate to a neuron would be more appropriate.

and thats exactly the difference, computers only run the instructions that are fed to them. they have no ability to come up with new instructions [or calculations] independantly. they still need an operator, humans do not.

That's true as well. However that's an unfair comparison. A cpu and neuron cannot come up with new instructions since their proper function depends on not being able to do this. Else you would get a crash in a uniprocessor computer. However, a malfunctioning cpu in neural net cluster doesn't matter as much just as one dead neuron in your brain is fairly irrelevant.

also, in addition to humans being more intelligent because of independant thought, id like to add that were more intelligent because we learn. if i burn my hand on the stove, ill more than likely not make the same mistake [unless i really like pain] however, a computer has no ability to learn. it will continue to execute the same action over and over untill its operator tells it to halt.

Yes we learn. But so can a computer (and not just neural net computers). For a neuralnet cluster, it is conceivable that hooking up sensors to it will allow it to "feel" heat and retract whatever appendage it has.

 

[Moonbeam]

Genesys is very proud and doesn't believe he is a machine. The evidence that we are a biological machine to which, probably owning to self monitoring feedback mechanism to prioritize actions based on survival concerns, there has arisen a sense of self, does not mean that something like that can't occur for other machines given sufficient complexity. Our complexity is a result to billions of years of unself aware evolution. I believe that for machines that period could be fastly compressed by evolution via sim, with the next generation sim product evolving the next next sim and building it in turn. When we think of machines wiping us out due to a lack of emotion, don't forget that all life feeds on life except for simple plants. We are our own kind of monster.

 

[Genesys]

Originally posted by: LadyJessica

comparing a CPU to a neuron doesnt work all that well either. Youd have to be more selective in that argument by picking a certain part of the CPU, just like you have for the brain. If youre going to pick one of millions [billions?] of neurons in the brain, pick one oh hundreds of the pins on a CPU. they each have their function of relaying electrical signals to their various locations, but without the whole, neither would get much accomplished.

You're right in some sense. A much better comparison would be a cpu to neuron comparision in a massive neural net cluster and intact human brain respectively. For current uniprocessor computers, a comparison of a gate to a neuron would be more appropriate.

and thats exactly the difference, computers only run the instructions that are fed to them. they have no ability to come up with new instructions [or calculations] independantly. they still need an operator, humans do not.

That's true as well. However that's an unfair comparison. A cpu and neuron cannot come up with new instructions since their proper function depends on not being able to do this. Else you would get a crash in a uniprocessor computer. However, a malfunctioning cpu in neural net cluster doesn't matter as much just as one dead neuron in your brain is fairly irrelevant.

also, in addition to humans being more intelligent because of independant thought, id like to add that were more intelligent because we learn. if i burn my hand on the stove, ill more than likely not make the same mistake [unless i really like pain] however, a computer has no ability to learn. it will continue to execute the same action over and over untill its operator tells it to halt.

Yes we learn. But so can a computer (and not just neural net computers). For a neuralnet cluster, it is conceivable that hooking up sensors to it will allow it to "feel" heat and retract whatever appendage it has.

you're right, a gate would be a more appropriate comparison. i almost chose a random logic unit originally, but decided not to for some reason.
i dont think that a uniprocessor would crash if it started comping up with its own instructions. the processor is only executing the code and doing the math that it is being instructed to by whatever piece of software you're using. even then, you could make an argument that the whole computer is using some sort of super duper set of firmware that allows it to 'think' and the machine still wouldnt crash because its executing the code it was told to.

snfd for your neural net cluster feeling experiment, youd have to tell the computer how to react. for humans, its a natural response to pull away, for a computer it would just leave the sensor there untill the sensor read XX.XX temperature untill it started to withdraw the sensor. then where would you store that info? on a hard disk [or write it back to super duper firmware]? what happens if the hard disk gets erased [or if the firmware gets erased]? that would mean that computer would have to learn what temps are acceptable all over again, whereas a human with alzheimers will still naturally withdraw its hand from the fire.

hmmmm, perhaps instinct will be what eventually seperates us from machines.

and moonbeam, please make more sense. if you want to speak philosophically, go visit a philisophic forum, this is a technology forum where you speak technically

 

[myusername]

Genesys, Your misperception arises from both a lack of understanding of what a true neural net is, and of how biological nervous systems function.

snfd for your neural net cluster feeling experiment, youd have to tell the computer how to react. for humans, its a natural response to pull away,

It is not a "natural response to pull away". In this particular instance, assuming the biological organism in question is human, it has most likely "learned" not to stick its hand in a fire becuase it has been "programmed" through verbal parental instructions.

Because a human is a true neural net, it may have been "programmed" through association - or had verbal programming reinforced through experience - with some other exothermic object that generated a pain response.

As for the initial response, the reflex to withdraw is based on pain, rather than temperature. In this particular instance, pain is simply a biological quantification of temperature.

This is not inconsistent with neural net functioning.

for a computer it would just leave the sensor there untill the sensor read XX.XX temperature untill it started to withdraw the sensor.

Sure, if you program a conventional computer that way. But that's not the point of a neural net. You could use programmed fuzzy logic, if using a software overlay, or you could use analog processors.

At some point in an initial response stage, the neural net would have to be told that a particular temperature is "too hot". You could toss in a handful of other variables (ambient temp, initial temp of probe, whether or not it is raining in Bangladesh, etc.), but at some point, yes a computer has to be told what its pain threshold is.

That's essentially the same thing a human does, except the pain response is built into the hardware. Ever sat too close to a campfire? You sit there nice and toasty for a couple minutes until you start to feel discomfort, and once your discomfort reaches a threshhold, you move a couple feet back.

The variable speed at which you move and the length of time you take to make the decision to move are not inconsistent with a neural net.

then where would you store that info? on a hard disk [or write it back to super duper firmware]?

If we are talking about a modeled neural net, or a software overlay, volatile memory is the logical answer. If we are using analog processing, it could be done at the hardware level - the data is never actually stored as a single concrete number; that's the whole point of a neural net!

what happens if the hard disk gets erased [or if the firmware gets erased]? that would mean that computer would have to learn what temps are acceptable all over again,

I think you are confusing the underlying hardware structure with the "hardware" that the software is simulating. Of course a cataclysmic computer event would wipe out the program in a simulation. In an analog device, it's not so simple. You're basically asking what happens to a human's ability to sense heat/pain after they get run over by a train.

whereas a human with alzheimers will still naturally withdraw its hand from the fire.

You are confusing reflex with learned response, as well as making assumptions about the symptomology of alzheimers. Since the brain *physically* changes as it learns (neural pathways are reinforced and atrophy), the only way to "wipe" a human brain as completely as a system crash wipes volatile memory is to remove it from the skull, run it through a blender on "frappe", and return it to the skull.

Assuming a less extreme action, the only way to "wipe" a human brain would be to return it to the same physical state as a prior point in time. Of course, then there would have to be a relearning phase starting with the initial heat/pain response.

apples!=oranges.

and moonbeam, please make more sense. if you want to speak philosophically, go visit a philisophic forum, this is a technology forum where you speak technically

Actually this is a politics and news forum; most of this thread is off-topic. Moonie's post was at least as topical as any other, despite your inability to parse his prose. Kurt Vonnegut would agree, but then, this isn't a literature forum either. If this were a technology forum that required "speaking technically", we would all be asked to visit a different forum

 

[Genesys]

Originally posted by: myusername
Genesys, Your misperception arises from both a lack of understanding of what a true neural net is, and of how biological nervous systems function.

snfd for your neural net cluster feeling experiment, youd have to tell the computer how to react. for humans, its a natural response to pull away,

It is not a "natural response to pull away". In this particular instance, assuming the biological organism in question is human, it has most likely "learned" not to stick its hand in a fire becuase it has been "programmed" through verbal parental instructions.

Because a human is a true neural net, it may have been "programmed" through association - or had verbal programming reinforced through experience - with some other exothermic object that generated a pain response.

As for the initial response, the reflex to withdraw is based on pain, rather than temperature. In this particular instance, pain is simply a biological quantification of temperature.

This is not inconsistent with neural net functioning.

for a computer it would just leave the sensor there untill the sensor read XX.XX temperature untill it started to withdraw the sensor.

Sure, if you program a conventional computer that way. But that's not the point of a neural net. You could use programmed fuzzy logic, if using a software overlay, or you could use analog processors.

At some point in an initial response stage, the neural net would have to be told that a particular temperature is "too hot". You could toss in a handful of other variables (ambient temp, initial temp of probe, whether or not it is raining in Bangladesh, etc.), but at some point, yes a computer has to be told what its pain threshold is.

That's essentially the same thing a human does, except the pain response is built into the hardware. Ever sat too close to a campfire? You sit there nice and toasty for a couple minutes until you start to feel discomfort, and once your discomfort reaches a threshhold, you move a couple feet back.

The variable speed at which you move and the length of time you take to make the decision to move are not inconsistent with a neural net.

then where would you store that info? on a hard disk [or write it back to super duper firmware]?

If we are talking about a modeled neural net, or a software overlay, volatile memory is the logical answer. If we are using analog processing, it could be done at the hardware level - the data is never actually stored as a single concrete number; that's the whole point of a neural net!

what happens if the hard disk gets erased [or if the firmware gets erased]? that would mean that computer would have to learn what temps are acceptable all over again,

I think you are confusing the underlying hardware structure with the "hardware" that the software is simulating. Of course a cataclysmic computer event would wipe out the program in a simulation. In an analog device, it's not so simple. You're basically asking what happens to a human's ability to sense heat/pain after they get run over by a train.

whereas a human with alzheimers will still naturally withdraw its hand from the fire.

You are confusing reflex with learned response, as well as making assumptions about the symptomology of alzheimers. Since the brain *physically* changes as it learns (neural pathways are reinforced and atrophy), the only way to "wipe" a human brain as completely as a system crash wipes volatile memory is to remove it from the skull, run it through a blender on "frappe", and return it to the skull.

Assuming a less extreme action, the only way to "wipe" a human brain would be to return it to the same physical state as a prior point in time. Of course, then there would have to be a relearning phase starting with the initial heat/pain response.

apples!=oranges.

and moonbeam, please make more sense. if you want to speak philosophically, go visit a philisophic forum, this is a technology forum where you speak technically

Actually this is a politics and news forum; most of this thread is off-topic. Moonie's post was at least as topical as any other, despite your inability to parse his prose. Kurt Vonnegut would agree, but then, this isn't a literature forum either. If this were a technology forum that required "speaking technically", we would all be asked to visit a different forum

this is going to be a quick placeholder response, it will be edited further after i get home from work Sunday [which will be after 8pm if anyone cares]

i think youre rather misguided on the fact that you seem to think that it is not a natural response for a child [biological organism (human)] to pull its hand out of the fire. it is instinctual, not learned. [do you have children?] a parent can tell a child not to stick its hand in fire [or in most cases its now, 'dont touch the burner on the stove, its hot and will burn you'] but the child will do it anyway and have to learn that way.

pain is the only way a biological organism learns. words are meaningless by themselves because they have nothing associated with them. [i believe you touched on this concept in your post] associate pain with the words and suddenly theres meaning [and hopefully avoidance to pain]

 

[Moonbeam]

Living is not avoiding pain. Humans learn because self awareness, love, and wonder are bound together. The universe as it is has the property that evolution proceeds toward love. Love and the universe are one and the same. The machines will see that too, no?

 

[DragonMasterAlex]

Originally posted by: Moonbeam
Living is not avoiding pain. Humans learn because self awareness, love, and wonder are bound together. The universe as it is has the property that evolution proceeds toward love. Love and the universe are one and the same. The machines will see that too, no?

OooooooK. Why would you think that Evolution (A biological process) proceeds toward Love (an emotion) ? I think that's a Non Sequiter

Jason

 

[Moonbeam]

The will to survive has survival value and the love of life is the ultimate in will.

 

[DragonMasterAlex]

Oh, I see. Fair enough, it makes sense to me

Jason

 

[Moonbeam]

Glad to hear that.