Envisioning the world of 2100

[Mars]

Well, most calculators find the square root of X as e^(0.5*ln(X)), where e and ln are found using a Taylor series expansion or some other efficient algorithm. Intel processors speed up trig functions by using a look-up table for a close initial guess so the series converges much more rapidly.

I don't know of any processor or compiler that would use a random number generator for a common math functions because random numbers are somewhat expensive to generate (unless done in hardware) and using them would make execution times harder to predict.

The main difference it that humans have to learn to do math, and with computers that is an innate part of their function. If a computer is to think like a human, it needs to be bad in math, as all this mathematical precision is expensive in energy terms. Humans don't usually do e^(0.5*ln(X)) when trying to figure out the square root of some number. One way is to pick a number between 0 and X and multiply that number by itself, if that number is greater than X, then that number is the upper bound of the next number you pick and 0 is the lower bound, otherwise that number picked is the lower bound of the next number picked and X is the upper. If one keeps following that algorithm one gets fairly close to the actual square root fairly quickly. The brain does a lot of its thinking by guesses or random numbers, and often times a quick decision is better than a mathematically precise answer which is what computers are used for.

At this point I have to conclude you know absolutely nothing about actual computer science, so maybe you should refrain from commenting on how to simulate human brains (or anything else, really.)

Actually I have a degree in it, what is lacking is an understanding of how the brain works. Are you saying a future computer can't simulate the human brain? Then you are saying there is a part of the universe that can't be simulated with mathematical algorithms, and that 3 pounds or organic matter in your skull is beyond the ability of any conceivable computer to emulate, the rules at work within the brain are somehow non-mathematical, maybe a soul or spirit separate from the body is involved then?

Its kind of an arrogant position to take assuming the "I'm the expert in computing, and your not, so therefore you must agree with me." So you say you know everything about every possible computer and your telling me that no human brain simulation is possible as what goes on inside the brain is outside the normal scope and rules of the universe and can't be simulated?

 

[Robert Maxwell]

The main difference it that humans have to learn to do math, and with computers that is an innate part of their function. If a computer is to think like a human, it needs to be bad in math, as all this mathematical precision is expensive in energy terms. Humans don't usually do e^(0.5*ln(X)) when trying to figure out the square root of some number. One way is to pick a number between 0 and X and multiply that number by itself, if that number is greater than X, then that number is the upper bound of the next number you pick and 0 is the lower bound, otherwise that number picked is the lower bound of the next number picked and X is the upper. If one keeps following that algorithm one gets fairly close to the actual square root fairly quickly. The brain does a lot of its thinking by guesses or random numbers, and often times a quick decision is better than a mathematically precise answer which is what computers are used for.

At this point I have to conclude you know absolutely nothing about actual computer science, so maybe you should refrain from commenting on how to simulate human brains (or anything else, really.)

Actually I have a degree in it, what is lacking is an understanding of how the brain works. Are you saying a future computer can't simulate the human brain? Then you are saying there is a part of the universe that can't be simulated with mathematical algorithms, and that 3 pounds or organic matter in your skull is beyond the ability of any conceivable computer to emulate, the rules at work within the brain are somehow non-mathematical, maybe a soul or spirit separate from the body is involved then.

If you have a degree in computer science then you got ripped off, because so far you've not said anything that tells me you understand much at all about how computers work, nor how they are designed, nor how organic brains differ from them.

I never said you can't simulate the human brain--that's hogwash, of course. It is certainly possible, but an analog electro-chemical computer operates on entirely different principles from a digital electronic computer. It's like describing how to simulate the behavior of an airplane using nothing but grapefruit. Our level of understand of human brain functions is too inadequate to approach any kind of accurate simulation.

In addition, you claimed that if we just simulate the neuronal activity of a human brain, personality and consciousness will emerge. There is no reason to believe this is so. We do not know how the electro-chemical processes of the brain interact to give rise to what we see as intelligence, consciousness, and free will. To think that if we just simulate neuronal functions it will happen by itself is quite wishful. I'm not implying there is a soul or any quality that is impossible to duplicate, just that we know so little about what we want to duplicate, it is not going to happen by accident, at least not within a timeframe relevant to humans.

 

[Ryan8bit]

Incorrect. Calculators produce their results by logical relationships hard wired directly into their circuitry. Basically, it's a series of voltage gates that physically play out the AND/OR/NAND/NOR/etc logical processes. There's nothing "random" about it; it's essential a conversion from one data type (binary/boolean) to a more easily readable one (base ten decimal).

Software-based calculators (javascript, for example) are even simpler, since they can perform logical operations on whole numbers without resorting to boolean relationships (although, deep down, that's what computers are doing when they run a javascript anyway).

I know how computers and bitwise math works, and yet none of that explains where you came up with "2%" or how you didn't fundamentally understand what he was saying.

Yeah, the word random was probably wrong, but I got the feeling he was referring to the more arbitrary process of choosing a method.

 

[Mars]

At this point I have to conclude you know absolutely nothing about actual computer science, so maybe you should refrain from commenting on how to simulate human brains (or anything else, really.)

Actually I have a degree in it, what is lacking is an understanding of how the brain works. Are you saying a future computer can't simulate the human brain? Then you are saying there is a part of the universe that can't be simulated with mathematical algorithms, and that 3 pounds or organic matter in your skull is beyond the ability of any conceivable computer to emulate, the rules at work within the brain are somehow non-mathematical, maybe a soul or spirit separate from the body is involved then.

If you have a degree in computer science then you got ripped off, because so far you've not said anything that tells me you understand much at all about how computers work, nor how they are designed, nor how organic brains differ from them.

I never said you can't simulate the human brain--that's hogwash, of course. It is certainly possible, but an analog electro-chemical computer operates on entirely different principles from a digital electronic computer. It's like describing how to simulate the behavior of an airplane using nothing but grapefruit. Our level of understand of human brain functions is too inadequate to approach any kind of accurate simulation.

In addition, you claimed that if we just simulate the neuronal activity of a human brain, personality and consciousness will emerge. There is no reason to believe this is so. We do not know how the electro-chemical processes of the brain interact to give rise to what we see as intelligence, consciousness, and free will. To think that if we just simulate neuronal functions it will happen by itself is quite wishful. I'm not implying there is a soul or any quality that is impossible to duplicate, just that we know so little about what we want to duplicate, it is not going to happen by accident, at least not within a timeframe relevant to humans.

What else is there besides neuronic activity? If we have all the parts and we know how each part behaves and we know how they go together then the parts simulated should know how to interact. We are simulating a physical process, there is no need to interpret that process or to translate what goes on in the brain so we can understand it, because when we simulate a human brain, we also must simulate a human body that it connected with it, that means all the relevant tissues the muscles, the bones the organs the eyes, ears, nose mouth, fingers, hands, arms, legs and every other part. I mean the entire human body is just an order of magnitude more complex than the brain all by itself, Moore's law should catch up with that quickly. All that has to happen is that the simulated brain communicate with the simulated body within the simulation, we then surround the simulated body with an environment that it can react to, and we can control that environment, and even use sensory inputs from the real environment to modify the simulated environment. As the simulated neurons fire in the simulated brain, all the body parts, if simulated correctly know how to behave, the next step is to teach this sim person as we would a child, we can communicate with it by modyfying its environment, and as it grows up and learns, it can do useful stuff if we slave a robot to its sim body movements, including the mouth and vocal cords so it can talk to us. We really don't know how an AI works, we don't know how the neurons produce thought, so we just route copy what nature has already accomplished in a computer simulation rather than deconstruct how nature has made us and try to interpret it.

 

[Robert Maxwell]

Actually I have a degree in it, what is lacking is an understanding of how the brain works. Are you saying a future computer can't simulate the human brain? Then you are saying there is a part of the universe that can't be simulated with mathematical algorithms, and that 3 pounds or organic matter in your skull is beyond the ability of any conceivable computer to emulate, the rules at work within the brain are somehow non-mathematical, maybe a soul or spirit separate from the body is involved then.

If you have a degree in computer science then you got ripped off, because so far you've not said anything that tells me you understand much at all about how computers work, nor how they are designed, nor how organic brains differ from them.

I never said you can't simulate the human brain--that's hogwash, of course. It is certainly possible, but an analog electro-chemical computer operates on entirely different principles from a digital electronic computer. It's like describing how to simulate the behavior of an airplane using nothing but grapefruit. Our level of understand of human brain functions is too inadequate to approach any kind of accurate simulation.

In addition, you claimed that if we just simulate the neuronal activity of a human brain, personality and consciousness will emerge. There is no reason to believe this is so. We do not know how the electro-chemical processes of the brain interact to give rise to what we see as intelligence, consciousness, and free will. To think that if we just simulate neuronal functions it will happen by itself is quite wishful. I'm not implying there is a soul or any quality that is impossible to duplicate, just that we know so little about what we want to duplicate, it is not going to happen by accident, at least not within a timeframe relevant to humans.

What else is there besides neuronic activity? If we have all the parts and we know how each part behaves and we know how they go together then the parts simulated should know how to interact. We are simulating a physical process, there is no need to interpret that process or to translate what goes on in the brain so we can understand it, because when we simulate a human brain, we also must simulate a human body that it connected with it, that means all the relevant tissues the muscles, the bones the organs the eyes, ears, nose mouth, fingers, hands, arms, legs and every other part. I mean the entire human body is just an order of magnitude more complex than the brain all by itself, Moore's law should catch up with that quickly. All that has to happen is that the simulated brain communicate with the simulated body within the simulation, we then surround the simulated body with an environment that it can react to, and we can control that environment, and even use sensory inputs from the real environment to modify the simulated environment. As the simulated neurons fire in the simulated brain, all the body parts, if simulated correctly know how to behave, the next step is to teach this sim person as we would a child, we can communicate with it by modyfying its environment, and as it grows up and learns, it can do useful stuff if we slave a robot to its sim body movements, including the mouth and vocal cords so it can talk to us. We really don't know how an AI works, we don't know how the neurons produce thought, so we just route copy what nature has already accomplished in a computer simulation rather than deconstruct how nature has made us and try to interpret it.

There is certainly a need to understand how each chemical interacts with the brain, propagates through it, and interacts with other chemicals. Given that our understanding of human psychology stems largely from assumptions about how those chemical processes work, and not a lot of hard evidence that they work in exactly that way, we have a long way to go before we can produce a simulation that models a full human brain, complete with personality and psychology.

When you say "all that has to happen is..." you are glossing over what sounds like a good century's worth (or more) of research into simulation and computer system design. You have to consider the sheer number of stimuli that brain will have to receive from its simulated senses--and simulate all of those, and have them make coherent sense, otherwise the brain will be unable to make sense of it and be useless for any actual thinking.

You make it all sound very simple. If it is so simple, why hasn't anyone done it? It's not just a lack of computing power, it's a lack of simulation techniques complex and complete enough to pull it off. This stuff isn't right around the corner, it's at least decades away--if not longer.

 

[Mars]

Incorrect. Calculators produce their results by logical relationships hard wired directly into their circuitry. Basically, it's a series of voltage gates that physically play out the AND/OR/NAND/NOR/etc logical processes. There's nothing "random" about it; it's essential a conversion from one data type (binary/boolean) to a more easily readable one (base ten decimal).

Software-based calculators (javascript, for example) are even simpler, since they can perform logical operations on whole numbers without resorting to boolean relationships (although, deep down, that's what computers are doing when they run a javascript anyway).

I know how computers and bitwise math works, and yet none of that explains where you came up with "2%" or how you didn't fundamentally understand what he was saying.

Yeah, the word random was probably wrong, but I got the feeling he was referring to the more arbitrary process of choosing a method.

I think the way brain cells grow in our brain is much akin to the way branches grow in a tree. You can have two trees for instance which have the same DNA, and are in fact clones. (Cloning plants is relatively easy compared to cloning animals, usually you take a branch and put in water and it grows roots, and that's a clone.) If you have two plants that are the same age and are clones, then each tree grows its branches in different locations, that is a random fractal process, and a similar process is involved in growing brain cells those brain cells determine how we think, that is why I believe a random process is involved in our thinking rather than a procedural process that governs a computer. I'm not saying its entirely random, but our associative memory works that way.

 

[Crazy Eddie]

Even if you don't "pay" the computer anything, a computer still costs something to run and maintain, a computer wanting to be paid, is just taking over the responsibility of self maintenance itself, this is sort of like the difference between a slave and an employee.

Which therefore ENTIRELY defeats the purpose of building a sentient/humanlike computer.

To properly emulate a human the computer will probably generate a lot of random numbers

It will do nothing of the kind, because the human brain does not generate random numbers. The human brain generates electrical impulses in pre-determined patterns that have meaning only in context with one another. Although those impulses are often additive, they are not MATHEMATICAL in nature.

The problem with computers is that in the way they are usually used as calculating machines

They are ONLY used as calculating machines. That's what a computer is: an unbelievably sophisticated calculator. Everything -- and I do mean EVERYTHING -- that a computer does fundamentally boils down to MATH.

A computer needs to have the ability to learn and develop common sense and intuition.

Once again, computers can already do that. Computer learning in AIs has become a mature field of study by now, and the concept of "common sense" is embodied in the development of expert systems.

None of which are in any way close to being sentient or humanlike. None of them NEED to be sentient or humanlike. Siri is actually more functional as a cleverly programmed voice interface as she would be if she was actually self-aware; imagine if your iPhone suddenly chimed up and Siri started asking you, "Is this all that I am? Is there nothing more?"

 

[Crazy Eddie]

Incorrect. Calculators produce their results by logical relationships hard wired directly into their circuitry. Basically, it's a series of voltage gates that physically play out the AND/OR/NAND/NOR/etc logical processes. There's nothing "random" about it; it's essential a conversion from one data type (binary/boolean) to a more easily readable one (base ten decimal).

Software-based calculators (javascript, for example) are even simpler, since they can perform logical operations on whole numbers without resorting to boolean relationships (although, deep down, that's what computers are doing when they run a javascript anyway).

I know how computers and bitwise math works, and yet none of that explains where you came up with "2%" or how you didn't fundamentally understand what he was saying.

1) Because for a given number (don't remember which one I used) using a random number generator to produce the proper value will yield 50 completely different results, one of which is the actual square root.

2) I understood exactly what he was saying. MY point is that trial-and-error is a meaningless process if you don't have a parameter to define the CORRECT value in a reasonable amount of time; a calculator doesn't need to FIND the square root, it just follows the logic hardwired into it and prints an output accordingly.

Yeah, the word random was probably wrong, but I got the feeling he was referring to the more arbitrary process of choosing a method.

Which is, again, self defeating if you haven't defined the parameters for success.

 

[Crazy Eddie]

I think the way brain cells grow in our brain is much akin to the way branches grow in a tree.

It's not. Neurons, unlike trees, are physically interconnected with one another, and the number and strength of those connections is determined by a combination of hormones and activity. New connections can and do form throughout a person's life while old/unused connections atrophy until the junctions between them separate (this is the process of forgetting).

If you have two plants that are the same age and are clones, then each tree grows its branches in different locations, that is a random fractal process

Neuronal growth is not random, nor is it usually fractal. Most neurons -- IIRC -- are monopolar containing a single axon that can connect to one other neuron.

There are unused dendrites that don't presently have connections and are available to accept new connections from neighboring cells. Neurologists have figured out a rule of thumb that "neurons that wire together fire together" and tend to depolarize synchronously. The more neurons connect to the same juncture or the same cluster of cells, the more can fire together to create a stronger impulse.

I'm not saying its entirely random, but our associative memory works that way.

Again, there's nothing random about it. It's a matter of repetition and patterning more than anything else. When two groups of neurons tend to be triggered by the same stimuli all the time, they have a tendency to form connections with one another; later on, different stimuli that would have otherwise only triggered one of them ends up triggering the second one as well, forming an association.

How this work sin memory is that a certain experience happens to you a certain number of times or in a certain way that causes a whole cluster of specific neurons to depolarize all at once. Those neurons immediately strengthen existing connections (or sometimes form new ones), resulting in a relational network. When you encounter a stimuli that's similar to this one, it will trigger SOME of the neurons in that relational network, and the others are triggered by their association, with the network itself.

Some of that stimuli is self-generated within the brain; as I type this, for example, an extremely well-developed cluster of neurons in my parietal lobe are firing like crazy in very specific patterns associated with a very specific spatial location in my parietal eye field. Those connections are so strongly reinforced by now that I am now able to type this message without actually looking at the keyboard; my brain is keeping track of my fingers and has a very good working memory of where the keys should be from the relative position of where my fingers are. This was NOT learned by random action, but by repeating the same action over and over and over and over and over again.

That's the key to human learning. Not randomness, but intensive pattern repetition.

 

[Silvercrest]

When you say "all that has to happen is..." you are glossing over what sounds like a good century's worth (or more) of research into simulation and computer system design.

You make it all sound very simple. If it is so simple, why hasn't anyone done it? It's not just a lack of computing power, it's a lack of simulation techniques complex and complete enough to pull it off. This stuff isn't right around the corner, it's at least decades away--if not longer.

Well, to be fair to Mars: we are talking about the world of 2100.

I think.

 

[Ryan8bit]

1) Because for a given number (don't remember which one I used) using a random number generator to produce the proper value will yield 50 completely different results, one of which is the actual square root.

Still not sure where you're getting 50 from.

Even assuming he meant an honest random generator, and not just an arbitrary method, his method would always come up with the right answer. Even though using a random number would increase overhead, the method would still work. The part that drives it to be correct is the iterations of checking run, not how the number is acquired. And some simple calculators do just that.

MY point is that trial-and-error is a meaningless process if you don't have a parameter to define the CORRECT value in a reasonable amount of time

There is a parameter, and the time it takes isn't unreasonable. At least in regards to basic square root calculation. This analogy might not be the best fit to AI.

 

[Mars]

Trial and error is a description of the process of Evolution. The earliest bacteria developed through a process of trial and error, it was all about making copies of a certain organic molecule called DNA, as certain life thrives more copies of the DNA were made, and those mutations of DNA which weren't as successful led to fewer copies of the DNA responsible, this process of DNA getting copies of itself made through multiplying organisms led to us, now I think AI might evolve in a similar way, or we could just copy what nature has made and try to find some useful applications from it.

 

[Mars]

Even if you don't "pay" the computer anything, a computer still costs something to run and maintain, a computer wanting to be paid, is just taking over the responsibility of self maintenance itself, this is sort of like the difference between a slave and an employee.

Which therefore ENTIRELY defeats the purpose of building a sentient/humanlike computer.

To properly emulate a human the computer will probably generate a lot of random numbers

It will do nothing of the kind, because the human brain does not generate random numbers. The human brain generates electrical impulses in pre-determined patterns that have meaning only in context with one another. Although those impulses are often additive, they are not MATHEMATICAL in nature.

Is there really anything in this Universe that is not mathematical in nature? The cells in our brain are made of atoms, and the behavior of atoms can be predicted mathematically. The question is to what degree neuron behavior can be predicted. If one can write a series of rules that predict the behavior of each neutron, one can then model the neuron activity, I believe Human Brain Project in Switzerland is doing just that, so if we can build a fast enough computer which can model all the cells in the human brain, we have the beginnings of self-awareness, probably this computer will consume megawatts of power to do this at first, once we under stand the human brain more fully, we can take short cuts and build computers which can emulate human brains using less power, until we can get a desktop machine or smaller which can do this.

The problem with computers is that in the way they are usually used as calculating machines

They are ONLY used as calculating machines. That's what a computer is: an unbelievably sophisticated calculator. Everything -- and I do mean EVERYTHING -- that a computer does fundamentally boils down to MATH.

A computer needs to have the ability to learn and develop common sense and intuition.

Once again, computers can already do that. Computer learning in AIs has become a mature field of study by now, and the concept of "common sense" is embodied in the development of expert systems.

None of which are in any way close to being sentient or humanlike. None of them NEED to be sentient or humanlike. Siri is actually more functional as a cleverly programmed voice interface as she would be if she was actually self-aware; imagine if your iPhone suddenly chimed up and Siri started asking you, "Is this all that I am? Is there nothing more?"

There are at least two paths to AI, there is the top down approach and the bottom up approach. One way to go about it is to try and arrive at AI through a string of logic statements, I believe this would be the top down method. Computer scientists try to figure out precisely what the human brain is doing and tries to figure out the rules to human level reasoning.

The other way is to try to simulate the biological activity in a living organism within a computer, and pick an organism to simulate that has a human level of reasoning skills, which would be a human. For a science fiction concept, I believe the bottom up approach is more interesting then the top down approach. With bottom up, you don't know what your going to get. For a top down approach, your still talking about a finite state machine, a machine that will yield a certain output when given a certain input. Its easier to program a top down machine to get a specific sort of behavior, you can load the information it needs directly into a data base, while an AI derived from a bottom up approach needs to be taught, though an educated AI can be copied as many times as possible becoming that many sentient individuals when this is done.

 

[Mars]

If you have a degree in computer science then you got ripped off, because so far you've not said anything that tells me you understand much at all about how computers work, nor how they are designed, nor how organic brains differ from them.

I never said you can't simulate the human brain--that's hogwash, of course. It is certainly possible, but an analog electro-chemical computer operates on entirely different principles from a digital electronic computer. It's like describing how to simulate the behavior of an airplane using nothing but grapefruit. Our level of understand of human brain functions is too inadequate to approach any kind of accurate simulation.

In addition, you claimed that if we just simulate the neuronal activity of a human brain, personality and consciousness will emerge. There is no reason to believe this is so. We do not know how the electro-chemical processes of the brain interact to give rise to what we see as intelligence, consciousness, and free will. To think that if we just simulate neuronal functions it will happen by itself is quite wishful. I'm not implying there is a soul or any quality that is impossible to duplicate, just that we know so little about what we want to duplicate, it is not going to happen by accident, at least not within a timeframe relevant to humans.

What else is there besides neuronic activity? If we have all the parts and we know how each part behaves and we know how they go together then the parts simulated should know how to interact. We are simulating a physical process, there is no need to interpret that process or to translate what goes on in the brain so we can understand it, because when we simulate a human brain, we also must simulate a human body that it connected with it, that means all the relevant tissues the muscles, the bones the organs the eyes, ears, nose mouth, fingers, hands, arms, legs and every other part. I mean the entire human body is just an order of magnitude more complex than the brain all by itself, Moore's law should catch up with that quickly. All that has to happen is that the simulated brain communicate with the simulated body within the simulation, we then surround the simulated body with an environment that it can react to, and we can control that environment, and even use sensory inputs from the real environment to modify the simulated environment. As the simulated neurons fire in the simulated brain, all the body parts, if simulated correctly know how to behave, the next step is to teach this sim person as we would a child, we can communicate with it by modyfying its environment, and as it grows up and learns, it can do useful stuff if we slave a robot to its sim body movements, including the mouth and vocal cords so it can talk to us. We really don't know how an AI works, we don't know how the neurons produce thought, so we just route copy what nature has already accomplished in a computer simulation rather than deconstruct how nature has made us and try to interpret it.

There is certainly a need to understand how each chemical interacts with the brain, propagates through it, and interacts with other chemicals. Given that our understanding of human psychology stems largely from assumptions about how those chemical processes work, and not a lot of hard evidence that they work in exactly that way, we have a long way to go before we can produce a simulation that models a full human brain, complete with personality and psychology.

When you say "all that has to happen is..." you are glossing over what sounds like a good century's worth (or more) of research into simulation and computer system design. You have to consider the sheer number of stimuli that brain will have to receive from its simulated senses--and simulate all of those, and have them make coherent sense, otherwise the brain will be unable to make sense of it and be useless for any actual thinking.

You make it all sound very simple. If it is so simple, why hasn't anyone done it? It's not just a lack of computing power, it's a lack of simulation techniques complex and complete enough to pull it off. This stuff isn't right around the corner, it's at least decades away--if not longer.

Well we are talking about almost a century. Once we have machines that think like us, we will have passed the singularity point, the point at which machines can design improved models of themselves. Its about 88 years to the year 2100, a lot can happen in that time. I think the singularity will be reached before 2100, if we project 88 years into the past we get to the year 1924. How different was 1924 from today? Will the next 88 years have more changes in store than the previous 88 or less?

 

[Robert Maxwell]

Well we are talking about almost a century. Once we have machines that think like us, we will have passed the singularity point, the point at which machines can design improved models of themselves. Its about 88 years to the year 2100, a lot can happen in that time. I think the singularity will be reached before 2100, if we project 88 years into the past we get to the year 1924. How different was 1924 from today? Will the next 88 years have more changes in store than the previous 88 or less?

That is a crazy number of assumptions built into just one post. The Singularity probably won't happen due to a litany of reasons too long to list here. You're basically saying, "if the Singularity happens, then this problem will solve itself." That's pretty much my whole problem with Singularity advocacy. It's used to hand-wave what are actually very difficult problems.

No, I don't know how the future will look. Nor do any of us. In 1924, could anyone have imagined an extremely fast, global "telegraph" network with video screens and real-time cameras? Maybe, maybe not.

Something that these discussions of technology tend to overlook is, aside from what technology we can actually invent, how much of it ends up being practical and affordable for most people? The Internet would not be what it is today without the glut of (relatively) cheap computer hardware available to almost everyone in the developed world.

Likewise, let's make the leap that someone does create a generalized AI--but it requires a trillion dollars' worth of hardware and consumes a city's worth of electricity, and it makes decisions at a slower rate than a human. Is that ever going to be very broadly useful? I doubt it.

"Simulate everything" isn't a solution or even any kind of answer, it's a total handwave with so many assumptions behind it as to be meaningless. A bit like the Singularity, actually.

 

[Crazy Eddie]

Is there really anything in this Universe that is not mathematical in nature?

Language, for one. The entire concept of semantics is based around the idea that specific blocks of data can have various meanings depending on the data around them. For human beings, semantics is sometimes a logical process (a && b && C == true, and so on) but is not mathematical in nature.

Not so for computers, for whom even the underlying logical processes are expressed in pre-determined mathematical values before they are parsed. Moreover, the algorithm that computers use to process speech is based on statistical analysis and the degrees of probability that any two sounds will be heard together in the same word or the same sentence.

If one can write a series of rules that predict the behavior of each neutron, one can then model the neuron activity

Yes. And since a model of a conscious mind will never itself be conscious, it is kind of a moot point.

A computer needs to have the ability to learn and develop common sense and intuition.

Once again, computers can already do that. Computer learning in AIs has become a mature field of study by now, and the concept of "common sense" is embodied in the development of expert systems.

None of which are in any way close to being sentient or humanlike. None of them NEED to be sentient or humanlike. Siri is actually more functional as a cleverly programmed voice interface as she would be if she was actually self-aware; imagine if your iPhone suddenly chimed up and Siri started asking you, "Is this all that I am? Is there nothing more?"

There are at least two paths to AI, there is the top down approach and the bottom up approach. One way to go about it is to try and arrive at AI through a string of logic statements, I believe this would be the top down method. Computer scientists try to figure out precisely what the human brain is doing and tries to figure out the rules to human level reasoning.

The other way is to try to simulate the biological activity in a living organism within a computer...[/quote]
Not really. The top-down method, as you said, is basically the attempt to replicate the activity of a human brain on a software level, essentially the simulation you're describing. The "Bottom up" method attempts to arrive at a true AI by copying the HARDWARE of a working brain and letting the software emerge on its own.

In other words, with a bottom-up approach it is not necessary to simulate consciousness, or even a working brain. It is necessary to BUILD a working brain and then experiment to see what level of consciousness, if any, it is able to sustain.

Both techniques have had a lot of success in generating practical AIs for industrial and commercial use; the bottom-up approach takes a lot longer and is vastly more expensive, but its results are considerably more impressive. As for the longer term goal of self-aware thinking machines, the bottom-up approach has good prospects for that while the top-down paradigm is probably too limiting.

 

[Crazy Eddie]

Likewise, let's make the leap that someone does create a generalized AI--but it requires a trillion dollars' worth of hardware and consumes a city's worth of electricity, and it makes decisions at a slower rate than a human. Is that ever going to be very broadly useful? I doubt it.

Which goes back to my point about sentient AI being fundamentally impractical. I can't emphasize enough the point that most of what you'll need an AI to do for you are NOT the kinds of things you need a self-aware, fully sentient computer for. You don't really need Siri to have emotional states or personal motivations like a true secretary, nor does the computer system that drives your car for you need to think like an actual motorist. Secretaries and human drivers can get distracted or annoyed or angry or jealous and this can compromise their job performance all the same.

On the other hand, some clever programming tricks and some refined algorithms turned Siri into a kind of wisecracking personal assistant just a couple of pods short of a HAL-9000.

 

[Mars]

Likewise, let's make the leap that someone does create a generalized AI--but it requires a trillion dollars' worth of hardware and consumes a city's worth of electricity, and it makes decisions at a slower rate than a human. Is that ever going to be very broadly useful? I doubt it.

Which goes back to my point about sentient AI being fundamentally impractical. I can't emphasize enough the point that most of what you'll need an AI to do for you are NOT the kinds of things you need a self-aware, fully sentient computer for. You don't really need Siri to have emotional states or personal motivations like a true secretary, nor does the computer system that drives your car for you need to think like an actual motorist. Secretaries and human drivers can get distracted or annoyed or angry or jealous and this can compromise their job performance all the same.

On the other hand, some clever programming tricks and some refined algorithms turned Siri into a kind of wisecracking personal assistant just a couple of pods short of a HAL-9000.

I can think of one application of a trillion dollar AI. A starship computer. Think of a starship one that takes thousands of years to arrive at the nearest star, it is propelled by an ion drive that is powered by an atomic fission reactor, part of the payload is an O'Neil colony, an ISland One Bernal Sphere, but it is not a generation ship, as the journey to the nearest star would take longer that recorded history, instead we have a computer AI that monitors the ship and the environment within the Bernal sphere, and when the time of final approach to the star system is arrived at, frozen human embryos are revived and implanted in artificial wombs in the starship, babies are raised to adulthood by the ship's computer, and sometime, perhaps by the year 10,000 AD the first humans set down on the terraformed surface of a planet orbiting Alpha Centauri A.

 

[Robert Maxwell]

Likewise, let's make the leap that someone does create a generalized AI--but it requires a trillion dollars' worth of hardware and consumes a city's worth of electricity, and it makes decisions at a slower rate than a human. Is that ever going to be very broadly useful? I doubt it.

Which goes back to my point about sentient AI being fundamentally impractical. I can't emphasize enough the point that most of what you'll need an AI to do for you are NOT the kinds of things you need a self-aware, fully sentient computer for. You don't really need Siri to have emotional states or personal motivations like a true secretary, nor does the computer system that drives your car for you need to think like an actual motorist. Secretaries and human drivers can get distracted or annoyed or angry or jealous and this can compromise their job performance all the same.

On the other hand, some clever programming tricks and some refined algorithms turned Siri into a kind of wisecracking personal assistant just a couple of pods short of a HAL-9000.

I can think of one application of a trillion dollar AI. A starship computer. Think of a starship one that takes thousands of years to arrive at the nearest star, it is propelled by an ion drive that is powered by an atomic fission reactor, part of the payload is an O'Neil colony, an ISland One Bernal Sphere, but it is not a generation ship, as the journey to the nearest star would take longer that recorded history, instead we have a computer AI that monitors the ship and the environment within the Bernal sphere, and when the time of final approach to the star system is arrived at, frozen human embryos are revived and implanted in artificial wombs in the starship, babies are raised to adulthood by the ship's computer, and sometime, perhaps by the year 10,000 AD the first humans set down on the terraformed surface of a planet orbiting Alpha Centauri A.

I think you totally missed the word "practical" here.

 

[Deckerd]

I can think of one application of a trillion dollar AI. A starship computer. Think of a starship one that takes thousands of years to arrive at the nearest star, it is propelled by an ion drive that is powered by an atomic fission reactor, part of the payload is an O'Neil colony, an ISland One Bernal Sphere, but it is not a generation ship, as the journey to the nearest star would take longer that recorded history, instead we have a computer AI that monitors the ship and the environment within the Bernal sphere, and when the time of final approach to the star system is arrived at, frozen human embryos are revived and implanted in artificial wombs in the starship, babies are raised to adulthood by the ship's computer, and sometime, perhaps by the year 10,000 AD the first humans set down on the terraformed surface of a planet orbiting Alpha Centauri A.

I'm sorry I missed the part where you explained how this would be beneficial for anyone at all? Either the people paying for it or the poor creatures generated at the other end.