Some thoughts on overconfidence in the Singularity, from Jaron Lanier.
84 thoughts on “The First Church Of Robotics”
Who’s “hanging on”? I come here all the time. Unlike you, our host has a mind open to all reasonable argument.
Besides, you called _me_, girlfriend. I wasn’t even talking to you at 6:59 a.m. today when you squeezed your womanly frame in between me and the person with whom I was conversing. You sought me out, asked my opinion, then insulted me when I replied. As they say on 4ch, LOL WUT?
Your fascination with me, while flattering, is, I suspect, becoming a bit tiresome for the others who read these comments. I therefore offer you the hospitality of my private e-mail (plybel@gmail.com); if you wish to further pursue my attention, kindly do so by that means.
You believe machines are necessarily deterministic.
All physical things. Ok, so adding a random number generator makes it non-deterministic…. whoa, not so fast. That’s why they call them pseudo random… they are in fact deterministic.
Ah, but what about truly random events? I don’t know. Are they truly random? What exactly does that mean? Are they at some level not perceived, pseudo random as well? I truly don’t know.
I understand and would like to say I agree completely. The problem is I believe free will means I get to choose. If I’m just a complex random number generator then choosing is an illusion. Perhaps it is. I choose not to believe that. I could be wrong. If I am wrong, then life is just one big pachinko machine and has no point. We don’t really choose anything after all. There is no good or evil… just the bouncing ball.
First of all, why assume a hard decision you take is the result of random chance? I’m sure you agree that your decision depends on your tastes–your “inate” and nurtured sense of ethics, aesthetics, etc.–previous decisions, and circumstances you face and the knowledge you have of them. It is highly unlikely that the probability of you making a particular choice is evenly distributed. I doubt that the distribution itself is even constant. More importantly, assigning a stochastic quality to decisionmaking explains “indecision” better than anything else I’ve ever seen.
From where I sit, this explanation doesn’t detract from the power of free will. It confirms it.
I think you may have just put the finger on the problem.
What problem? I’d think AI researchers are free to define the boundaries of their inquiry.
A.I. The A stands for artificial.
And the Fed in Fedex stands for Federal, though I doubt we’d argue that Fedex is a federal agency. This, of course, is wordplay, which has little to no bearing on what AI research is about, is accomplishing, or can accomplish.
No. The question is multipart. Can we define intelligence (write it down, not point to it.)
Yes, we can. I just did. In AI, this is a trivial point, and the only question is whether or not an operating definition is useful for our purposes. If the philosophers want to debate whether a machine is or isn’t intelligent, that’s well and good; it’s not an argument that attracts a lot of attention in the working community for the reasons I described earlier.
Can we produce it in a machine (software and hardware, not biological which already has intelligence… you pointed at it, remember?)
Based on your definition of machine, the answer is we don’t know. That’s why we’re doing this work in the first place.
Artificial means something but is perhaps not in vogue anymore. Saying a baby is A.I. has got to be the ultimate example of moving the goal posts.
AI isn’t a business requirement. It’s an active area of research. There are no goalposts to be moved.
Then you don’t understand Hawking or Quantum physics. Reconsider, local phenomena that may be predictable for an unspecifiable time.
I don’t claim to understand Hawking’s personally, but I’m comfortable with my grasp on QM. And once again, I’m not about debate something Hawking said extemporaneously for a popular audience, especially when the quote is provided without citation. That’s neither here nor there. Hawking certainly doesn’t believe that the universe within the classical limit is lawless. His entire work presupposes that it is lawful within that domain.
why assume a hard decision you take is the result of random chance?
It either is or it isn’t. I don’t as a matter of fact. What does hard have to do with anything?
So what is a decision and where does it come from? Answer that and I can build you an A.I. It’s not the simple question you may imagine it to be.
His entire work presupposes that it is lawful within that domain.
Yes, for an unspecified time at which no prediction can be made. This is consistent with everything I’ve ever read about QM. I’ve also read that nobody understands QM and anybody that says they do is a liar. I think that was Feynman.
There are no goalposts to be moved.
After reconsideration of this statement I realize there is no hope.
It either is or it isn’t. I don’t as a matter of fact. What does hard have to do with anything?
That’s not even close to true. There’s any number of distributions that might characterize any choice before you; the uniform case is simply the most trivial. As for relevance, you presented this demonstrably false dichotomy in the first place.
So what is a decision and where does it come from? Answer that and I can build you an A.I. It’s not the simple question you may imagine it to be.
Well, neuroscientists are pretty sure its in the frontal lobe.
Yes, for an unspecified time at which no prediction can be made.
One, you’re garbling your own quote. You previously stated Hawking believed that the universe was unlawful outside of a local region of space for an unspecified slice of time. Two, the answer is no, the universe is lawful within a general region of space-time. This is general relativity applied to cosmology, specifically the FLRW metric.
This is consistent with everything I’ve ever read about QM.
Which is why I said the universe is lawful within the classical limit. Even beyond that, it is certainly lawful in the semiclassical sense. Hence, QM. The difference between QM and GR is that QM presents statistical laws and GR does not.
I’ve also read that nobody understands QM and anybody that says they do is a liar. I think that was Feynman.
Feynman was making a philosophical point about interpreting the formalism of quantum mechanics. Is a wave function real? Are non-local interactions real? Is the indeterminacy of position and momentum below the Planck scale real? These are philosophical points that motivate interpretations of QM’s laws. Nevertheless, the laws are very well understood, well enough for its obvious application in our everyday life. That clicker isn’t magic, you know.
You previously stated Hawking believed that the universe was unlawful outside of a local region of space for an unspecified slice of time.
Did I? Let’s look again…
“…the universe isn’t lawful, never will be lawful, never can be lawful … at best we study local phenomena that may be predictable for an unspecifiable time.”
Sooo….
“unlawful outside of a local region of space” = “local phenomena that may be predictable”
I’ll have to leave that for others to decide.
He said the “universe isn’t lawful” which you change to “unlawful outside” which I suppose you could do? Except he’s not really saying, “only in safe places where we never really see it so everything’s alright and we can all go back to sleep now.”
He’s saying anything is possible and don’t have too much faith in your ability to predict anything. Because one day it’s going to bite you in the ass.
That’s not even close to true.
I say something is either random or it’s not. So what’s my third choice? If you say partly random I’m going to reach right through this internet and grab you by the neck.
So what’s my third choice?
Any of these, and countably infinite others and their variations.
Ya did! You should be feeling my boney fingers reaching for your neck right about now…
Ok. Stepping back a bit…
It is highly unlikely that the probability of you making a particular choice is evenly distributed.
This is making an assumption about my argument that really misses the point. The real question is are we just a meat machine? If so, making a copy shouldn’t (in theory) be such a big deal and will happen eventually. You are making the assumption, a priori, that we simply are meat machines. So it follows there is a probability distribution.
But I’m saying we are not meat machines… well, perhaps mostly we are, but with just the tiniest bit of something more. So you’re probability distribution has no place in the argument. It’s a non sequitur.
Are we meat machines or something more? The answer has consequences. Meat machines do not have free will. It’s just input/output. Choice is a complete illusion. A perfect copy kept in the exact same environment (to cancel out the butterfly effect) would follow an identical path. Free will and choice simply would not exist.
This I have a problem with.
This is getting tiresome, so let’s get to the point.
Meat machines do not have free will. It’s just input/output.
Says who?
I made my argument in the last paragraph. Feel free to refute it… accept it… ignore it… whatever.
That’s easy. You won’t defend the premise that “meat machines” necessarily lack free will.
Essentially, the problem we’re having is with the imprecision your use of terms you either appropriate or make up. Your argument boils down to “AI is bunk because ‘real’ humans have this thing called free will that we can’t sense or measure.” When I try to nail you down on a specific definition, it changes. First it was the capacity to make a choice in opposition to determinism. I present to you the example of a probabilistic computer. You then deride its indeterminacy. So which is it? Is free will deterministic and illusion, or indeterministic and troubling to you?
I haven’t changed anything. Free will is the ability to choose.
Machines don’t choose, they only follow the path they are put upon. That doesn’t change if you make it non deterministic with a random number generator. It doesn’t change no matter how analytical the program or what parameters are considered. Only a buffoon would say a computer has free will. It doesn’t. It doesn’t think. It’s choices are programmed into it.
Mine are not. I do think. I do choose. I am not just a machine.
So if I’m reading you right, machines don’t and can’t have this thing free will because by definition free will is something only intelligent agents in nature have.
It used to be thought that if a computer could play chess it would be thinking. Then they did it and realized what they’d come up did not involve thinking (except by the programmer that wrote the code.)
I have free will. Machines don’t. Refute that.
by definition free will is something only intelligent agents in nature have.
Hey I’m only pointing my finger at it… like you did to define intelligence.
Well the easiest refutation would be to point out your lack of verifiable evidence for whatever it is that you call free will and consign to the sole province of Man.
Hey I’m only pointing my finger at it… like you did to define intelligence.
Actually, you’re pointing your finger at something you figure is a feature of free will, and then pointing at something else when shown a non-human example of the same thing. That is the very essence of shifting the goalposts.
Actually, you’re pointing your finger at something you figure is a feature of free will, and then pointing at something else when shown a non-human example of the same thing.
What??? I think you need to up your meds.
I’m pointing my finger at the same baby you did.
You say it is the definition of intelligence. I say it has intelligence and you haven’t defined anything at all.
I say the baby has free will. A machine does not. I have no idea what you’re saying.
My goalposts are firmly set. Showing that a machine does not have intelligence or free will doesn’t move the posts at all.
Try again.
shown a non-human example of the same thing
What non-human example?
Well the easiest refutation would be to point out your lack of verifiable evidence for whatever it is that you call free will
Let me get this straight… are you saying free will doesn’t exist???
What??? I think you need to up your meds.
Says the guy defending the supernatural.
I’m pointing my finger at the same baby you did.
No, you’re not. You’re pointing your finger at something that supercedes nature entirely.
You say it is the definition of intelligence. I say it has intelligence and you haven’t defined anything at all.
On the contrary, I said intelligence is some collection of traits associated with this child. I did not exhaustive go beyond that since our argument at the time was whether or not intelligence could be defined, and if so to your satisfaction. Clearly, that was an impossible task.
I say the baby has free will. A machine does not. I have no idea what you’re saying.
Very well, given the only definition of free will you’ve offered so far, all I need do is show you a machine that evolves its states stochastically to suffice. Very well, I choose a slot machine.
My goalposts are firmly set. Showing that a machine does not have intelligence or free will doesn’t move the posts at all.
No, your goal posts will move again. First it’s to ability to make a choice. I wonder what condition you will impose next?
What non-human example?
Your computer chess player.
Let me get this straight… are you saying free will doesn’t exist???
Not at all. The brain is an extraordinarily complicated, numerous ensemble, and I’d find highly unlikely such an ensemble reduces to something determinate (and therefore causal) and acyclic (and therefore incapable of introspection and control). However bounded by self-awareness, instinctual and nurtured values, etc., the brain can still exercise considerable freedom of action faced with any given choice.
Says the guy defending the supernatural.
You’re hallucinating. When did I defend supernatural? What I said was, “Supernatural is a misnomer. … which is just natural but beyond our normal experience.” In other words, if anything can be identified as supernatural it’s not some kind of magic, it’s just something natural that we don’t understand.
Arthur C. Clarke made a law of it. Perhaps you’re familiar?
Clearly, that was an impossible task.
Clearly, because if you could exhaustively define it, we should be able to reproduce it. Essential elements of it are clearly undefined.
all I need do is show you a machine that evolves its states stochastically to suffice. Very well, I choose a slot machine.
You’re saying a slot machine has free will and is an example of thought. Sorry, I (and all those that are sane) must disagree.
Your computer chess player.
…was a comparison to show contrast. How is this difficult to understand and what has this to do with moving goal posts?
You’re hallucinating. When did I defend supernatural? What I said was, “Supernatural is a misnomer. … which is just natural but beyond our normal experience.” In other words, if anything can be identified as supernatural it’s not some kind of magic, it’s just something natural that we don’t understand.
Nah, you don’t get a pass for redefining natural to include the supernatural.
Arthur C. Clarke made a law of it. Perhaps you’re familiar?
No, I’m not. I’m familiar with his dictum “any sufficiently advanced technology is indistinguishable from magic.” This is a haughty restatement of the principle that scientific knowledge does not claim completeness. To defend something inherently out of the reach of scientific understanding is to defend the supernatural.
Clearly, because if you could exhaustively define it, we should be able to reproduce it. Essential elements of it are clearly undefined.
Why? We can exhaustively define stars, but we can’t presently reproduce them. That’s not to say that a machine with a degree of intelligence we’d assign to a human being is that far out of reach.
You’re saying a slot machine has free will and is an example of thought. Sorry, I (and all those that are sane) must disagree.
I’m saying a slot machine has free will. Who cares whether or not it thinks?
…was a comparison to show contrast.
Through incredulity. Yet this computer chess player must be able to consider its decisions, your decisions, and when presented with a choice between two or more equally desirable moves must still make a decision. Within the context of the game, only chauvinism denies that the behavior exhibited is distinct from self-awareness or free will. The only thing missing is choice in goal orientation, something easily rectified by implementing a checkers extension, a system that encodes some rudimentary values, and a random number generator to permits evolution of those values and choice between the two games. You need look no farther than a MMOG with non-player characters to find a system that does this and more.
How is this difficult to understand…
It’s not. Your definition evolves when presented with a machine that meets it.
…and what has this to do with moving goal posts?
It only serves to show that you’re about to move them again.
We can exhaustively define stars, but we can’t presently reproduce them.
In theory we can so you are wrong. In theory, we are not able to create an electronic brain. It’s people in the A.I. community that have been moving the goal posts because of all the times they embarrassed themselves. Saying artificial has no meaning is just one example. ‘Electronic brain’ being another.
Nowhere did I redefine natural. What I said was if you identify something as beyond natural is more likely a case of misunderstanding. I suspect, in your case, you misunderstand on purpose.
I’m saying a slot machine has free will.
Thus endeth the argument.
In theory we can so you are wrong.
Seriously, those goalposts must have a million frequent flier miles seeing how often you’re moving them.
In theory, we are not able to create an electronic brain.
It’s people in the A.I. community that have been moving the goal posts because of all the times they embarrassed themselves.
Exactly when have they ever moved any goalposts?
Saying artificial has no meaning is just one example.
One that apparently exists only in your imagination.
‘Electronic brain’ being another.
Which, as pointed out here is apparently well beyond “reproducing…in theory.”
Nowhere did I redefine natural. What I said was if you identify something as beyond natural is more likely a case of misunderstanding.
And yet you consistently point to some qualia concerning human intelligence and free will that is beyond natural understanding.
I suspect, in your case, you misunderstand on purpose.
I’m certain you do.
Good day sir… I said good day.
… [a slot machine has free will.]
Jaron Lanier? Really? Getting his opinion on transhumanist topics is like asking a satanists opinion of the Pope.
Yet our host felt the article was worth linking to Mike. Even a satanist could have some worthwhile insights regarding the pope. Are there any points in the article you disagree with?
It’s people in the A.I. community that have been moving the goal posts because of all the times they embarrassed themselves.
I don’t know that it’s goalpost moving as collapsing expectations. Early on, they thought that things like general problem solving algorithms, theorem proving, and “common sense” relational knowledge would in themselves achieve genuine artificial intelligence. Later it was neural networks, genetic algorithms, and similar things. Now, I think it’s down to finding algorithms for data mining and knowledge bases which probably is where the expectations will linger for a while until someone gets something that works above that level.
I don’t know that it’s goalpost moving as collapsing expectations.
Perhaps so. They’ve had a lot of cold water splashed in there face by reality. Even today they make claims (giving us a date for the singularity… please) that they probably hope we forget when the time comes. They insist it’s just a matter of getting the transistor count up enough as if complexity were the only question… it’s an article of faith.
Who’s “hanging on”? I come here all the time. Unlike you, our host has a mind open to all reasonable argument.
Besides, you called _me_, girlfriend. I wasn’t even talking to you at 6:59 a.m. today when you squeezed your womanly frame in between me and the person with whom I was conversing. You sought me out, asked my opinion, then insulted me when I replied. As they say on 4ch, LOL WUT?
Your fascination with me, while flattering, is, I suspect, becoming a bit tiresome for the others who read these comments. I therefore offer you the hospitality of my private e-mail (plybel@gmail.com); if you wish to further pursue my attention, kindly do so by that means.
You believe machines are necessarily deterministic.
All physical things. Ok, so adding a random number generator makes it non-deterministic…. whoa, not so fast. That’s why they call them pseudo random… they are in fact deterministic.
Ah, but what about truly random events? I don’t know. Are they truly random? What exactly does that mean? Are they at some level not perceived, pseudo random as well? I truly don’t know.
First of all, why assume a hard decision you take is the result of random chance? I’m sure you agree that your decision depends on your tastes–your “inate” and nurtured sense of ethics, aesthetics, etc.–previous decisions, and circumstances you face and the knowledge you have of them. It is highly unlikely that the probability of you making a particular choice is evenly distributed. I doubt that the distribution itself is even constant. More importantly, assigning a stochastic quality to decisionmaking explains “indecision” better than anything else I’ve ever seen.
From where I sit, this explanation doesn’t detract from the power of free will. It confirms it.
What problem? I’d think AI researchers are free to define the boundaries of their inquiry.
And the Fed in Fedex stands for Federal, though I doubt we’d argue that Fedex is a federal agency. This, of course, is wordplay, which has little to no bearing on what AI research is about, is accomplishing, or can accomplish.
Yes, we can. I just did. In AI, this is a trivial point, and the only question is whether or not an operating definition is useful for our purposes. If the philosophers want to debate whether a machine is or isn’t intelligent, that’s well and good; it’s not an argument that attracts a lot of attention in the working community for the reasons I described earlier.
Based on your definition of machine, the answer is we don’t know. That’s why we’re doing this work in the first place.
AI isn’t a business requirement. It’s an active area of research. There are no goalposts to be moved.
I don’t claim to understand Hawking’s personally, but I’m comfortable with my grasp on QM. And once again, I’m not about debate something Hawking said extemporaneously for a popular audience, especially when the quote is provided without citation. That’s neither here nor there. Hawking certainly doesn’t believe that the universe within the classical limit is lawless. His entire work presupposes that it is lawful within that domain.
why assume a hard decision you take is the result of random chance?
It either is or it isn’t. I don’t as a matter of fact. What does hard have to do with anything?
So what is a decision and where does it come from? Answer that and I can build you an A.I. It’s not the simple question you may imagine it to be.
His entire work presupposes that it is lawful within that domain.
Yes, for an unspecified time at which no prediction can be made. This is consistent with everything I’ve ever read about QM. I’ve also read that nobody understands QM and anybody that says they do is a liar. I think that was Feynman.
There are no goalposts to be moved.
After reconsideration of this statement I realize there is no hope.
That’s not even close to true. There’s any number of distributions that might characterize any choice before you; the uniform case is simply the most trivial. As for relevance, you presented this demonstrably false dichotomy in the first place.
Well, neuroscientists are pretty sure its in the frontal lobe.
One, you’re garbling your own quote. You previously stated Hawking believed that the universe was unlawful outside of a local region of space for an unspecified slice of time. Two, the answer is no, the universe is lawful within a general region of space-time. This is general relativity applied to cosmology, specifically the FLRW metric.
Which is why I said the universe is lawful within the classical limit. Even beyond that, it is certainly lawful in the semiclassical sense. Hence, QM. The difference between QM and GR is that QM presents statistical laws and GR does not.
Feynman was making a philosophical point about interpreting the formalism of quantum mechanics. Is a wave function real? Are non-local interactions real? Is the indeterminacy of position and momentum below the Planck scale real? These are philosophical points that motivate interpretations of QM’s laws. Nevertheless, the laws are very well understood, well enough for its obvious application in our everyday life. That clicker isn’t magic, you know.
You previously stated Hawking believed that the universe was unlawful outside of a local region of space for an unspecified slice of time.
Did I? Let’s look again…
“…the universe isn’t lawful, never will be lawful, never can be lawful … at best we study local phenomena that may be predictable for an unspecifiable time.”
Sooo….
“unlawful outside of a local region of space” = “local phenomena that may be predictable”
I’ll have to leave that for others to decide.
He said the “universe isn’t lawful” which you change to “unlawful outside” which I suppose you could do? Except he’s not really saying, “only in safe places where we never really see it so everything’s alright and we can all go back to sleep now.”
He’s saying anything is possible and don’t have too much faith in your ability to predict anything. Because one day it’s going to bite you in the ass.
That’s not even close to true.
I say something is either random or it’s not. So what’s my third choice? If you say partly random I’m going to reach right through this internet and grab you by the neck.
Any of these, and countably infinite others and their variations.
Ya did! You should be feeling my boney fingers reaching for your neck right about now…
Ok. Stepping back a bit…
It is highly unlikely that the probability of you making a particular choice is evenly distributed.
This is making an assumption about my argument that really misses the point. The real question is are we just a meat machine? If so, making a copy shouldn’t (in theory) be such a big deal and will happen eventually. You are making the assumption, a priori, that we simply are meat machines. So it follows there is a probability distribution.
But I’m saying we are not meat machines… well, perhaps mostly we are, but with just the tiniest bit of something more. So you’re probability distribution has no place in the argument. It’s a non sequitur.
Are we meat machines or something more? The answer has consequences. Meat machines do not have free will. It’s just input/output. Choice is a complete illusion. A perfect copy kept in the exact same environment (to cancel out the butterfly effect) would follow an identical path. Free will and choice simply would not exist.
This I have a problem with.
This is getting tiresome, so let’s get to the point.
Says who?
I made my argument in the last paragraph. Feel free to refute it… accept it… ignore it… whatever.
That’s easy. You won’t defend the premise that “meat machines” necessarily lack free will.
Essentially, the problem we’re having is with the imprecision your use of terms you either appropriate or make up. Your argument boils down to “AI is bunk because ‘real’ humans have this thing called free will that we can’t sense or measure.” When I try to nail you down on a specific definition, it changes. First it was the capacity to make a choice in opposition to determinism. I present to you the example of a probabilistic computer. You then deride its indeterminacy. So which is it? Is free will deterministic and illusion, or indeterministic and troubling to you?
I haven’t changed anything. Free will is the ability to choose.
Machines don’t choose, they only follow the path they are put upon. That doesn’t change if you make it non deterministic with a random number generator. It doesn’t change no matter how analytical the program or what parameters are considered. Only a buffoon would say a computer has free will. It doesn’t. It doesn’t think. It’s choices are programmed into it.
Mine are not. I do think. I do choose. I am not just a machine.
So if I’m reading you right, machines don’t and can’t have this thing free will because by definition free will is something only intelligent agents in nature have.
It used to be thought that if a computer could play chess it would be thinking. Then they did it and realized what they’d come up did not involve thinking (except by the programmer that wrote the code.)
I have free will. Machines don’t. Refute that.
by definition free will is something only intelligent agents in nature have.
Hey I’m only pointing my finger at it… like you did to define intelligence.
Well the easiest refutation would be to point out your lack of verifiable evidence for whatever it is that you call free will and consign to the sole province of Man.
Actually, you’re pointing your finger at something you figure is a feature of free will, and then pointing at something else when shown a non-human example of the same thing. That is the very essence of shifting the goalposts.
Actually, you’re pointing your finger at something you figure is a feature of free will, and then pointing at something else when shown a non-human example of the same thing.
What??? I think you need to up your meds.
I’m pointing my finger at the same baby you did.
You say it is the definition of intelligence. I say it has intelligence and you haven’t defined anything at all.
I say the baby has free will. A machine does not. I have no idea what you’re saying.
My goalposts are firmly set. Showing that a machine does not have intelligence or free will doesn’t move the posts at all.
Try again.
shown a non-human example of the same thing
What non-human example?
Well the easiest refutation would be to point out your lack of verifiable evidence for whatever it is that you call free will
Let me get this straight… are you saying free will doesn’t exist???
Says the guy defending the supernatural.
No, you’re not. You’re pointing your finger at something that supercedes nature entirely.
On the contrary, I said intelligence is some collection of traits associated with this child. I did not exhaustive go beyond that since our argument at the time was whether or not intelligence could be defined, and if so to your satisfaction. Clearly, that was an impossible task.
Very well, given the only definition of free will you’ve offered so far, all I need do is show you a machine that evolves its states stochastically to suffice. Very well, I choose a slot machine.
No, your goal posts will move again. First it’s to ability to make a choice. I wonder what condition you will impose next?
Your computer chess player.
Not at all. The brain is an extraordinarily complicated, numerous ensemble, and I’d find highly unlikely such an ensemble reduces to something determinate (and therefore causal) and acyclic (and therefore incapable of introspection and control). However bounded by self-awareness, instinctual and nurtured values, etc., the brain can still exercise considerable freedom of action faced with any given choice.
Says the guy defending the supernatural.
You’re hallucinating. When did I defend supernatural? What I said was, “Supernatural is a misnomer. … which is just natural but beyond our normal experience.” In other words, if anything can be identified as supernatural it’s not some kind of magic, it’s just something natural that we don’t understand.
Arthur C. Clarke made a law of it. Perhaps you’re familiar?
Clearly, that was an impossible task.
Clearly, because if you could exhaustively define it, we should be able to reproduce it. Essential elements of it are clearly undefined.
all I need do is show you a machine that evolves its states stochastically to suffice. Very well, I choose a slot machine.
You’re saying a slot machine has free will and is an example of thought. Sorry, I (and all those that are sane) must disagree.
Your computer chess player.
…was a comparison to show contrast. How is this difficult to understand and what has this to do with moving goal posts?
Nah, you don’t get a pass for redefining natural to include the supernatural.
No, I’m not. I’m familiar with his dictum “any sufficiently advanced technology is indistinguishable from magic.” This is a haughty restatement of the principle that scientific knowledge does not claim completeness. To defend something inherently out of the reach of scientific understanding is to defend the supernatural.
Why? We can exhaustively define stars, but we can’t presently reproduce them. That’s not to say that a machine with a degree of intelligence we’d assign to a human being is that far out of reach.
I’m saying a slot machine has free will. Who cares whether or not it thinks?
Through incredulity. Yet this computer chess player must be able to consider its decisions, your decisions, and when presented with a choice between two or more equally desirable moves must still make a decision. Within the context of the game, only chauvinism denies that the behavior exhibited is distinct from self-awareness or free will. The only thing missing is choice in goal orientation, something easily rectified by implementing a checkers extension, a system that encodes some rudimentary values, and a random number generator to permits evolution of those values and choice between the two games. You need look no farther than a MMOG with non-player characters to find a system that does this and more.
It’s not. Your definition evolves when presented with a machine that meets it.
It only serves to show that you’re about to move them again.
We can exhaustively define stars, but we can’t presently reproduce them.
In theory we can so you are wrong. In theory, we are not able to create an electronic brain. It’s people in the A.I. community that have been moving the goal posts because of all the times they embarrassed themselves. Saying artificial has no meaning is just one example. ‘Electronic brain’ being another.
Nowhere did I redefine natural. What I said was if you identify something as beyond natural is more likely a case of misunderstanding. I suspect, in your case, you misunderstand on purpose.
I’m saying a slot machine has free will.
Thus endeth the argument.
Seriously, those goalposts must have a million frequent flier miles seeing how often you’re moving them.
Says who?
Exactly when have they ever moved any goalposts?
One that apparently exists only in your imagination.
Which, as pointed out here is apparently well beyond “reproducing…in theory.”
And yet you consistently point to some qualia concerning human intelligence and free will that is beyond natural understanding.
I’m certain you do.
Good day sir… I said good day.
… [a slot machine has free will.]
Jaron Lanier? Really? Getting his opinion on transhumanist topics is like asking a satanists opinion of the Pope.
Yet our host felt the article was worth linking to Mike. Even a satanist could have some worthwhile insights regarding the pope. Are there any points in the article you disagree with?
It’s people in the A.I. community that have been moving the goal posts because of all the times they embarrassed themselves.
I don’t know that it’s goalpost moving as collapsing expectations. Early on, they thought that things like general problem solving algorithms, theorem proving, and “common sense” relational knowledge would in themselves achieve genuine artificial intelligence. Later it was neural networks, genetic algorithms, and similar things. Now, I think it’s down to finding algorithms for data mining and knowledge bases which probably is where the expectations will linger for a while until someone gets something that works above that level.
I don’t know that it’s goalpost moving as collapsing expectations.
Perhaps so. They’ve had a lot of cold water splashed in there face by reality. Even today they make claims (giving us a date for the singularity… please) that they probably hope we forget when the time comes. They insist it’s just a matter of getting the transistor count up enough as if complexity were the only question… it’s an article of faith.