Limitations of Turing tests

[Kharakov]

It's possible that the physical correlates of consciousness will be discovered. Woowho!

It's possible that they won't.

I was recently reading PK Dick's electric sheep book. PKPKPK! The idea is that certain types of psychological tests could determine whether someone was an android or not (Blade Runner, right??).

I'm thinking about dystopian future societies in which one's ability and willingness to satisfy human desires is a reflection of whether or not one is conscious.

What side of the line will the conscious beings be on- will it be determined by what they are willing to do? Will it be determined by what their apparent desires are?

Gotta go.

 

[untermensche]

Turing never thought much of the so-called "Turing Test".

It is only lesser minds that have turned into some great endeavor.

If I make a machine that can run it tells me absolutely nothing about how a horse runs, just as a machine that appears to have human attributes tells us nothing about how humans run.

 

[Bronzeage]

There is an old BBC show called Private Schultz, set in WW2. Schultz is a petty criminal who concocts a scheme to get himself and all his buddies into a cushy situation, by convincing the SS to start a counterfeiting operation, which will make British 10pound notes. The standard they aim for is "a note the Bank of England will accept." It is not possible for a gang of criminals in a German prison to produce a real 10 pound note, but they can make one that will fool the arbiter of these matters.

It's a financial Turing test. When the counterfeit cannot be distinguished from the genuine, it's clear the test is no longer adequate for the task, not that counterfeits are real.

 

[Juma]

The turing test is a attemot to define what we mean when saying that someone is thinking. Not to discretize berween humans and machines.

 

[Kharakov]

It's a financial Turing test. When the counterfeit cannot be distinguished from the genuine, it's clear the test is no longer adequate for the task, not that counterfeits are real.

My concerned, at least for the Turing test side of things, would be that conscious beings might be thought of as non-conscious and treated as if they felt nothing, and were too beaten down to act as if they felt something.

So a counterfeit consciousness would still be a consciousness. Something that feels nothing would be another story altogether...

 

[untermensche]

The turing test is a attemot to define what we mean when saying that someone is thinking. Not to discretize berween humans and machines.

How could a machine that merely mimics human behavior ever be evidence of human thinking?

Maybe someday somebody will grow a machine that thinks.

You can't build one. We weren't built, we grew.

 

[Juma]

The turing test is a attemot to define what we mean when saying that someone is thinking. Not to discretize berween humans and machines.

How could a machine that merely mimics human behavior ever be evidence of human thinking?

What is "human thinking"? What is the difference between "human thinking" and "thinking"? What does "thinking" mean?

Those are the questions Turing wanted to answer.

 

[untermensche]

How could a machine that merely mimics human behavior ever be evidence of human thinking?

What is "human thinking"? What is the difference between "human thinking" and "thinking"? What does "thinking" mean?

Those are the questions Turing wanted to answer.

I agree that "What is thinking?" is a good question but I don't think Turing ever said we could understand it with this so-called Turing Test.

It was a throw away idea that he never took seriously nor should anyone else.

That is not to say that the field of artificial intelligence is worthless. On the contrary it is highly worthwhile to try to mimic human capacities.

It won't tell us anything about how humans work but it leads to valuable things like voice recognition.

And human thinking is the kind of thinking humans are capable of.

And it is difficult to make many objective statements about what "thinking" is, because it's very nature is subjective and unobservable, even by the thinker. I can experience my thoughts but I can't see them, and most definitely can't see how they are constructed. Even a mental image is not "seen". It is experienced. When we say we "see" a mental image it is only a metaphor. To see is to experience the information hitting the eye. It is a similar experience but it has limits, it conforms to how the universe actually behaves. While mental imagery has limits too it has different limits than "seeing". I can only "see" the way the universe behaves, but I can imagine all kinds of behaviors, like flying cats, that I can't "see".

But all of this is very familiar to most people, because while thinking may be difficult to explain, they all think.

 

[Bronzeage]

It's a financial Turing test. When the counterfeit cannot be distinguished from the genuine, it's clear the test is no longer adequate for the task, not that counterfeits are real.

My concerned, at least for the Turing test side of things, would be that conscious beings might be thought of as non-conscious and treated as if they felt nothing, and were too beaten down to act as if they felt something.

So a counterfeit consciousness would still be a consciousness. Something that feels nothing would be another story altogether...

A consciousness that feels nothing is a contradiction in terms.

There is nothing which says a machine could not have a consciousness. A dog has a consciousness, so does a cat. It's plain for any observer to see there is a great difference in the consciousness of these two animals, but both have a degree of memory, both can anticipate the future, and both are able to interpret the feelings of others, which we call empathy.

Whether the machine can be built which equals a dog's consciousness is yet to seen. Since we don't have a viable model of our own consciousness, It will probably be one of those, "we'll know it when we see it," things.

 

[Juma]

What is "human thinking"? What is the difference between "human thinking" and "thinking"? What does "thinking" mean?

Those are the questions Turing wanted to answer.

I agree that "What is thinking?" is a good question but I don't think Turing ever said we could understand it with this so-called Turing Test.

It was a throw away idea that he never took seriously nor should anyone else.

That is not to say that the field of artificial intelligence is worthless. On the contrary it is highly worthwhile to try to mimic human capacities.

It won't tell us anything about how humans work but it leads to valuable things like voice recognition.

And human thinking is the kind of thinking humans are capable of.

Just the fact that you use the term "human thinking" shows that you are burying yourself in a corner.

 

[untermensche]

I agree that "What is thinking?" is a good question but I don't think Turing ever said we could understand it with this so-called Turing Test.

It was a throw away idea that he never took seriously nor should anyone else.

That is not to say that the field of artificial intelligence is worthless. On the contrary it is highly worthwhile to try to mimic human capacities.

It won't tell us anything about how humans work but it leads to valuable things like voice recognition.

And human thinking is the kind of thinking humans are capable of.

Just the fact that you use the term "human thinking" shows that you are burying yourself in a corner.

I don't have a clue what you're talking about.

Human thinking is the thinking humans do. A lot of it involves human language.

But all it expresses is that humans are not the only thing that thinks.

Squirrels think. Their thinking would be described as squirrel thinking.

 

[Speakpigeon]

There is an old BBC show called Private Schultz, set in WW2. Schultz is a petty criminal who concocts a scheme to get himself and all his buddies into a cushy situation, by convincing the SS to start a counterfeiting operation, which will make British 10pound notes. The standard they aim for is "a note the Bank of England will accept." It is not possible for a gang of criminals in a German prison to produce a real 10 pound note, but they can make one that will fool the arbiter of these matters.

It's a financial Turing test. When the counterfeit cannot be distinguished from the genuine, it's clear the test is no longer adequate for the task, not that counterfeits are real.

Yeah, well, FDI would say that you can buy stuff with counterfeit money so it is useful and that's all that matters.

Objectively, he's right of course. If there is not test that can detect a counterfeit bill then it could be used to buy stuff legally no questions asked.

It would still be counterfeit money, though.
EB

 

[Speakpigeon]

How could a machine that merely mimics human behavior ever be evidence of human thinking?

What is "human thinking"? What is the difference between "human thinking" and "thinking"? What does "thinking" mean?

Those are the questions Turing wanted to answer.

And never did.

I always thought of the test as completely idiotic. Basically, it said that it's inventor thought that if he could be fooled by smart computer-generated gibberish he would be convinced that the machine could think even though it's clear enough that the brain does a whole lot beside generating gibberish that at least some other people would find convincing. We will be fooled, if we are not already, but that still doesn't amount to thinking.

Another problem now however is that real thinking humans are using, or really abusing, the test to promote their wares and gadgets to win dollars. How pathetic.
EB

 

[Speakpigeon]

And human thinking is the kind of thinking humans are capable of.

Just the fact that you use the term "human thinking" shows that you are burying yourself in a corner.

A corner? Could you please explain your thinking behind this interesting remark?

Me I thought it was an articulate and effective use of a human language (if you don't mind that qualification talking about English) to convey one of the speaker's ideas. For example that even if we grant that a computer could think this would still not entail that it could do something close enough to human thinking. How would you express this idea without using the expression "human thinking"?
EB

 

[Kharakov]

I always thought of the test as completely idiotic. Basically, it said that it's inventor thought that if he could be fooled by smart computer-generated gibberish he would be convinced that the machine could think even though it's clear enough that the brain does a whole lot beside generating gibberish that at least some other people would find convincing. We will be fooled, if we are not already, but that still doesn't amount to thinking.

Until the physics of consciousness (more aptly, the consciousness of the physical) is understood sufficiently, they only test we have is "does it seem conscious to me?". Of course, some conscious beings like pretending to be robots.

beep.

 

[Juma]

Just the fact that you use the term "human thinking" shows that you are burying yourself in a corner.

A corner? Could you please explain your thinking behind this interesting remark?

Me I thought it was an articulate and effective use of a human language (if you don't mind that qualification talking about English) to convey one of the speaker's ideas. For example that even if we grant that a computer could think this would still not entail that it could do something close enough to human thinking. How would you express this idea without using the expression "human thinking"?
EB

"Human thinking" is as silly as flogiston. What do you think is so special human about thinking?

 

[Speakpigeon]

A corner? Could you please explain your thinking behind this interesting remark?

Me I thought it was an articulate and effective use of a human language (if you don't mind that qualification talking about English) to convey one of the speaker's ideas. For example that even if we grant that a computer could think this would still not entail that it could do something close enough to human thinking. How would you express this idea without using the expression "human thinking"?
EB

"Human thinking" is as silly as flogiston.

Your sortie here is as idiotic as ever. You are really the Kind of the Idiotic Repartie. The flogiston is supposed not to exist. So, are you trying to say there's actually no human thinking?

What do you think is so special human about thinking?

Probably you have this assumption that thinking is fundamentally the same if it's done by a machine or a by a human being. Ok, maybe that's true but maybe it isn't. But again, your sortie is totally beside the point. I was pointing out to you that the expression "human thinking" is meaningful, for example to express the idea that thinking by a human is fundamentally different from thinking by a machine. This idea is in fact one point you are debating with untermensche. So, ask him what is different.

Me, I don't know and you don't know either and actually nobody knows, and therefore everything in this respect remains to be discovered. It's also a question of vocabulary, again. If you define thinking as what a computer does, well, broadly, a human being can do it so you can say that there's no substantial difference but your point is trivial and idiotic, again.

But untermensche starts from a different point entirely. He calls "human thinking" whatever human beings do when they think, and that includes things way beyond what computers do, in particular the subjective experience of thoughts. I don't believe for a moment that you or anybody else could prove that computers do human thinking as defined here.
EB

 

[Juma]

"Human thinking" is as silly as flogiston.

Your sortie here is as idiotic as ever. You are really the Kind of the Idiotic Repartie. The flogiston is supposed not to exist. So, are you trying to say there's actually no human thinking?

What do you think is so special human about thinking?

Probably you have this assumption that thinking is fundamentally the same if it's done by a machine or a by a human being. Ok, maybe that's true but maybe it isn't. But again, your sortie is totally beside the point. I was pointing out to you that the expression "human thinking" is meaningful, for example to express the idea that thinking by a human is fundamentally different from thinking by a machine. This idea is in fact one point you are debating with untermensche. So, ask him what is different.

Me, I don't know and you don't know either and actually nobody knows, and therefore everything in this respect remains to be discovered. It's also a question of vocabulary, again. If you define thinking as what a computer does, well, broadly, a human being can do it so you can say that there's no substantial difference but your point is trivial and idiotic, again.

But untermensche starts from a different point entirely. He calls "human thinking" whatever human beings do when they think, and that includes things way beyond what computers do, in particular the subjective experience of thoughts. I don't believe for a moment that you or anybody else could prove that computers do human thinking as defined here.
EB

Humans do a lot of things when they think. Should really all that be labelled thinking?

 

[Speakpigeon]

Humans do a lot of things when they think. Should really all that be labelled thinking?

Not scratching your nose, breathing, etc. But Descartes adopted a large definition, which, if you think about it, is reflected in untermensche's definition.

There are also plenty of words for what computer's do (computation, calculation, text edition, simulation, etc. etc. etc.). Where is the need to redefine the word "thinking" as what computers do? This is hubris. Scientists are human beings, if they think they can escape human nature merely by being scientific in their outlook they are wrong. I suggest some of them should try to visit their nearest church to ask the Lord for forgiveness even if they don't believe in God. Works even without a God this.
EB

 

[Juma]

Humans do a lot of things when they think. Should really all that be labelled thinking?

Not scratching your nose, breathing, etc. But Descartes adopted a large definition, which, if you think about it, is reflected in untermensche's definition.

There are also plenty of words for what computer's do (computation, calculation, text edition, simulation, etc. etc. etc.). Where is the need to redefine the word "thinking" as what computers do? This is hubris. Scientists are human beings, if they think they can escape human nature merely by being scientific in their outlook they are wrong. I suggest some of them should try to visit their nearest church to ask the Lord for forgiveness even if they don't believe in God. Works even without a God this.
EB

There is no hybris. Just a decent wish to create a useful definition of "thinking" without including everything.