The Hard Problem of Gravity

[westprog]

A) First of all, AI is a very broad topic. It doesn't have the sole goal of reproducing (or having it's own flavor) of consciousness. If anything (and I'm no expert, Rocket Dodger knows much more than I), it's telling a computer/robot/embedded system "Do something as needed done in an efficient manner."

B) What makes you say that AI is flawed in "the approach" to consciousness? For one, there are different ways to go about it. Two, on what grounds do you have to make this accusation?

C) Disappointment? I know we can't be well read in every discipline, but maybe you a least do a little research before you construct an argument against a very large field of research.

When I say that AI research has been a disappointment, I'm not referring to RD's production of acid-hurling zombies who throw to allow for splash damage. That's generally been quite satisfactory. I mean something that could even approach passing the Turing test.

I remember what the aspirations were for the field back in the 1970's, when people were talking about ELIZA as if the next step would be a program that really understood things. That just hasn't happened.

It's the areas where AI has not succeeded as expected that we are dealing with here.

 

[westprog]

And why exactly do you think AI doesn't do this? Plenty of AI research is about trying to replicate biological mechanism in analog systems as well as research into the "pure" mathematical foundations of thought in logical systems.

I'm not opposing any kind of research, and certainly trying to duplicate certain functionality of the brain might be productive. But until we fully understand how the mind works, we will find it very difficult to emulate.

 

[westprog]

Where have I claimed this ?

If you are saying that information processing, as a concept, has been fuzzily defined, then I agree with that.

 

[westprog]

No, they don't. This has been explained to you many times over.

And it's been explained to you precisely why they do.

 

[rocketdodger]

I suppose that if you are really, really convinced that you are right, then the only explanation for disagreement is that the other person has a shallow and uninformed grasp of the subject, and hasn't thought about it enough.

The fundamental flaw of AI research is its basic assumption that its approach is the right way to produce consciousness. That's why, for all its assumptions, and achievemetns, the field has been such a disappointment.

See?

This is your typical response. Not only are you extremely uninformed but you seem to be unwilling to remedy that situation.

Instead of offering some sort of mathematical argument that some aspect of consciousness can't be reduced to simpler components, you offer a vague and sweeping critique of an entire branch of science.

If that is how you advance your knowledge of a subject, more power to ya.

 

[rocketdodger]

And why exactly do you think AI doesn't do this?

Because he has almost no knowledge of what is actually going on in AI research and furthermore he has no desire to learn.

Three people have already posted these links in this thread that invalidate just about everything westprog has said:

http://www.guardian.co.uk/technology/2007/dec/20/research.it

http://www.idsia.ch/~juergen/

http://news.bbc.co.uk/2/hi/science/nature/7979113.stm

Yet, we have heard zero commentary from him regarding the content of these links. I wonder why?

 

[rocketdodger]

When you say and understood by almost everyone you mean understood by conscious human beings.

Or anyone else using the language in question, yes. Obviously, if something isn't using a language, then the meaning of terms in that language is irrelevant to it.

They are the only constituents in the universe that actually care whether the AC is on or off. Indeed, they are the only constituents in the universe that we are aware of caring about anything*.

Oh, so now monkeys, dogs, dolphins, etc, aren't conscious anymore? They are incapable of caring about stuff?

Because Mr Jones sets his thermostat to a chilly 55 degrees F, why should we assume that the thermostat cares what the temperature is?

Because when the temperature passes 55 the behavior of the thermostat changes drastically.

What additional understanding of thermostat operation do we gain by using anthropomorphic language about it?

None. It is just easier to talk about because we are, ya know, humans.

When we say something "switches" we mean it satisfies a certain set of properties. For convenience, that set of properties has a default definition. A thermostat switches according to the default definition. Like I said, you can override the default definition in order to pigeonhole any entity into satisfying this set of properties.

But when you do that, you change the original meaning of "switches" slightly. If you want to say the rocks in the grand canyon switch, then you are drastically changing the default meaning of "switch."

I don't care what you call it, but the behavior of a thermostat is very different from the behavior of a rock. To get a rock to behave anything like a thermostat you would have to make it some kind of a doped semiconductor device -- which of course is done all over the world. Electronics aren't made from rock switches.

And this bleeds into what Pixy is trying to tell you -- if a rock could switch, people would use rocks as switches. We don't. All the word games in the world won't change that fact. You can redefine "switch" to your heart's content and it won't ever allow a single rock to take the place of a thermostat or a transistor or even a manual mechanical switch.

Furthermore, people have nothing to do with it. If you removed all the humans from the universe, a rock still wouldn't behave like a transistor. The "meaning" attributed to these entities by people is completely irrelevant.

 

[westprog]

See?

This is your typical response. Not only are you extremely uninformed but you seem to be unwilling to remedy that situation.

Instead of offering some sort of mathematical argument that some aspect of consciousness can't be reduced to simpler components, you offer a vague and sweeping critique of an entire branch of science.

If that is how you advance your knowledge of a subject, more power to ya.

My criticisms of AI are based on the hubris of the early days. It's not something to be taken personally.

You seem to be under the misapprehension that by knowing a great deal about a very specialised area, that you are entitled to lecture people who don't share your rose-tinted view of its wider applicability. And that can be considered a critique of your field as well.

 

[westprog]

Oh, so now monkeys, dogs, dolphins, etc, aren't conscious anymore? They are incapable of caring about stuff?

I'd actually put in a '*' for a footnote because I knew that someone would keep making THE SAME IRRELEVANT POINT. But I forgot to put the footnote in.

I've stated repeatedly that I think that animals are probably conscious to some degree. But it still gets jumped on.

 

[Belz...]

Sure. The laws of thermodynamics are suspended at the decree of talking heads on the internet.

Aku, nobody is decreeing anything.

Your behaviour is getting more and more obnoxious. It would help if you actually tried to understand what other people are saying and not hand-wave it by assuming that nobody knows squat.

You said:

Rocks respond to temperature changes just as thermostats do.

No, they don't. Thermostats add another layer of complexity as they respond to heat by actually doing something. Rocks don't.

 

[Belz...]

...........................................................

Belz... I posted a lengthy post restating the definition I've been going by, along with links to several other lengthy post in which I elaborate on it further. There is are several pages worth of me defining and defining what is meant by consicousness to the nth degree. Just because you can't be bothered with actually reading my responses does not make them magically disappear.

Amaze me. Where's that definition ?

I suggest you educate yourself on current physics and information theory.

Of course! It's such an easy answer. Why didn't I think of that ?!

No, wait! Why don't you instead tell me what you think you mean by "process information" and then I can hopefully amend my opinion on the matter to match yours.

 

[Belz...]

The obvious answer is that apparently consciousness is tied to the physical processes going on in a human being, in particular but not solely the brain. We don't have enough knowledge to define it more tightly than that.

You can remove "apparently" from your statements. It's absolutely tied. Completely ? Most probably. But tied.

If you are saying that information processing, as a concept, has been fuzzily defined, then I agree with that.

And where have I claimed that ?

Maybe you should stop trying to make up what I'm saying and actually read what I'm saying.

I'm not opposing any kind of research, and certainly trying to duplicate certain functionality of the brain might be productive. But until we fully understand how the mind works, we will find it very difficult to emulate.

Perhaps we'll never fully understand it until we replicate it. Did you think about that ?

And it's been explained to you precisely why they do.

Really ? No. You guys said that rocks expand with heat. Fine. If that's information processing to you, then everything processes information and the term is useless.

The way I see it, "information processing" means that information enters the system and is transformed into some other information. Thermostats do that because they take heat and transform it into "Heater ON". Rocks just expand, but they do squat with the data they get.

 

[westprog]

Because when the temperature passes 55 the behavior of the thermostat changes drastically.

And when you drop a rock off a cliff its behaviour changes drastically when it hits the bottom. Does that mean that rocks care about being dropped off cliffs?

It would help when you make these statements if you are actually attempting precise definitions. Is a drastic change in behaviour an indication that some inanimate object "cares"?

None. It is just easier to talk about because we are, ya know, humans.

When we say something "switches" we mean it satisfies a certain set of properties. For convenience, that set of properties has a default definition. A thermostat switches according to the default definition. Like I said, you can override the default definition in order to pigeonhole any entity into satisfying this set of properties.

But when you do that, you change the original meaning of "switches" slightly. If you want to say the rocks in the grand canyon switch, then you are drastically changing the default meaning of "switch."

I'm trying to get a physical definition. You are using an engineering definition. You can't hop from one realm to another using whatever definitions you have to hand.

I don't care what you call it, but the behavior of a thermostat is very different from the behavior of a rock. To get a rock to behave anything like a thermostat you would have to make it some kind of a doped semiconductor device -- which of course is done all over the world. Electronics aren't made from rock switches.

And this bleeds into what Pixy is trying to tell you -- if a rock could switch, people would use rocks as switches. We don't. All the word games in the world won't change that fact. You can redefine "switch" to your heart's content and it won't ever allow a single rock to take the place of a thermostat or a transistor or even a manual mechanical switch.

And you couldn't replace a rock with a thermostat either. If you can give a scientific definition of a switch that excludes the rock - or components within the rock - and includes thermostats, I'd be interested to see it.

Here's what I consider a rough and ready definition-

SWITCH - a physical object that allows or prevents an event from taking place according to its physical state.

What's yours?

Furthermore, people have nothing to do with it. If you removed all the humans from the universe, a rock still wouldn't behave like a transistor. The "meaning" attributed to these entities by people is completely irrelevant.

It's not a matter of whether a rock - or a part of a rock - behaves like a transistor. It's whether it behaves like a switch.

 

[Belz...]

I've stated repeatedly that I think that animals are probably conscious to some degree. But it still gets jumped on.

Where, along the line, do they stop being conscious ?

You can't answer that unless you have an operational definition of "consciousness". This is what we've been trying to do here and what you are incapable of doing.

Pixy's and Dodger's definitions work because they allow you to determine what is and what isn't based on behaviour, which you yourself admitted was the only criterion. The downside is that things are conscious under that definition that you, for some reason, refuse to aknowlege as such.

 

[westprog]

You can remove "apparently" from your statements. It's absolutely tied. Completely ? Most probably. But tied.



And where have I claimed that ?

Maybe you should stop trying to make up what I'm saying and actually read what I'm saying.

Maybe you should try saying it. Is information processing precisely or fuzzily defined?

Perhaps we'll never fully understand it until we replicate it. Did you think about that ?

Then replicate it. I'm not stopping you. And when you have a walkin' talkin' livin' doll, be sure and let me know. But so far, you don't.

Really ? No. You guys said that rocks expand with heat. Fine. If that's information processing to you, then everything processes information and the term is useless.

If you want to produce a physical definition of information processing that isn't dependent on the use human beings make of the information, then the onus is on you. Information processing is a useful term in computing. Whether it means anything in the real world is another matter.

The way I see it, "information processing" means that information enters the system and is transformed into some other information. Thermostats do that because they take heat and transform it into "Heater ON". Rocks just expand, but they do squat with the data they get.

I'm still not seeing a precise definition.

I would say that the Grand Canyon is a fairly clear chunk of information. A great big hole in the ground caused by liquid being channelled by rocks.

If you want to argue that the AC coming on is real information, and that the Grand Canyon isn't, then you will have to explain exactly why.

 

[yy2bggggs]

The way I see it, "information processing" means that information enters the system and is transformed into some other information.

By that definition, what about the fact that rocks convert light into heat? Or would things such as the conversion of time into ratios of potassium to argon count?

 

[westprog]

Where, along the line, do they stop being conscious ?

You can't answer that unless you have an operational definition of "consciousness". This is what we've been trying to do here and what you are incapable of doing.

It's also what you are incapable of doing.

Pixy's and Dodger's definitions work because they allow you to determine what is and what isn't based on behaviour, which you yourself admitted was the only criterion. The downside is that things are conscious under that definition that you, for some reason, refuse to aknowlege as such.

I could produce any number of definitions of consciousness. Things that breathe are conscious. Things made of cells are conscious. Human beings are conscious but nothing else is.

It's very easy to make such claims.

 

[westprog]

By that definition, what about the fact that rocks convert light into heat? Or would things such as the conversion of time into ratios of potassium to argon count?

If rocks weren't full of information, what would the geologists and archeologists be doing with them?

We still need a precise definition of information and information processing that excludes rocks and includes thermostats. I have a feeling that we'll be waiting quite some time.

 

[westprog]

Thermostats add another layer of complexity as they respond to heat by actually doing something. Rocks don't.

You must know that that's not true. Perhaps you would like to rephrase it.

- a rock responding to heat by doing something.

 

[yy2bggggs]

We still need a precise definition of information and information processing that excludes rocks and includes thermostats. I have a feeling that we'll be waiting quite some time.

Well, I still see useful distinctions between the "switch" as defined earlier, information storage, and information conversion. Whatever significance other people tie to these distinctions is another issue.