Computers as the Philosopher's Stone of AI
--------------

Those who are seriously interested in ai are faced with two choices:

Computers are useful, which nobody doubts, and

Computers are intelligent, which nobody believes.

And 99+% of all the verbiage devoted to explaining computers as the
agency of artificial intelligence are just so much rationalization of
the proposition that computers are intelligent, which no one believes.
Computers are simply the modern philosopher's stone of intelligence.

Regards - Lester

From:	rick++
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	10 Jan 2005 08:32:42 -0800

Many of mankind's technologies have been metaphors
for the human mind, especially since the Enlightenment
and most especially since electronic computing.
But the debate has always been there how good of
a metaphor these actually are. Not so great in my
opinion, if 60 years of effort hasnt gotten very far in
manufacturing an A.I.

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 17:19:00 GMT

On 10 Jan 2005 08:32:42 -0800, "rick++" in
comp.ai.philosophy wrote:

>Many of mankind's technologies have been metaphors
>for the human mind, especially since the Enlightenment
>and most especially since electronic computing.
>But the debate has always been there how good of
>a metaphor these actually are. Not so great in my
>opinion, if 60 years of effort hasnt gotten very far in
>manufacturing an A.I.

Yeah, but that certainly hasn't stopped academics from pretending
universities and taxpayers have gotten their money's worth out of the
failure.

Regards - Lester

From:	RJS
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	16 Jan 2005 09:20:39 -0800

Yeaa looks like everyone else has left.
Intelligence is a very tricky word. But as far as computers are
concerned i think we would lower the requirements for it being
intelligent. After all we would only want them to be so much inteligent
that they do not form unions and start asking for rights ( the last
thing we need is another race, when we hardly have space for ourselfs
on this planet).
Also this is only a philosophical question. As you said machines
cannot be intelligent.

But the overall objective is to have a machine which can do farely
complex tasks without supervision. Now this is a very broad
definitation and also ends up including the robotic arms and other such
things.
This further brings it down to what are complex things that humans can
do. This also is in incoherence with your comment about computers being
a mirrors. The thing is if we want something to learn we can only teach
it from our own experince. Thus an intelligent machine should be able
to do what a regular human can do ( not including rocket scientists
8-) ). A truly intelligent machine would be one, ( atleast what we
would like it to be ) that is a menial worker. This discussion can get
into a full social issue over this comment.
But i believe all the scientific community wants is a deeper
understanding of how humans work and this can be achieved by modeling a
machine based on it.

From:	pensul
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	19 Jan 2005 19:15:27 GMT

> But i believe all the scientific community wants is a deeper
> understanding of how humans work and this can be achieved by modeling a
> machine based on it.
>
In a sense this already exists because people are many times judged to be fit
for a job if they have access to computers and the internet.

--
"The world of existence is an emanation of the merciful attribute of God."
Abdul-Baha
http://www.costarricense.cr/pagina/ernobe

From:	RJS
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	16 Jan 2005 12:05:00 -0800

Hey i posted a reply to this and it is not there. Strange

From:	JGCASEY
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	9 Jan 2005 19:16:18 -0800

Lester Zick wrote:
> On 9 Jan 2005 12:47:45 -0800, "JGCASEY" in
> comp.ai.philosophy wrote:
>
> >
> >Lester Zick wrote:
> >> Computers as the Philosopher's Stone of AI
> >> --------------
> >>
> >> Those who are seriously interested in ai are faced with two
choices:
> >>
> >> Computers are useful, which nobody doubts, and
> >>
> >> Computers are intelligent, which nobody believes.
> >>
> >> And 99+% of all the verbiage devoted to explaining computers as
the
> >> agency of artificial intelligence are just so much rationalization
of
> >> the proposition that computers are intelligent, which no one
> >believes.
> >> Computers are simply the modern philosopher's stone of
intelligence.
> >>
> >> Regards - Lester
> >
> >Let us not confuse the computer with the software which
> >at this stage is open ended.
>
> So are computers. I don't see what this is in aid of unless you're
> trying to contend that computers aren't yet intelligent but soon will
> be as soon as someone can figure out what intelligence means that
> computers and software can do.
>
> >Let us not confuse current software with what computers
> >can and cannot do.
>
> Same observation. Suggesting that computers are not but will become
> intelligent is the same as suggesting that computers are intelligent.
>
> Regards - Lester

You make statements such as "the proposition that computers
are intelligent" and I was making the point computers aren't
anything. People say the computer this, the computer that,
when they mean software and they generalize, from what that
particular piece (or pieces) of software do, about what is
or is not possible.

When you write "computers can't ... whatever" it makes
as much sense as saying a collection of amino acids, DNA
etc cannot make a human because all we have so far is a
virus that cannot do anything by itself.

A computer has to be "wired up" to be something. This
is called programming which connects the various parts
of the computer together.

A general purpose computer can be programmed (wired up)
to become isomorphic with any dynamic system *providing*
the computer has enough components that work fast enough
and we know how the system in question works.

I don't think we have to figure out what intelligence
means so much as how we can get a machine to do the
things that fall into the category of intelligence.

Regards - John

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 17:13:43 GMT

On 9 Jan 2005 19:16:18 -0800, "JGCASEY" in
comp.ai.philosophy wrote:

>
>Lester Zick wrote:
>> On 9 Jan 2005 12:47:45 -0800, "JGCASEY" in
>> comp.ai.philosophy wrote:
>>
>> >
>> >Lester Zick wrote:
>> >> Computers as the Philosopher's Stone of AI
>> >> --------------
>> >>
>> >> Those who are seriously interested in ai are faced with two
>choices:
>> >>
>> >> Computers are useful, which nobody doubts, and
>> >>
>> >> Computers are intelligent, which nobody believes.
>> >>
>> >> And 99+% of all the verbiage devoted to explaining computers as
>the
>> >> agency of artificial intelligence are just so much rationalization
>of
>> >> the proposition that computers are intelligent, which no one
>> >believes.
>> >> Computers are simply the modern philosopher's stone of
>intelligence.
>> >>
>> >> Regards - Lester
>> >
>> >Let us not confuse the computer with the software which
>> >at this stage is open ended.
>>
>> So are computers. I don't see what this is in aid of unless you're
>> trying to contend that computers aren't yet intelligent but soon will
>> be as soon as someone can figure out what intelligence means that
>> computers and software can do.
>>
>> >Let us not confuse current software with what computers
>> >can and cannot do.
>>
>> Same observation. Suggesting that computers are not but will become
>> intelligent is the same as suggesting that computers are intelligent.
>>
>> Regards - Lester
>
>
>You make statements such as "the proposition that computers
>are intelligent" and I was making the point computers aren't
>anything.

Maybe you should be explaining this to IBM.

> People say the computer this, the computer that,
>when they mean software and they generalize, from what that
>particular piece (or pieces) of software do, about what is
>or is not possible.

It doesn't change the terms of the observations I made to change the
term hardware to software. How about if I just change the original
statements to:

Ware is useful, which nobody doubts, and

Ware is intelligent, which nobody believes.

Then you can argue to your hearts content over whether qualifying the
term ware as hard or soft changes the nature of the argument. Nor do I
try to generalize from particular ware. That's your claim.

>When you write "computers can't ... whatever" it makes
>as much sense as saying a collection of amino acids, DNA
>etc cannot make a human because all we have so far is a
>virus that cannot do anything by itself.

Yeah, yeah, maybe you could point out exactly where I said "computers
can't . . ."

>A computer has to be "wired up" to be something. This
>is called programming which connects the various parts
>of the computer together.

Thanks for the programming lesson.

>A general purpose computer can be programmed (wired up)
>to become isomorphic with any dynamic system *providing*
>the computer has enough components that work fast enough
>and we know how the system in question works.

Yeah, well isomorphism with any dynamic system is just swell. Maybe
you should consider changing the name from AI to AIDS, Artificial
Isomorphism with Dynamic Systems. Then you wouldn't even have to
address sticky questions like intelligence.

>I don't think we have to figure out what intelligence
>means so much as how we can get a machine to do the
>things that fall into the category of intelligence.

Yeah, well, if we knew what those were we would be home free, wouldn't
we? Kinda like restricting our definition of prime numbers to ones we
already know are prime like 1, 2, and 3. Then we could develop a
formula for prime numbers out of hand instead of having to wade
through the tedious process of factoring. Someone once actually
proposed a law, I believe it was in Indiana, setting the value of
pi=3.14 period so schoolchildren wouldn't have to overtax their
imaginations trying to understand the concept.

So if we restrict our definition of intelligent applications to
playing chess, then any ware that can play chess is ipso facto
intelligent. Or maybe you'd rather go back to the mathematical
calculation of the number of toes on three toed sloths. Same
difference. Like I said above:

>"And 99+% of all the verbiage devoted to explaining computers as
> the agency of artificial intelligence is just so much rationalization of
> the proposition that computers are intelligent, which no one believes."

Regards - Lester

From:	pensul
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	18 Jan 2005 08:40:26 GMT

> I don't think we have to figure out what intelligence
> means so much as how we can get a machine to do the
> things that fall into the category of intelligence.
>
No amount of programming can get it to compare itself to a virus that cannot
do anything by itself, because the concept that it is dependent on us cannot
be understood by it. It cannot usefully represent an unknown -- us, who can
tell it to do anything we want. For that to happen it would need the capacity
for analogic reasoning.

--
"The world of existence is an emanation of the merciful attribute of God."
Abdul-Baha
http://www.costarricense.cr/pagina/ernobe

From:	synapse001 at gmail.com
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	17 Jan 2005 12:11:51 -0800

The arguement as far as I can decypher it from Lester seems to be as
follows.
Intelligence can not be defined.
Machiens can not be built to display somethign that can not be defined.
Thus Machines can not ever be intelligent.

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 17 Jan 2005 22:49:03 GMT

On 17 Jan 2005 12:11:51 -0800, "synapse001@gmail.com"
in comp.ai.philosophy wrote:

>The arguement as far as I can decypher it from Lester seems to be as
>follows.
>Intelligence can not be defined.
>Machiens can not be built to display somethign that can not be defined.
>Thus Machines can not ever be intelligent.

This is about as far removed from my position as I can imagine and I
have no idea how you got this impression. My definition for machine
intelligence was summarily outlined in Apocalypse Now. What I think
you may be referring to is my analysis of the ways people go about
defining intelligence as if one definition were as good as another. Of
course one cannot program machine intelligence without understanding
what intelligence is first any more than one could do it without first
understanding what a machine is.

What I see is people recognizing they don't understand intelligence in
any mechanical sense and then trying to mechanize it anyway as if
whatever we put on a machine were intelligence because we put it on a
machine without first understanding what intelligence is or can be. Or
people run off in all directions at once trying to define intelligence
in some completely superficial way without taking into account that
intelligence is what underlies definitions and not vice versa so that
trying to get at the definition of intelligence without understanding
the nature of intelligence on which definition relies is futile.

Do I think machine intelligence is practical? Probably not. But it
would give us considerable insight into our own mechanics. To me
robotics is a much more practical field in purely utilitarian terms.

Regards - Lester

From:	JPL Verhey
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 03:31:23 +0100

"Lester Zick" wrote in message
news:41e1933f.13551144@netnews.att.net...
>
> Computers as the Philosopher's Stone of AI
> --------------
>
> Those who are seriously interested in ai are faced with two choices:
>
> Computers are useful, which nobody doubts, and
>
> Computers are intelligent, which nobody believes.
>
> And 99+% of all the verbiage devoted to explaining computers as the
> agency of artificial intelligence are just so much rationalization of
> the proposition that computers are intelligent, which no one believes.
> Computers are simply the modern philosopher's stone of intelligence.

From my reply to John Casey and if I still agree with myself, it is as
legitimate (or illegitimate) to say that computers are intelligent as
they are cool, a scourge, dumb, usefull or useless.

A bit aside here, Lester, but I got a little riddle for you (perhaps
not, and you debunk it on e the spot):

The universal P(differences), is it the same as the other universal
P(differences) we talked about?

Nuff for now,
Cheers

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 16:25:39 GMT

On Mon, 10 Jan 2005 03:31:23 +0100, "JPL Verhey"
in comp.ai.philosophy wrote:

>
>"Lester Zick" wrote in message
>news:41e1933f.13551144@netnews.att.net...
>>
>> Computers as the Philosopher's Stone of AI
>> --------------
>>
>> Those who are seriously interested in ai are faced with two choices:
>>
>> Computers are useful, which nobody doubts, and
>>
>> Computers are intelligent, which nobody believes.
>>
>> And 99+% of all the verbiage devoted to explaining computers as the
>> agency of artificial intelligence are just so much rationalization of
>> the proposition that computers are intelligent, which no one believes.
>> Computers are simply the modern philosopher's stone of intelligence.
>
>From my reply to John Casey and if I still agree with myself, it is as
>legitimate (or illegitimate) to say that computers are intelligent as
>they are cool, a scourge, dumb, usefull or useless.
>
>A bit aside here, Lester, but I got a little riddle for you (perhaps
>not, and you debunk it on e the spot):
>
>The universal P(differences), is it the same as the other universal
>P(differences) we talked about?

Not sure what you're referring to, JPL. All I remember discussing
was P "differences" as universal because alternatives are inherently
self contradictory.

Regards - Lester

From:	Kobrinsky
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	14 Jan 2005 17:27:05 -0800

RJS:
> Interesting discussion going on here.

ROFL, we got some case of lost train, it seems the convoy
already left discussion. :->>

> Refering to computers as intelligent is still a very far
> away target.

Let us get a common definition of intelligence. Questions
about AI always belongs to and break down over a wrong
dominated & applied idea of the word 'intelligence'.

> Computers are merely machines which can be made to do a
> variety of tasks. The advantage of computers is like cars,
> humans get tired after walking a few kilometers whereas cars
> can drive for as long as you keep feeding them gas. This is
> not a complete analogy ( before anyone else points towards
> it) but something similar.

Example is fine, but the purpose of using computers doesn't
seems relevant to discute about artifical intelligence.

> Computers can add, match, sort etc and one of the biggest
> advantage they have is the capacity to work with a huge
> number of variables.

You're kidding?

> This is one of the feature that makes us think that they
> are more intelligent but in actuality all it means is they
> have a bigger slate to write on.

ACK.

> The behavior of computers is nowhere near intelligent, but
> there is a work in progress which will make them behave in
> an intelligent-like fashion.

Never, its an awful, very awful idea of intelligence that
lets people call something intelligent or its opposition.
And here we got the main point of the basicly ideological
troubles resulted to almost started discutions about AI.

> Computers are tools in hands of humans for to experiement
> with the representation of knowledge and developing
> alogrithms on how to efficiently operate on that knowledge.
> Once we have mastered this then atleast we have something
> which will show intelligent-like features.
> This is the point till we can visualize and strive for,
> after this point what differentiates real intelligence from
> this constructed one is the ability to adapt according to
> environment. One can argue a module for adpatation can also
> be programmed in it, but the problem is adpat to what?

Nice, now we're getting very close to the definition of get
handled something intelligent. My primary intention is to
get defined the /basic/ meaning of the word intelligence.
There are always the same questions and discusions over AI,
the same ideological deputys with the same old rhetorical
arguments, but an useful definition of the most interpreted
word I'm still missing. A definition that will give the
urgent needed ground to push through futur acts to progress
scientifical computing & AI.

> Now given the fact that in 60 years we still are not
> that advanced to achieve the first part should tell
> us that its a long way before computers can be truly
> intelligent.

A lot of scientists and people are working very hard to
push computers forward, to progress, to get them onto
fullfill acts we are calling intelligent. Please tell
me which you think are points needed to call a computer
be intelligent?

Machines itself will never be intelligent.

> But they will continue to amaze people.

Exactly. And u know why? Computers now are the colossal
structured mirror peoples own reality. Getting amazed by
recognize someones or your own face while using its
programmed software can makes us delirious, right.
But if we are able to watch our reality out of computers
we are allowed to drink the fountain of intelligence and
will prosurvive.

Did you follow my hint about artifical intelligence?
Have you got my basic definition of real intelligence?

I.

From:	JGCASEY
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	9 Jan 2005 12:47:45 -0800

Lester Zick wrote:
> Computers as the Philosopher's Stone of AI
> --------------
>
> Those who are seriously interested in ai are faced with two choices:
>
> Computers are useful, which nobody doubts, and
>
> Computers are intelligent, which nobody believes.
>
> And 99+% of all the verbiage devoted to explaining computers as the
> agency of artificial intelligence are just so much rationalization of
> the proposition that computers are intelligent, which no one
believes.
> Computers are simply the modern philosopher's stone of intelligence.
>
> Regards - Lester

Let us not confuse the computer with the software which
at this stage is open ended.

Let us not confuse current software with what computers
can and cannot do.

Regards - John

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 00:58:58 GMT

On 9 Jan 2005 12:47:45 -0800, "JGCASEY" in
comp.ai.philosophy wrote:

>
>Lester Zick wrote:
>> Computers as the Philosopher's Stone of AI
>> --------------
>>
>> Those who are seriously interested in ai are faced with two choices:
>>
>> Computers are useful, which nobody doubts, and
>>
>> Computers are intelligent, which nobody believes.
>>
>> And 99+% of all the verbiage devoted to explaining computers as the
>> agency of artificial intelligence are just so much rationalization of
>> the proposition that computers are intelligent, which no one
>believes.
>> Computers are simply the modern philosopher's stone of intelligence.
>>
>> Regards - Lester
>
>Let us not confuse the computer with the software which
>at this stage is open ended.

So are computers. I don't see what this is in aid of unless you're
trying to contend that computers aren't yet intelligent but soon will
be as soon as someone can figure out what intelligence means that
computers and software can do.

>Let us not confuse current software with what computers
>can and cannot do.

Same observation. Suggesting that computers are not but will become
intelligent is the same as suggesting that computers are intelligent.

Regards - Lester

From:	RJS
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	12 Jan 2005 11:04:20 -0800

Interesting discussion going on here.
Refering to computers as intelligent is still a very far away target.
Computers are merely machines which can be made to do a variety of
tasks. The advantage of computers is like cars, humans get tired after
walking a few kilometers whereas cars can drive for as long as you keep
feeding them gas. This is not a complete analogy ( before anyone else
points towards it) but something similar. Computers can add, match,
sort etc and one of the biggest advantage they have is the capacity to
work with a huge number of variables. This is one of the feature that
makes us think that they are more intelligent but in actuality all it
means is they have a bigger slate to write on.

The behavior of computers is nowhere near intelligent, but there is a
work in progress which will make them behave in an intelligent-like
fashion. Computers are tools in hands of humans for to experiement
with the representation of knowledge and developing alogrithms on how
to efficiently operate on that knowledge. Once we have mastered this
then atleast we have something which will show intelligent-like
features.
This is the point till we can visualize and strive for, after this
point what differentiates real intelligence from this constructed one
is the ability to adapt according to environment. One can argue a
module for adpatation can also be programmed in it, but the problem is
adpat to what?
Now given the fact that in 60 years we still are not that advanced to
achieve the first part should tell us that its a long way before
computers can be truly intelligent. But they will continue to amaze
people.

RJS

From:	pensul
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	18 Jan 2005 09:05:56 GMT

> .... But they will continue to amaze
> people.

Computers cannot adapt to anything without self-awareness. Animals seem to
have this, but it could actually be nothing but an accidental appearance.

--
"The world of existence is an emanation of the merciful attribute of God."
Abdul-Baha
http://www.costarricense.cr/pagina/ernobe

--
"The world of existence is an emanation of the merciful attribute of God."
Abdul-Baha
http://www.costarricense.cr/pagina/ernobe

From:	RJS
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	17 Jan 2005 20:36:00 -0800

I agree with Lester: to have an intelligent machine we need to define
what intelligence is. But the problem is that its an np-hard problem.
No optimal solution exists for it. On the other hand we can have
tending to optimality solution which can depend on two things.

1) Defining it in terms what can be achieved.
2) defining it in terms what it might be.

The first gives as a goal to work towards and the second creates enough
room to make the field charismatic for people to join and work in it.
Always if nothing else this exercise will make us understand humans
more.

RJS

From:	JGCASEY
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	10 Jan 2005 10:31:35 -0800

Lester Zick wrote:
> On 9 Jan 2005 19:16:18 -0800, "JGCASEY" in
> comp.ai.philosophy wrote:
[...]

>> People say the computer this, the computer that,
>> when they mean software and they generalize,
>> from what that particular piece (or pieces)
>> of software do, about what is or is not possible.
>
> It doesn't change the terms of the observations
> I made to change the term hardware to software.
> How about if I just change the original
> statements to:
>
> Ware is useful, which nobody doubts, and
>
> Ware is intelligent, which nobody believes.
>
> Then you can argue to your hearts content
> over whether qualifying the term ware as
> hard or soft changes the nature of the
> argument.

No, nothing to do with hard or soft. I was
pointing out that a "computer" was a collection
of parts and wasn't anything useful until
those parts were put together to make some
kind of machine.

> Nor do I try to generalize from
> particular ware. That's your claim.

That comes from your second statement, "ware
is intelligent, which nobody believes."

People don't believe ware is intelligent for
the simple reason software isn't intelligent
in the human sense at this point of time.
There is a *third* possibility one day they
might believe it if a robot actually behaves
intelligently, in the human sense.

>> When you write "computers can't ... whatever"
>> it makes as much sense as saying a collection
>> of amino acids, DNA etc cannot make a human
>> because all we have so far is a virus that
>> cannot do anything by itself.
>
>
> Yeah, yeah, maybe you could point out exactly
> where I said "computers can't . . ."

Your topic line was "Computers as the Philosopher's
Stone of AI". I thought you meant that just as the
philosopher's stone can't turn something to gold,
computers can't turn something into intelligence.
If I misunderstood your meaning I apologize.

[...]

Regards - John

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 19:14:22 GMT

On 10 Jan 2005 10:31:35 -0800, "JGCASEY" in
comp.ai.philosophy wrote:

>
>Lester Zick wrote:
>> On 9 Jan 2005 19:16:18 -0800, "JGCASEY" in
>> comp.ai.philosophy wrote:
>[...]
>
>>> People say the computer this, the computer that,
>>> when they mean software and they generalize,
>>> from what that particular piece (or pieces)
>>> of software do, about what is or is not possible.
>>
>> It doesn't change the terms of the observations
>> I made to change the term hardware to software.
>> How about if I just change the original
>> statements to:
>>
>> Ware is useful, which nobody doubts, and
>>
>> Ware is intelligent, which nobody believes.
>>
>> Then you can argue to your hearts content
>> over whether qualifying the term ware as
>> hard or soft changes the nature of the
>> argument.
>
>No, nothing to do with hard or soft. I was
>pointing out that a "computer" was a collection
>of parts and wasn't anything useful until
>those parts were put together to make some
>kind of machine.
>
>> Nor do I try to generalize from
>> particular ware. That's your claim.
>
>That comes from your second statement, "ware
>is intelligent, which nobody believes."

Just the statement of alternatives for forensic purposes.

>People don't believe ware is intelligent for
>the simple reason software isn't intelligent
>in the human sense at this point of time.
>There is a *third* possibility one day they
>might believe it if a robot actually behaves
>intelligently, in the human sense.

This is exactly the crux of the problem. People say that there is a
third alternative without knowing what they mean. The problem is
that if there is such an alternative, it is already implicit in what
we have today. I don't say that intelligence cannot be implemented on
computers ever. I do say that based on current state of the art
approaches to the mechanization of intelligence we will never know
when, whether, of if what is implemented on computers is intelligence
or not. And I'm quite certain there are those who fully expect that if
implementations are sufficiently complex, no one will ever be able to
tell one way or the other. Many of those who advocate Turing criteria
for intelligent machines base their claims on such notions.

>>> When you write "computers can't ... whatever"
>>> it makes as much sense as saying a collection
>>> of amino acids, DNA etc cannot make a human
>>> because all we have so far is a virus that
>>> cannot do anything by itself.
>>
>>
>> Yeah, yeah, maybe you could point out exactly
>> where I said "computers can't . . ."
>
>Your topic line was "Computers as the Philosopher's
>Stone of AI". I thought you meant that just as the
>philosopher's stone can't turn something to gold,
>computers can't turn something into intelligence.
>If I misunderstood your meaning I apologize.

Well, the title line only states a theme. I don't know that a
philosopher's stone can't turn base metal into gold. I just know it
represents an apt analogy to the way people tend to regard computers
as substitutes for the proper analysis of a subject like intelligence.

Regards - Lester

From:	Stargazer
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 17:39:36 -0200

Lester Zick wrote:
> This is exactly the crux of the problem. People say that there is a
> third alternative without knowing what they mean. The problem is
> that if there is such an alternative, it is already implicit in what
> we have today. I don't say that intelligence cannot be implemented on
> computers ever. I do say that based on current state of the art
> approaches to the mechanization of intelligence we will never know
> when, whether, of if what is implemented on computers is intelligence
> or not.

Some time ago, a dog named "Rico" made the news worldwide. It is a
"special dog", able to differentiate the meaning of nearly 200 words.

http://www.guardian.co.uk/life/thisweek/story/0,12977,1239988,00.html

Any robot (built with today's "state of the art approaches to
mechanization of intelligence") that happens to present behavior
comparable to Rico's is, in my vision, a demonstration that
intelligence can be created with technology the way it is being
developed today.

*SG*

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Mon, 10 Jan 2005 22:07:34 GMT

On Mon, 10 Jan 2005 17:39:36 -0200, "Stargazer"
in comp.ai.philosophy wrote:

>Lester Zick wrote:
>> This is exactly the crux of the problem. People say that there is a
>> third alternative without knowing what they mean. The problem is
>> that if there is such an alternative, it is already implicit in what
>> we have today. I don't say that intelligence cannot be implemented on
>> computers ever. I do say that based on current state of the art
>> approaches to the mechanization of intelligence we will never know
>> when, whether, of if what is implemented on computers is intelligence
>> or not.
>
>Some time ago, a dog named "Rico" made the news worldwide. It is a
>"special dog", able to differentiate the meaning of nearly 200 words.

I'm not sure I'd be interested. I already know 200 words.

>http://www.guardian.co.uk/life/thisweek/story/0,12977,1239988,00.html
>
>Any robot (built with today's "state of the art approaches to
>mechanization of intelligence") that happens to present behavior
>comparable to Rico's is, in my vision, a demonstration that
>intelligence can be created with technology the way it is being
>developed today.

I think we can safely conclude there are programs able to distinguish
200 words, in which case we already have artificial intelligence. This
is just an argument from Turing's criterion, which says as I recollect
that an inability to say something is not intelligent is equivalent to
saying that it is intelligent. Very peculiar rationalization. Typical
positivistic rationale. Now using analogous reasoning we can get to
work on prime numbers and pi.

Regards - Lester

From:	Stargazer
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Tue, 11 Jan 2005 09:11:44 -0200

Lester Zick wrote:
> On Mon, 10 Jan 2005 17:39:36 -0200, "Stargazer"
> in comp.ai.philosophy wrote:
>
> > Lester Zick wrote:
> > > This is exactly the crux of the problem. People say that there is
> > > a third alternative without knowing what they mean. The problem is
> > > that if there is such an alternative, it is already implicit in
> > > what we have today. I don't say that intelligence cannot be
> > > implemented on computers ever. I do say that based on current
> > > state of the art approaches to the mechanization of intelligence
> > > we will never know when, whether, of if what is implemented on
> > > computers is intelligence or not.
> >
> > Some time ago, a dog named "Rico" made the news worldwide. It is a
> > "special dog", able to differentiate the meaning of nearly 200
> > words.
>
> I'm not sure I'd be interested. I already know 200 words.
>
> > http://www.guardian.co.uk/life/thisweek/story/0,12977,1239988,00.html
> >
> > Any robot (built with today's "state of the art approaches to
> > mechanization of intelligence") that happens to present behavior
> > comparable to Rico's is, in my vision, a demonstration that
> > intelligence can be created with technology the way it is being
> > developed today.
>
> I think we can safely conclude there are programs able to distinguish
> 200 words, in which case we already have artificial intelligence. This
> is just an argument from Turing's criterion, which says as I recollect
> that an inability to say something is not intelligent is equivalent to
> saying that it is intelligent. Very peculiar rationalization. Typical
> positivistic rationale. Now using analogous reasoning we can get to
> work on prime numbers and pi.

You don't get it, Lester. Rico is a dog and it doesn't say anything.
It, however, presents behavior (not verbal!) that demonstrates
understanding: "Come on, Rico, pick up the red fruit". And it goes
there and, among several fruits, it picks the reddish (fruit!). We don't
have a robot that does the same thing, and I daresay that most would
consider such a robot intelligent, were it to present such a behavior.

*SG*

From:	Lester Zick
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	Tue, 11 Jan 2005 16:36:24 GMT

On Tue, 11 Jan 2005 09:11:44 -0200, "Stargazer"
in comp.ai.philosophy wrote:

>Lester Zick wrote:
>> On Mon, 10 Jan 2005 17:39:36 -0200, "Stargazer"
>> in comp.ai.philosophy wrote:
>>
>> > Lester Zick wrote:
>> > > This is exactly the crux of the problem. People say that there is
>> > > a third alternative without knowing what they mean. The problem is
>> > > that if there is such an alternative, it is already implicit in
>> > > what we have today. I don't say that intelligence cannot be
>> > > implemented on computers ever. I do say that based on current
>> > > state of the art approaches to the mechanization of intelligence
>> > > we will never know when, whether, of if what is implemented on
>> > > computers is intelligence or not.
>> >
>> > Some time ago, a dog named "Rico" made the news worldwide. It is a
>> > "special dog", able to differentiate the meaning of nearly 200
>> > words.
>>
>> I'm not sure I'd be interested. I already know 200 words.
>>
>> > http://www.guardian.co.uk/life/thisweek/story/0,12977,1239988,00.html
>> >
>> > Any robot (built with today's "state of the art approaches to
>> > mechanization of intelligence") that happens to present behavior
>> > comparable to Rico's is, in my vision, a demonstration that
>> > intelligence can be created with technology the way it is being
>> > developed today.
>>
>> I think we can safely conclude there are programs able to distinguish
>> 200 words, in which case we already have artificial intelligence. This
>> is just an argument from Turing's criterion, which says as I recollect
>> that an inability to say something is not intelligent is equivalent to
>> saying that it is intelligent. Very peculiar rationalization. Typical
>> positivistic rationale. Now using analogous reasoning we can get to
>> work on prime numbers and pi.
>
>You don't get it, Lester. Rico is a dog and it doesn't say anything.
>It, however, presents behavior (not verbal!) that demonstrates
>understanding: "Come on, Rico, pick up the red fruit". And it goes
>there and, among several fruits, it picks the reddish (fruit!). We don't
>have a robot that does the same thing, and I daresay that most would
>consider such a robot intelligent, were it to present such a behavior.

Well, I've given the subject some more consideration, and I would
like to clear up some problems. First, I don't say that dogs aren't
intelligent. In fact just the reverse is true. All kinds of sentient
beings are intelligent including animals and plants. However, I do say
that intelligence in plants is not the same as in animals nor animal
intelligence the same (to the best of our knowledge) as in in homo s.

Second, your characterization of dogs' understanding of our verbal
concepts is pure anthropomorphosis, Stargazer. I don't say animals
like dogs have no understanding of concepts. In fact I think a dog
probably knows a lot more than a mere two hundred concepts. More
likely on the order of thousands.The criterion for human understanding
though is what one does with the concepts one understands. In other
words, intelligence in the human sense implies a mechanical reduction
of concepts in relation to one another that I don't see in animal
intelligence or at least only in haphazard ways.

A suitable test for mechanical reduction in such contexts would be the
ability to count thingss. We know many animals which can count one,
two, three, many, but not many if any which can count N events. So if
asked to get 43 shoes, I doubt there are many dogs or other animals
which could understand the request. (By the way, I doubt a dog could
distinguish red or reddish fruit in any event because as I recollect a
dog's vision is black and white.)

In other words the only relevant test for intelligence is the ability
to frame and draw knowledge through mechanical reduction. My
problem with the mechanical reductions in robots is that they are not
reductions of the robots but of the programmers instructing robots to
make them and how.

In other words, intelligence resides in the methods and methodology
and not in the things used to implement the methodology. Computers
are no more intelligent in this respect than the carbon, oxygen, etc.
molecules we use to mechanize our intelligence. We just don't know
enough of what it means to be intelligent to program intelligence on
machines yet. Computers are just tools in any event. And the typical
empiricist/positivist reliance on the inability to disprove doesn't
make them anything else.

Regards - Lester

From:	Kobrinsky
Subject:	Re: Computers as the Philosopher's Stone of AI
Date:	10 Jan 2005 09:09:37 -0800

JGCASEY:
> You make statements such as "the proposition that computers
> are intelligent" and I was making the point computers aren't
> anything. People say the computer this, the computer that,
> when they mean software and they generalize, from what that
> particular piece (or pieces) of software do, about what is
> or is not possible.

Computer can be brought to abilities that men just can
dream of. Thats their purpose. Its the system of men
and its diffenrent ways of perfectionism mens activity,
that some time will be called 'intelligent'. Intelligence
is more than each part of this men-engine-system.

> When you write "computers can't ... whatever" it makes
> as much sense as saying a collection of amino acids, DNA
> etc cannot make a human because all we have so far is a
> virus that cannot do anything by itself.
>
> A computer has to be "wired up" to be something. This
> is called programming which connects the various parts
> of the computer together.
>
> A general purpose computer can be programmed (wired up)
> to become isomorphic with any dynamic system *providing*
> the computer has enough components that work fast enough
> and we know how the system in question works.

Right, let me specify: a computer [also software or action itself]
cannot be defined as 'intelligent'. Means 'intelligence' just can
cristallice out of a *given* context. The situation that applies.
If computer/software is then possible to follow or [if its made able
to] create commands that specified as correct solution to exactly
*this* constellation of problem, it will be defined as an
-> intelligent /handling/ .

> I don't think we have to figure out what intelligence
> means so much as how we can get a machine to do the
> things that fall into the category of intelligence.

Well, in other words, the exact reaction out of a given pack of human
problems makes them almost right[intelligent]. Possible ways will
provide solution that organized more or fever created by mens mind.
So strictly speaking, only the prosurvival composition of (re-)
actions out of a real problemship, its the hole system [all
components of handling - may include computing - to solute a problem
as best as actual possible] that defines the part of used intelligence.

Any handling itself (and may it called the 'best') is simply stupid!
There ever will be unique problemships thought out by menkind and
only an ideal created complex of people AND machine can make a system
called intelligent.

I. Kobrinsky.