Should Robots Replace Humans?



 Religions > Atheism > Should Robots Replace Humans?

LINK TO THIS PAGE  


rating :  0   |  0


  Page 1 of 2

1

 

2

 
Topic: Religions > Atheism
User: ""
Date: 01 Aug 2005 05:14:59 PM
Object: Should Robots Replace Humans?
In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?
Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.
Some sources:
http://www.transtopia.org/
http://www.neoeugenics.com/
.

User: "Denis Loubet"

Title: Re: Should Robots Replace Humans? 01 Aug 2005 07:25:44 PM
<xmarkjonesx@yahoo.com> wrote in message
news:1122934499.687887.39720@g14g2000cwa.googlegroups.com...

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

But if you're going to propose super-advanced robotics, then why not propose
super-advanced medical technology that allows humans to be upgraded.
--
Denis Loubet
dloubet@io.com
http://www.io.com/~dloubet
http://www.ashenempires.com
.
User: "sam ende"

Title: Re: Should Robots Replace Humans? 04 Aug 2005 07:05:58 PM
Denis Loubet wrote:

But if you're going to propose super-advanced robotics, then why not
propose super-advanced medical technology that allows humans to be
upgraded.

we tried that;
http://www.chiprowe.com/articles/bionic.html
sammi
.


User: "Fred Stone"

Title: Re: Should Robots Replace Humans? 01 Aug 2005 08:39:46 PM
wrote in
news:1122934499.687887.39720@g14g2000cwa.googlegroups.com:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase
and simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news:
Out with the old, in with the new.

Impractical for several reasons, not least of which is the lack of
flexibility of identical pre-programmed "adult" minds. The human species
benefits from the diversity of its individuals' mental faculties, in
spite of the occasional irrationality that results from our early
learning experiences.
--
Fred Stone
aa# 1369
Political correctness is a form of moral and emotional blackmail
whereby if you call a spade a spade you must apologize to the spade.
.
User: ""

Title: Re: Should Robots Replace Humans? 01 Aug 2005 08:54:40 PM

Impractical for several reasons, not least of which is the lack of
flexibility of identical pre-programmed "adult" minds. The human species
benefits from the diversity of its individuals' mental faculties, in
spite of the occasional irrationality that results from our early
learning experiences.

Humans are pre-programmed as well by nature. We too are all
mechanical, like robots, we just have more complex programming and
better hardware for the time being, but technology may change that. I
would rather my children by silicon-based than organic.
.
User: "sam ende"

Title: Re: Should Robots Replace Humans? 04 Aug 2005 07:02:20 PM
wrote:

Humans are pre-programmed as well by nature. We too are all
mechanical, like robots, we just have more complex programming and
better hardware for the time being, but technology may change that.
I would rather my children by silicon-based than organic.

http://www.healthy.net/scr/article.asp?ID=2067
sammi
.



User: "R. Pierce Butler"

Title: Re: Should Robots Replace Humans? 01 Aug 2005 05:23:34 PM
wrote in news:1122934499.687887.39720
@g14g2000cwa.googlegroups.com:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.

Some sources:

http://www.transtopia.org/
http://www.neoeugenics.com/


That works for me. Let me know when you have volunteered to be
assimilated.
rj
.

User: "Paul Hovnanian P.E."

Title: Re: Should Robots Replace Humans? 07 Aug 2005 09:10:22 PM
wrote:


In humans, the higher the IQ, the less children they have.

Not necessarily. The higher IQ males impregnate more females, but find
ways to get the lower IQ males to pick up the child support, thinking
that the offspring belong to them.
--
Paul Hovnanian mailto:Paul@Hovnanian.com
------------------------------------------------------------------
668: The Neighbor of the Beast
.

User: "sam ende"

Title: Re: Should Robots Replace Humans? 04 Aug 2005 06:32:14 PM
wrote:

In humans, the higher the IQ, the less children they have.

really ?
sammi
.

User: "Richo"

Title: Re: Should Robots Replace Humans? 02 Aug 2005 01:37:47 AM
wrote:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Define "better".
Better for whom?
Mark.
.

User: "gigantul"

Title: Re: Should Robots Replace Humans? 28 Aug 2005 03:34:28 AM
wrote:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.

Some sources:

http://www.transtopia.org/
http://www.neoeugenics.com/

I think to achieve human consciousness from a CPU, we need a lot more
development of processors. CPU is a single processor that can do
millions of calculations in a second. However, the way the human mind
works is totally different. Human brain is a collection of billions of
processors(albait sluggish and inaccurate) working in parallel. even
the most advanced supercomputer cannot be more interconnected with its
processors than neural network of an insect. As intel dropped the idea
of increasing speed of the processor, and trying to produce a computer
with more of parallel processing, this might be achievable but i guess
not anytime soon as much as 30 years.
.
User: "Hans-Georg Michna"

Title: Re: Should Robots Replace Humans? 28 Aug 2005 11:06:50 AM
On 28 Aug 2005 01:34:28 -0700, "gigantul" <gigantul@gmail.com>
wrote:

I think to achieve human consciousness from a CPU, we need a lot more
development of processors. CPU is a single processor that can do
millions of calculations in a second. However, the way the human mind
works is totally different. Human brain is a collection of billions of
processors(albait sluggish and inaccurate) working in parallel. even
the most advanced supercomputer cannot be more interconnected with its
processors than neural network of an insect. As intel dropped the idea
of increasing speed of the processor, and trying to produce a computer
with more of parallel processing, this might be achievable but i guess
not anytime soon as much as 30 years.

More like 5 to 15 years. A serial processor can emulate a
network of parallel processors if it is fast enough, so the two
approaches are equivalent or can be made equivalent.
We will apparently get more parallelism soon, because single
processor speed is hitting some physical limits. On the other
hand, there may be new technologies like optical computing that
increase the speed again.
Hans-Georg
--
No mail, please.
.


User: ""

Title: Re: Should Robots Replace Humans? 01 Aug 2005 08:56:40 PM
wrote:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Better at doing what?
Consciousness is what it is all about,
of pleasure and of pain. The ability to
live happy lives and to pass this on to
future generations. This is what makes
life worth living.
If robots were to reproduce it is possible
that after many generation consciousness
might arise. Humans of coarse would be
steering the evolution of robots for traits
that benefit man. The most probable out come
would be that robots take pleasure in serving
human beings.
Larry
.
User: "sam ende"

Title: Re: Should Robots Replace Humans? 04 Aug 2005 06:55:50 PM
wrote:

The most probable out come
would be that robots take pleasure in serving
human beings.

well i think we ought to insist on that :)
sammi
.

User: "Uncle Buck"

Title: Re: Should Robots Replace Humans? 01 Aug 2005 10:39:43 PM
On 1 Aug 2005 18:56:40 -0700,
wrote:

If robots were to reproduce it is possible
that after many generation consciousness
might arise. Humans of coarse would be
steering the evolution of robots for traits
that benefit man. The most probable out come
would be that robots take pleasure in serving
human beings.

Not necessarily. We don't know, first of all, what it would take to give birth
to "consciousness" - it's entirely possible that the simplest of programs are in
some way "conscious" (at least if they're on analog systems). But secondly,
robots wouldn't even need to be conscious to become intelligent enough to begin
steering their own evolution and thus, ending our control of their existence.
--
L8r,
Uncle Buck
BAAWA
a.a. #88
UU Minister # "=" - "Grand Equivocator of the Balanced Equal Sign"
~=^*=^=~-=^=-~=^=*^=~=-^=-=~^=*=^~=-=^-=~=^*=^=~-=^=-~=^=*^=~=-^=-=~^=*=^~=
"Surrendering To The Fall" - A blog about - what else? - me:
http://surrenderingtothefall.blogspot.com
(perpetually under construction)
~=^*=^=~-=^=-~=^=*^=~=-^=-=~^=*=^~=-=^-=~=^*=^=~-=^=-~=^=*^=~=-^=-=~^=*=^~=
.
User: ""

Title: Re: Should Robots Replace Humans? 02 Aug 2005 02:08:55 PM
Uncle Buck wrote:

On 1 Aug 2005 18:56:40 -0700,

wrote:

If robots were to reproduce it is possible
that after many generation consciousness
might arise. Humans of coarse would be
steering the evolution of robots for traits
that benefit man. The most probable out come
would be that robots take pleasure in serving
human beings.


Not necessarily. We don't know, first of all, what it would take to
give birth to "consciousness" - it's entirely possible that
the simplest of programs are in some way "conscious"

Well that might be however then so might rocks and
trains as well as other dead things. The fact is however
that there is no evidence of consciousness.
If robots could create themselves it is likely that
the design could be modified, then favorable
characteristics could be passed on. If emotions
and consciousness are deemed favorable then
it is likely that they would develop.

(at least if they're on analog systems). But secondly,
robots wouldn't even need to be conscious to become
intelligent enough to begin
steering their own evolution and thus, ending our
control of their existence.

The design of self building robots, would have to start by human
beings. We would be there from the beginning selecting which
robot types we liked best and constantly guiding the evolution
like dogs or cats. Through the years we have turned wolves
into mans best friend. I think it would be most likely that
If robots did develop consciousness, that they would friendly
to humans because they had been steered that direction.
Larry
.



User: "Graham Kennedy"

Title: Re: Should Robots Replace Humans? 01 Aug 2005 07:29:24 PM
wrote:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

I wish a robot would replace me. I could get a bit of peace then.
.

User: "Dante Alighieri"

Title: Re: Should Robots Replace Humans? 05 Aug 2005 11:42:52 PM
Approx. 1 Aug 2005 15:14:59 -0700, someone calling themselves
xmarkjonesx@yahoo.com let their cat run across the keyboard resulting in:
:-)= Low fertility of robots is not a problem
Humans make more robots?
Reminds me of the federal/private fund grantee who developed a
a walking robot that had to learn to walk all over again every time
time it was turned on. A program interested in developing a robot that
that walked like a human. That particular participant in the program
program certainly demonstrated human behavior at its most human
human level. Why did the inventor not develop a robot that would
would remember how to walk once it had learned. Similarly why not
not develop a robot for each stage of an assembly line that made
made robots? Have the robots equiped with vision and with progs
progs that would interpret the digital video input making them more
more efficient because they could "see" what they were doing thus
thus making them capable (with more software) of detecting an error
error and making them capable of seeing something if it dropped and
and with proper programming identify the dropped item and know what
what should be done with the dropped article. Did you ever think about
about that, now, didja?
HTH
YAWIA
BTW
The most amusing sight in the world is a human being
acting just like a human being.
.

User: "Hans-Georg Michna"

Title: Re: Should Robots Replace Humans? 03 Aug 2005 12:42:24 PM
On 1 Aug 2005 15:14:59 -0700,
wrote:

In humans, the higher the IQ, the less children they have.

Questionable. Doesn't seem to be entirely true. But is also
unimportant.

Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.

The question is intelligence. As long as robots are as stupid as
they are, they are of correspondingly limited use. As machine
intelligence rises, they will become more useful and more
widespread. I already have two, a vacuum cleaner robot, the
ubiquitous Roomba, and a lawn mower robot by Friendly Robotics,
Israel. But they are still quite stupid.
As machine intelligence rises further, the problem of surpassing
human intelligence arises, but that's another matter, discussed
in www.michna.com/transition.htm , among other places.
Hans-Georg
--
No mail, please.
.

User: "Dubh Ghall"

Title: Re: Should Robots Replace Humans? 02 Aug 2005 09:29:29 AM
On 1 Aug 2005 15:14:59 -0700,
wrote:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.

You put me in mind of a turkey, that is looking forward to xmas.
--
Puck Greenman
The spelling, Like any opinion stated here,
is purely my own
#162 BAAWA Knight.
Plonked by Rob Duncan

Na bister 500,000
.

User: "smozoma"

Title: Re: Should Robots Replace Humans? 02 Aug 2005 01:27:24 AM
On 1 Aug 2005 15:14:59 -0700,
wrote:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

What is "better," though? Robots are better at building cars, because
they are faster, more accurate, and so on. Maybe one day they will be
better at designing the cars, even.
However, a human has no inherent purpose, or "job," so if you
"replace" a human with a robot, then you have failed, because a robot
can not be a better human than a human can. ;) They can be better at
things humans do, but why do they need to replace us, just because
they are better than us at everything we do? The world isn't going to
lay us off.
It's a bit like humans and neanderthals. There was no "should" about
neanderthals dying out. Humans happened to survive and neanderthals
happened not to have.
--
Michael Capewell
--
I aim simply to entertain. Me.

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.

Some sources:

http://www.transtopia.org/
http://www.neoeugenics.com/

.
User: "smozoma"

Title: Re: Should Robots Replace Humans? 02 Aug 2005 01:51:15 AM
On Tue, 02 Aug 2005 02:27:24 -0400, smozoma
<smozomaNOSPAM@yahooBLAHBLAH.com> wrote:

On 1 Aug 2005 15:14:59 -0700,

wrote:

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?


What is "better," though? Robots are better at building cars, because
they are faster, more accurate, and so on. Maybe one day they will be
better at designing the cars, even.

However, a human has no inherent purpose, or "job," so if you
"replace" a human with a robot, then you have failed, because a robot
can not be a better human than a human can. ;) They can be better at
things humans do, but why do they need to replace us, just because
they are better than us at everything we do? The world isn't going to
lay us off.

It's a bit like humans and neanderthals. There was no "should" about
neanderthals dying out. Humans happened to survive and neanderthals
happened not to have.

Actually, it looks like the world laid the neanderthals off, doesn't
it? ;) But that's not because we were better than them at being
neanderthals than they were...
--
Michael Capewell
.


User: "Michael Gray"

Title: Re: Should Robots Replace Humans? 01 Aug 2005 06:45:08 PM
On 1 Aug 2005 15:14:59 -0700,
wrote:

Should Robots Replace Humans?

Visit any church, and you'll discover that they already have.
.

User: "Dave"

Title: Re: Should Robots Replace Humans? 28 Aug 2005 11:12:57 AM
Plenty of brainless bots on Usenet. You might be one.
.

User: "Mal Osborne"

Title: Re: Should Robots Replace Humans? 02 Aug 2005 07:01:17 AM
Robots are *STUPID*. For instance, no robot yet devised can use vision to
locate a path through complex surroundings and navigate along it. A brain
damaged sheep is generations ahead in performing this simple task
<xmarkjonesx@yahoo.com> wrote in message
news:1122934499.687887.39720@g14g2000cwa.googlegroups.com...

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.

Some sources:

http://www.transtopia.org/
http://www.neoeugenics.com/

.
User: ""

Title: Re: Should Robots Replace Humans? 02 Aug 2005 10:56:09 AM
Mal Osborne wrote:

Robots are *STUPID*. For instance, no robot yet devised can use vision to
locate a path through complex surroundings and navigate along it. A brain
damaged sheep is generations ahead in performing this simple task

Profoundly true; the really interesting thing is why no one has noticed
that intelligent machines are only found in fiction, despite 40 years
of well-funded AI research.

<xmarkjonesx@yahoo.com> wrote in message
news:1122934499.687887.39720@g14g2000cwa.googlegroups.com...

In humans, the higher the IQ, the less children they have. Also,
humans can't be upgraded, they are irrational, they are weak, and so
forth. Would a new species of AI robotics be better than our current
dysgenic human evolution?

Low fertility of robots is not a problem: simply manufacture more of
them in the factory production line. Skip the whole children phase and
simply create fully grown adult robots with all the education
programmed into their silicon brains already. Humans are old news: Out
with the old, in with the new.

Some sources:

http://www.transtopia.org/
http://www.neoeugenics.com/

.
User: "Hans-Georg Michna"

Title: Re: Should Robots Replace Humans? 03 Aug 2005 12:42:23 PM
On 2 Aug 2005 08:56:09 -0700,
wrote:

Mal Osborne wrote:

Robots are *STUPID*. For instance, no robot yet devised can use vision to
locate a path through complex surroundings and navigate along it. A brain
damaged sheep is generations ahead in performing this simple task

Profoundly true; the really interesting thing is why no one has noticed
that intelligent machines are only found in fiction, despite 40 years
of well-funded AI research.

Some have. The reason is simply that the required data
processing performance has not been available and is still not
available.
Current estimates are (see www.KurzweilAI.net, among others)
that the first supercomputer will reach the performance of a
human brain around 2010, while a cheap personal computer will
reach it roughly around 2020.
More info: www.michna.com/transition.htm
Hans-Georg
Including alt.extropians in the crosspost now.
--
No mail, please.
.
User: "Mal Osborne"

Title: Re: Should Robots Replace Humans? 08 Aug 2005 07:55:53 AM
"Hans-Georg Michna" <hans-georgNoEmailPlease@michna.com> wrote in message
news:1pu1f1has7r33epgds1jojbsh6hrj4nn3b@4ax.com...

On 2 Aug 2005 08:56:09 -0700,

wrote:

Mal Osborne wrote:


Robots are *STUPID*. For instance, no robot yet devised can use vision
to
locate a path through complex surroundings and navigate along it. A
brain
damaged sheep is generations ahead in performing this simple task


Profoundly true; the really interesting thing is why no one has noticed
that intelligent machines are only found in fiction, despite 40 years
of well-funded AI research.


Some have. The reason is simply that the required data
processing performance has not been available and is still not
available.

Current estimates are (see www.KurzweilAI.net, among others)
that the first supercomputer will reach the performance of a
human brain around 2010, while a cheap personal computer will
reach it roughly around 2020.

More info: www.michna.com/transition.htm

Hans

I disagree.
A personal computer running 1000 times as fast as it does today will open up
Excel spreadsheets really, really fast. It will be no more self-aware than
the laptop I am typing on now, or the brick holding the door open. We have
NO IDEA how to program in consciousness, and no promising technologies
opening up.
.
User: "Hans-Georg Michna"

Title: Re: Should Robots Replace Humans? 10 Aug 2005 08:41:20 AM
On Mon, 8 Aug 2005 20:55:53 +0800, "Mal Osborne"
<noone@nowhere.com> wrote:

I disagree.

A personal computer running 1000 times as fast as it does today will open up
Excel spreadsheets really, really fast. It will be no more self-aware than
the laptop I am typing on now, or the brick holding the door open. We have
NO IDEA how to program in consciousness, and no promising technologies
opening up.

Mal,
first of all, I have no proof otherwise, and probably nobody
has, one way or another. But we can speculate.
Let me just cite from http://www.michna.com/transition.htm ,
from the end of the chapter "Instinct and Intelligence":
But does high data processing performance always lead to
intelligence? Not necessarily, but we have two good reasons to
assume that very high performance will assist us in creating
genuine intelligence in two different ways.
First, raw performance translates into intelligence in "brute
force" algorithms (simply trying and calculating a large number
of possibilities). In certain areas, for example, chess playing,
brute force can easily be more successful than the human way of
thinking. Given enough performance, many more areas could be
covered by this relatively simple method.
Second, with rising performance computers will assist their own
programming, and later, as they gain the ability to understand
written human language, increasingly program themselves. As soon
as we accept that in the computer and software industry the
structural and programming complexity of computers will grow
along with raw performance, it follows that intelligence will
rise and eventually reach and exceed human levels.
Hans-Georg
--
No mail, please.
.
User: ""

Title: Re: Should Robots Replace Humans? 10 Aug 2005 10:45:08 AM
Hans-Georg Michna wrote:

..does high data processing performance always lead to
intelligence? Not necessarily, but we have two good reasons to
assume that very high performance will assist us in creating
genuine intelligence in two different ways.

First, raw performance translates into intelligence in "brute
force" algorithms (simply trying and calculating a large number
of possibilities). In certain areas, for example, chess playing,
brute force can easily be more successful than the human way of
thinking.

But programmed chess requires a utility function (ie a function that
assigns a quality score to a board position) that has to be
hand-crafted by humans; in the case of the most successful chess
program of all time - IBM's Deep Blue - the parameters of the utility
function were set by comparing its choices to moves from actual master
play, and adjusting the parameters until the program made the same
response as the grandmaster. Likewise, its opening moves were directly
taken from standard openings worked out by humans over the centuries.
It is therefore an example of the automation of "the human way of
thinking", not of machine intelligence. It may be a world beater at
chess, but it could not learn to play any other game, even tic-tac-toe.

Given enough performance, many more areas could be
covered by this relatively simple method.

Other games, maybe, but no area that requires learning at the level of
human competency.

Second, with rising performance computers will assist their own
programming

This is where you go off the deep end into pure science fiction. Where
is the evidence that computers can program themselves in any
substantial way? Serious programming requires precisely the
capabilities that computers lack, so how can they get it by programming
themselves?

and later, as they gain the ability to understand
written human language

You are assuming a priori that understanding human written language is
just a matter of processing power. It is not. If that's all it were, it
would be possible now, even if it took a year for a machine to
understand a single page. Even such a slow rate of comprehension would
be a major advance in the AI world.
The problem with comprehending human language is not the problem of
understanding grammar and the definition of words; it is the problem of
grasping the way humans understand the world, which is implicit in
writing or utterance of any complexity. To understand what we say, you
have to be one of us.

increasingly program themselves. As soon
as we accept that in the computer and software industry the
structural and programming complexity of computers will grow

This may be where you are going wrong; though the modern desktop
hardware is vastly more complex than, say, that of an early IBM XT,
what it does is fundamentally the same, and, leaving aside add-on
features (graphics, audio), any program that a modern PC can run can
also be run, though far slower, on the XT.

along with raw performance, it follows that intelligence will
rise and eventually reach and exceed human levels.

Dream on.


Hans-Georg

--
No mail, please.

.
User: "Fred Stone"

Title: Re: Should Robots Replace Humans? 10 Aug 2005 11:45:52 AM
wrote in
news:1123688708.078149.305800@o13g2000cwo.googlegroups.com:

Hans-Georg Michna wrote:

..does high data processing performance always lead to
intelligence? Not necessarily, but we have two good reasons to
assume that very high performance will assist us in creating
genuine intelligence in two different ways.

First, raw performance translates into intelligence in "brute
force" algorithms (simply trying and calculating a large number
of possibilities). In certain areas, for example, chess playing,
brute force can easily be more successful than the human way of
thinking.


But programmed chess requires a utility function (ie a function that
assigns a quality score to a board position) that has to be
hand-crafted by humans; in the case of the most successful chess
program of all time - IBM's Deep Blue - the parameters of the utility
function were set by comparing its choices to moves from actual master
play, and adjusting the parameters until the program made the same
response as the grandmaster. Likewise, its opening moves were directly
taken from standard openings worked out by humans over the centuries.
It is therefore an example of the automation of "the human way of
thinking", not of machine intelligence. It may be a world beater at
chess, but it could not learn to play any other game, even
tic-tac-toe.

Given enough performance, many more areas could be
covered by this relatively simple method.


Other games, maybe, but no area that requires learning at the level of
human competency.

Second, with rising performance computers will assist their own
programming


This is where you go off the deep end into pure science fiction. Where
is the evidence that computers can program themselves in any
substantial way? Serious programming requires precisely the
capabilities that computers lack, so how can they get it by
programming themselves?

Neural networks and genetic algorithms can be combined to make a very
powerful learning "robot". Even with a very limited neural network being
"evolved" by the computer you can get quite sophisticated behaviors from
a game engine like this one:
http://nn.cs.utexas.edu/NERO/

and later, as they gain the ability to understand
written human language


You are assuming a priori that understanding human written language is
just a matter of processing power. It is not. If that's all it were,
it would be possible now, even if it took a year for a machine to
understand a single page. Even such a slow rate of comprehension would
be a major advance in the AI world.

Some articles at the academic site linked to that game site discuss
computer understanding of stories, their ability to answer questions
about what they just read, and the ability to distinguish between
different variations of the same story versus slightly different
stories.

The problem with comprehending human language is not the problem of
understanding grammar and the definition of words; it is the problem
of grasping the way humans understand the world, which is implicit in
writing or utterance of any complexity. To understand what we say, you
have to be one of us.

Or at least be able to *model* one of us.

increasingly program themselves. As soon
as we accept that in the computer and software industry the
structural and programming complexity of computers will grow


This may be where you are going wrong; though the modern desktop
hardware is vastly more complex than, say, that of an early IBM XT,
what it does is fundamentally the same, and, leaving aside add-on
features (graphics, audio), any program that a modern PC can run can
also be run, though far slower, on the XT.

If you don't count memory limitations and processor instruction sets.
But that's neither here nor there. Human mental architecture is notably
different from that of a sequential computer, but not to the extent that
a simulation is not possible.

along with raw performance, it follows that intelligence will
rise and eventually reach and exceed human levels.


Dream on.

Time will tell.
--
Fred Stone
aa# 1369
"Why, it appears that we appointed all of our worst generals
to command the armies and we appointed all of our best generals
to edit the newspapers. I mean, I found by reading a newspaper
that these editor generals saw all of the defects plainly from
the start but didn't tell me until it was too late. I'm willing
to yield my place to these best generals and I'll do my best
for the cause by editing a newspaper."
Robert E. Lee
.








  Page 1 of 2

1

 

2

 


Related Articles
 

NEWER

pg.3585     pg.2749     pg.2106     pg.1612     pg.1232     pg.940     pg.716     pg.544     pg.412     pg.311     pg.234     pg.175     pg.130     pg.96     pg.70     pg.50     pg.35     pg.24     pg.16     pg.10     pg.6     pg.3     pg.1

OLDER