Re: How Robots Will Steal Your Job

[Roedy Green]

Well of course awareness and consciousness are not the same in my
book.

Awareness is not all that interesting. Even the most primitive robot
could be aware you stepped on its toe, just because a pressure sensor
tripped.

Self consciousness is not all that interesting. It is just one of the
bazillion pet tricks you can do with a decent brain when you have
nothing better to do.

What is fascinating is that sense of experience that appears inside
myself when I am not under anaesthesia. This is something quite other
than computation and awareness.

It is important because morality is based on the notion it is ok to be
"cruel" to rocks because they are unconscious, and not ok to humans
because they are. Nothing would actually feel. There would be no
pain, just sham pain.

So I define consciousness as the ability to actually FEEL pain.

This definition suggests a much wider inclusion -- any thing that
reacts AS IF in pain until shown otherwise.

 

[Roedy Green]

He
understands how similar are the brain organizations of higher mammals,
and wishes to understand what it is that makes humans different from
the animals,

One argument in favour of humans being special is the primitiveness of
our consciousness. It can't seem to pay attention to many things at a
time. It is like a primitive single task OS.

 

[Lester Zick]

At least Edelman's and Humphrey's ideas about primary and higher-order
[ala 1st + 2nd order] C include only living animals, and they also
allow higher-animals besides humans some semblance of C. No
circular-logic argumentation regarding thermostats, quantum state
changes, and computers .... plus a blind man can take some comfort in
thinking that the guide-dog leading him across the street in the
spectre of rush-hour traffic has some awareness of what it's up to.

Well of course awareness and consciousness are not the same in my
book. What I note first in the context of the faculty of consciousness
is self consciousness, a comment I make in a collateral reply to Roedy
Green on this subject.

I am also dissatisfied with the kinds of approaches to the analysis of
consciousness like those you describe above because they are
speculative in nature and offer no potential mechanical insight as to
why things are one way and not another.

This is the reason I draw a definite distinction between non conscious
animals and conscious beings like homo sapiens.

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


In point of fact, Edelman has been investigating and simulating
neuronal mechanisms of perception and consciousness for many years.
His models are heavily based upon neuroanatomy and neurophysiology. He
understands how similar are the brain organizations of higher mammals,
and wishes to understand what it is that makes humans different from
the animals, in terms of the underlying neuronal mechanisms. This
approach trumps all philosophical speculations ever made since the
beginning of time, and provides a basis for making the distinction he
does between primary and higher-order C.

This approach makes great sense at this point in time, and holds great
interest to computer jockeys such as myself. We have enormous amounts
of computer power available to use in creating models and simulations.
In contrast, wet neurophysiology is still stuck in the doldrums of
single-unit recordings, etc. So, we can either wait 50 more years for
neurophysiology techniques to tell us more, or else use what we
currently know to create computer simulations that might give us
additional insight. This is what Edelman has been doing.
====================

Or we could just develop a better grasp of differential mechanics
right now. If we move with neurophysiology we'll just wind up in fifty
years with better neurophysiology. Which isn't a bad thing if you're a
neurophysiologist. But it won't explain the mechanics or nature of
sentience or consciousness.

Which probably wouldn't be too far away from Edelman bipartate
classifications of primary and higher-order consciousness.
===============


Of course the issue as to whether homo


In which case, I think Edleman's classification scheme has already
trumped you.

Except that Edleman has no mechanics justify his classification scheme
in which case my mechanics and classification scheme trump his.
;-)


Regards - Lester

 

[Lester Zick]

Awareness is not all that interesting. Even the most primitive robot
could be aware you stepped on its toe, just because a pressure sensor
tripped.

Self consciousness is not all that interesting. It is just one of the
bazillion pet tricks you can do with a decent brain when you have
nothing better to do.

And how would you do it? If all you're talking about is making a bot
act as if it were self conscious just as you indicate below with
respect to pain we could just as easily make a bot act as if it were
in pain too. Would that make it conscious or self conscious? That's no
kind of standard at all.

What is fascinating is that sense of experience that appears inside
myself when I am not under anaesthesia. This is something quite other
than computation and awareness.

It is important because morality is based on the notion it is ok to be
"cruel" to rocks because they are unconscious, and not ok to humans
because they are. Nothing would actually feel. There would be no
pain, just sham pain.

So I define consciousness as the ability to actually FEEL pain.

This definition suggests a much wider inclusion -- any thing that
reacts AS IF in pain until shown otherwise.

Regards - Lester

 

[Bent C Dalager]

And how would you do it? If all you're talking about is making a bot
act as if it were self conscious just as you indicate below with
respect to pain we could just as easily make a bot act as if it were
in pain too. Would that make it conscious or self conscious? That's no
kind of standard at all.

With robots it's generally easy. We're capable of understanding the
mechanics behind their functioning and by examining these we can
easily conclude that it's just a cheap trick. It is easy to find, with
a robot for which we have the detailed blueprints, that it doesn't
actually _feel_ the pain, it just behaves as if it did.

This becomes much more difficult with living beings or, indeed, with a
robot for which we didn't have the blueprint and for which - most
importantly - we are unable to understand its underlying technology.
For this sort of entities it would seem prudent to give them the
benefit of the doubt and assume them conscious for their reaction to
pain until we can gather better data.

We could certainly be wrong for any given entity. We will remedy the
situation if and when this becomes clear to us.

Cheers
Bent D

 

[Lester Zick]

With robots it's generally easy. We're capable of understanding the
mechanics behind their functioning and by examining these we can
easily conclude that it's just a cheap trick. It is easy to find, with
a robot for which we have the detailed blueprints, that it doesn't
actually _feel_ the pain, it just behaves as if it did.

This becomes much more difficult with living beings or, indeed, with a
robot for which we didn't have the blueprint and for which - most
importantly - we are unable to understand its underlying technology.
For this sort of entities it would seem prudent to give them the
benefit of the doubt and assume them conscious for their reaction to
pain until we can gather better data.

We could certainly be wrong for any given entity. We will remedy the
situation if and when this becomes clear to us.

Cheers
Bent D

Thanks for the input Bent. But I don't know if we can trust the
opinions of anyone using emacs. Seriously, though, this is really what
I'd call an ostensible and not an analytically sufficient judgment.

Regards - Lester

 

[Eugeny Kornienko]

The direct result of the faculty of consciousness is the production of
abstract information. Abstract information is abstract in the sense
that it is abstract from a being's own being.

The original poster of the subject is frustrated with our discussion. From
our talk he couldn't grasp "what is intelligence". Why has the theme turn to
"nature of consciousness"? I think that we can differentiate some close
words that relate to intelligence. On my view consciousness actually helps
us in making decisions. Who operates with abstract information is
"intellect", not consciousness.

I agree with Roedy Green that consciousness is feeling. Sentient means
conscious. Sentience is the basic and simplest sign of consciousness.

The notion of "intellect" doesn't refer to capacity for feeling. Intellect
works by rules, uses language, makes logical corollaries from statements.

A conscious being, a simple animal may possess very weak or zero intellect.
Analogously, an apparent intellectual agent, the CYC, an expert system may
have tiny or zero consciousness.

At first, like other animals, human being had capacity to feel. When we
started to use language and make machines then we started to accumulate
"abstract" knowledge and intellect that is capacity to manipulate symbols.
Symbols don't occur in nature. Consciousness wishes something, intellect
think how to behave to realize the wish. Mind (consciousness plus intellect
plus experience) invents new symbols (words) for intellect to think better.

Consciousness operates with immediate feelings or images of feelings. They
are not abstract. I see (feel) symbols "sigma" and "integral". They are
patterns on paper, not abstract information. Intellect understands their
meaning, the knowledge that is implemented to these pictures. So intellect
operates with "abstract information".

Intellect logically compares possible decisions and choose better one. If
the decisions are "intellectually" equal then consciousness (free will,
want) works instead of intellect.

Let a robot can operate quite good in environment. In many circumstances
its programmed "artificial intellect" behaves properly. Battery is low - to
charge, it is dark - to turn light on etc. Some actions are difficult to
preview. So the designer implemented a "self-learning engine" into the
robot. This engine is to increase a value of a goal function by finding
right behavior when circumstances are logically equal for the intellectual
subsystem of the robot. Intellectual part of the robot behavior may be
explained by implemented knowledge of the designer. New learnt part of
behavior is a result of random search. There is no "logical" explanation of
the new behavior (without special investigation).

After a while robot has learnt to behave better (to choose better decisions)
in logically indistinguishable situations. It makes a decision because of a
"wish" of its self-learning subsystem. Robot wants to do such and such.
Intellectually it doesn't grasp reasons why such behavior is better.

While it is learning this robot have a chance to be a sentient being in
concern to the senses that are involved in learning.

When the new behavior is included in "intellectual engine" as a new rule the
behavior becomes automatic. If the robot is learning constantly then it may
be sentient always.

EK

 

[David Postill]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Green

| On 15 Jan 2004 14:25:17 -0800, (e-mail address removed) (dan michaels) wrote
| or quoted :
|
| >He
| >understands how similar are the brain organizations of higher mammals,
| >and wishes to understand what it is that makes humans different from
| >the animals,
|
| One argument in favour of humans being special is the primitiveness of
| our consciousness. It can't seem to pay attention to many things at a
| time. It is like a primitive single task OS.

That is a male brain specialisation.

There is plenty of supporting research for this.

The female brain can multitask far more efficiently than the male brain
(eg they can watch TV, do the ironing and have a phone conversation all
at the same time).

When watching an engrossing drama/thriller on TV I am incapable of doing
anything else - apart from drinking beer/wine

If I am working on the computer (I am a programmer) and I am "in the flow"
I just grunt when my wife says something (much to her annoyance). I can
listen to music at the same time though (for some variety of music, usually
techno/trance which is repetitive and doesn't require active listening).

<davidp />

- --
David Postill

-----BEGIN PGP SIGNATURE-----
Version: PGP 8.0.3 - not licensed for commercial use: www.pgp.com
Comment: Get key from pgpkeys.mit.edu:11370

iQA/AwUBQAh403xp7q1nhFwUEQJNuwCgx5gvBW/ZZs2Q1FR0DyqpGyH1ZDMAoMmW
HjQlUEy++0/+WAAfh7wzyiCX
=wcR0
-----END PGP SIGNATURE-----

 

[dan michaels]

At least Edelman's and Humphrey's ideas about primary and higher-order
[ala 1st + 2nd order] C include only living animals, and they also
allow higher-animals besides humans some semblance of C. No
circular-logic argumentation regarding thermostats, quantum state
changes, and computers .... plus a blind man can take some comfort in
thinking that the guide-dog leading him across the street in the
spectre of rush-hour traffic has some awareness of what it's up to.


Well of course awareness and consciousness are not the same in my
book. What I note first in the context of the faculty of consciousness
is self consciousness, a comment I make in a collateral reply to Roedy
Green on this subject.

I am also dissatisfied with the kinds of approaches to the analysis of
consciousness like those you describe above because they are
speculative in nature and offer no potential mechanical insight as to
why things are one way and not another.

This is the reason I draw a definite distinction between non conscious
animals and conscious beings like homo sapiens.

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


In point of fact, Edelman has been investigating and simulating
neuronal mechanisms of perception and consciousness for many years.
His models are heavily based upon neuroanatomy and neurophysiology. He
understands how similar are the brain organizations of higher mammals,
and wishes to understand what it is that makes humans different from
the animals, in terms of the underlying neuronal mechanisms. This
approach trumps all philosophical speculations ever made since the
beginning of time, and provides a basis for making the distinction he
does between primary and higher-order C.

This approach makes great sense at this point in time, and holds great
interest to computer jockeys such as myself. We have enormous amounts
of computer power available to use in creating models and simulations.
In contrast, wet neurophysiology is still stuck in the doldrums of
single-unit recordings, etc. So, we can either wait 50 more years for
neurophysiology techniques to tell us more, or else use what we
currently know to create computer simulations that might give us
additional insight. This is what Edelman has been doing.
====================

Or we could just develop a better grasp of differential mechanics
right now. If we move with neurophysiology we'll just wind up in fifty
years with better neurophysiology. Which isn't a bad thing if you're a
neurophysiologist. But it won't explain the mechanics or nature of
sentience or consciousness.

Which probably wouldn't be too far away from Edelman bipartate
classifications of primary and higher-order consciousness.
===============


Of course the issue as to whether homo


In which case, I think Edleman's classification scheme has already
trumped you.

Except that Edleman has no mechanics justify his classification scheme
in which case my mechanics and classification scheme trump his.
;-)


Regards - Lester

Like I said above, Edelman has been using knowledge about neuroanatomy
and neurophysiology to produce large-scale computer models and
simulations of perceptual and consciousness mechanisms. He wants to
understand possible mechanisms underlying such brain processes. The
classification schemes are simply a way to set the foundation for his
simulations.

 

[dan michaels]

Awareness is not all that interesting. Even the most primitive robot
could be aware you stepped on its toe, just because a pressure sensor
tripped.

Self consciousness is not all that interesting. It is just one of the
bazillion pet tricks you can do with a decent brain when you have
nothing better to do.

What is fascinating is that sense of experience that appears inside
myself when I am not under anaesthesia. This is something quite other
than computation and awareness.

It is important because morality is based on the notion it is ok to be
"cruel" to rocks because they are unconscious, and not ok to humans
because they are. Nothing would actually feel. There would be no
pain, just sham pain.

So I define consciousness as the ability to actually FEEL pain.

This definition suggests a much wider inclusion -- any thing that
reacts AS IF in pain until shown otherwise.

If you kick a dog, it feels pain. It's howling after the fact makes
this very obvious.

 

[Programmer Dude]

If you kick a dog, it feels pain. It's howling after the fact makes
this very obvious.

With some dogs, you also get a display of the emotion: anger!

 

[Noah Roberts]

dan michaels wrote:




With some dogs, you also get a display of the emotion: anger!

Actually, I don't believe that dogs have this emotion. They can become
frustrated and/or defensive but I have never seen angry. When you kick
a dog and he attacks, if he attacks and doesn't just run, he is biting
out of a defensive mindset not out of anger.

NR

 

[Programmer Dude]

Actually, I don't believe that dogs have this emotion. They can
become frustrated and/or defensive but I have never seen angry.
When you kick a dog and he attacks, if he attacks and doesn't just
run, he is biting out of a defensive mindset not out of anger.

Well, I did mean that a bit tongue in cheek. (Single line replies
from me usually are at least a little so.)

Perhaps anger isn't the best description, but frustration or
defensiveness seems futher off the mark to my meaning. Some
dogs are very aggressive and their response is very close (in
my mind) to anger.

There's a pit we (my dog and I) walk past that just loses it (we're
a good ten yards away with a fence between). There's no reason for
the dog's reaction--indeed I've spoken gently to it and tried to
make friends.

But it just seems pissed off. (-:

In another case, I know a dog that will instantly attack ANY other
dog it can (it's a pussycat around humans). It just recently got
out of its yard, roamed around and killed a tied up dog much, much
smaller than it (i.e. no threat whatsoever).

OTOH, I'm convinced *that* dog is literally insane.

 

[Noah Roberts]

Noah Roberts wrote:




Well, I did mean that a bit tongue in cheek. (Single line replies
from me usually are at least a little so.)

Perhaps anger isn't the best description, but frustration or
defensiveness seems futher off the mark to my meaning. Some
dogs are very aggressive and their response is very close (in
my mind) to anger.

There's a pit we (my dog and I) walk past that just loses it (we're
a good ten yards away with a fence between). There's no reason for
the dog's reaction--indeed I've spoken gently to it and tried to
make friends.

You are encroaching on what it thinks of as its territory. Some are
more protective than others over their "territory" and will become quite
agressive in its defence. These dogs can be dangerous and I wouldn't
try to make friends too much. IMHO the owner has a responsibility to
train the dog out of this type of behavior, but too many don't.

But it just seems pissed off. (-:

In another case, I know a dog that will instantly attack ANY other
dog it can (it's a pussycat around humans). It just recently got
out of its yard, roamed around and killed a tied up dog much, much
smaller than it (i.e. no threat whatsoever).

Could be a hunter dog and the smaller dog triggered a prey/hunter type
response. I have a dog that chaces down and kills rabbits for no
apparent reason. My neighbor's dog killed two of our cats. Could also
have been a territorial dispute, the smaller dog tried to assert its
territory and lost.

Also, just because the animal is small does not mean the dog does not
perceive it as a threat. Small dogs can, and often do, behave very
aggressivly.

Most dogs won't kill another dog except by accident. There are of
course exceptions.

OTOH, I'm convinced *that* dog is literally insane.

It is very possible.

NR

 

[pete]

Actually, I don't believe that dogs have this emotion.
They can become frustrated and/or defensive but I have never
seen angry.

I've seen anger.

Briggs was a German shepard from the neighborhood.
Zeus was the larger resident. They were fine together alone,
but if you petted Briggs, Zeus would always attack him.
When Zeus was chained to the doghouse and sleeping,
I saw Briggs pick up Zeus' water bowl from right next to his face,
and take it just outside of the chain line. He didn't drink
or spill any water. Then he stood there looking at Zeus until
Zeus woke up. Zeus didn't react. Eventually Briggs went home.

 

[Noah Roberts]

I've seen anger.

Briggs was a German shepard from the neighborhood.
Zeus was the larger resident. They were fine together alone,
but if you petted Briggs, Zeus would always attack him.

Obviously Zeus feels he should be the dominant player, and so by
consequence, the one being pet. This is pack psychology, not anger.
Dog "attacks" like this are also usually not full out attacks at any
rate, but a way for one dog to communicate to the other that there is
something going on here that could develop into a fight. The bites are
inhibited and there is rarely any damage done. Real dog fights do not
happen like this.

When Zeus was chained to the doghouse and sleeping,
I saw Briggs pick up Zeus' water bowl from right next to his face,
and take it just outside of the chain line. He didn't drink
or spill any water. Then he stood there looking at Zeus until
Zeus woke up. Zeus didn't react. Eventually Briggs went home.

Yeah, I have seen dogs **** with each other. My sisters dog new to grab
my dogs collar and drag him all over the place. Without being there to
witness it I would guess this was done in play. By stealing something
of Zeus's, Briggs is trying to illicit an expected responce in which
Zeus chases or 'attacks'; he didn't get this responce and so got bored
and left. I have seen cats mess with dogs in a similar fassion; they
act in such a way as to get the dog to chase them.

Anyway, I am not a dog expert but it is my opinion that dogs don't get
angry. There is usually, from what I have seen, some other reason why
they are acting in a fassion that might be percieved by people as anger.

NR

 

[Programmer Dude]

You are encroaching on what it thinks of as its territory.

It's more than that. It's that it's a *dog* encroaching. I have no
data, but I'd still bet it doesn't react quite the same to a passing
human, and I'd bet it ignores cars completely.

Some are more protective [...] and will become quite agressive...

Yep. And I'd classify the expression of that agressiveness as
something very close to anger. Something that occurred to me last
night thinking about this was that *my* outbursts of anger are
these days almost solely due to frustration. If a dog can feel
frustrated, why not angry?

I'm *convinced* dogs feel joy and happiness. I've seen it in every
dog I've known. And I know my dog has a sense of humor--I've seen
it in action. So, if joy, happiness and teasing are part of a dog's
emotional palette, why not anger?

These dogs can be dangerous and I wouldn't try to make friends
too much.

You mentioned elsethread you're not super familiar with dogs. I am.
Been around them all my life, and--frankly--I prefer their company
to that of *most* humans. If you want a list of reasons, just check
out:
http://www.sonnack.com/Sam/DogFile1a.html

Anyway, I know my limits and I know how far I can go with a dog,
and I have indeed made friends with some of the agressive hounds
on our normal walk paths. This dog just isn't in a normal path,
so we don't see it that often. It's already calming down a bit
the last few times we've passed (e.g. last time, it didn't bother
following us to the furthest corner of its yard--it always has in
the past).

IMHO the owner has a responsibility to train the dog out of this
type of behavior, but too many don't.

VERY much agreed! I tend to be fairly lax training most of my
dogs--I don't want a highly-trained teddy bear--but one thing that
is *ABSOLUTE* is anti-bite training, respect for humans and obeying
when I tell them to "relax". The other very important command is
"Get Beer!"

Could be a hunter dog and the smaller dog triggered a prey/hunter
type response.

No, it's insane. (-: I know this dog and its history. It's nuts.
You can see it in its eyes.

 

[karen]

Obviously Zeus feels he should be the dominant player, and so by
consequence, the one being pet. This is pack psychology, not anger.
Dog "attacks" like this are also usually not full out attacks at any
rate, but a way for one dog to communicate to the other that there is
something going on here that could develop into a fight. The bites are
inhibited and there is rarely any damage done. Real dog fights do not
happen like this.


Yeah, I have seen dogs **** with each other. My sisters dog new to grab
my dogs collar and drag him all over the place. Without being there to
witness it I would guess this was done in play. By stealing something
of Zeus's, Briggs is trying to illicit an expected responce in which
Zeus chases or 'attacks'; he didn't get this responce and so got bored
and left. I have seen cats mess with dogs in a similar fassion; they
act in such a way as to get the dog to chase them.

Anyway, I am not a dog expert but it is my opinion that dogs don't get
angry. There is usually, from what I have seen, some other reason why
they are acting in a fassion that might be percieved by people as anger.

NR

Not that this has anything to do with programming, but...
And cats can apparently feel jealousy. One cat ignores a warm lap, wants
nothing to do with you. So, you convince the other cat to come to your lap,
and it gets petted. Suddenly, the first cat is right there, sometimes
pushing the second cat away.

Back to the subject of intelligence, and maybe this has been covered
else-thread, but wouldn't a reasonable definition that fits humans only
be a species that not only makes and uses tools, but then uses those tools
daily? A species that makes tools based on other tools that have been made,
and uses tools to an extent that they become dependant upon those tools.

My cats do have intelligence*, they do express emotion, and I can even
understand some of their language. (Most cat people know the word "bird" in
cat, even if they can't say it.) But their intelligence isn't on a level
that makes tools (making them less intelligent than primates and crows), nor
is it on a level that relies on tools (making them less intelligent than me
using my news browser).

That still doesn't define intelligence, of course, but it does show a
distinction between us and the other animals. Whatever that distinction may
mean.

-karen

* distinct and varying levels of intelligence. We've had a smart cat that
knew about 20 words in English and understood the concept of "want". We've
also had a dumb cat that was noticebly less bright than the average cat.
They express desire, and not just for food. The show jealousy, annoyance
(or anger), fear, contentment. Some of my assertion is probably me
interpreting their actions, I admit.

 

[Programmer Dude]

And cats can apparently feel jealousy.

I've seen dogs do the same.

 

[Noah Roberts]

My cats do have intelligence*, they do express emotion, and I can even
understand some of their language. (Most cat people know the word "bird" in
cat, even if they can't say it.) But their intelligence isn't on a level
that makes tools (making them less intelligent than primates and crows), nor
is it on a level that relies on tools (making them less intelligent than me
using my news browser).

I myself wouldn't say cats or dogs have a spoken language per se, but
there is a type of language that uses body and tone of voice to express
rudimentary ideas like, "Stay away" or, "I'm scared." I have had, and
do have, several cats and I have never percieved any "words" in their
language.

I have noticed much intelegence in both dogs and cats. Either or can
learn to open doors or fences. Most dogs never really learn how to
untangle a chain, but mine has. Mine has also learned how to take it
off if I put it on in a certain way that allows it, which I stopped
doing. There are of course stupid dogs and cats as well, my dog's
brother got ran over by my mom because he ran under the car from behind!
I don't know if it made him even stupider or not, but he is of obvious
less intelegence than my own. In my observation, the stupid dogs bark more.

NR