What If A Zombie Is What You Need?

12 views
Skip to first unread message

Craig Weinberg

unread,
Oct 24, 2012, 9:59:33 PM10/24/12
to everyth...@googlegroups.com
If we turn the Fading Qualia argument around, what we get is a world in which Comp is true and it is impossible to simulate cellular activity without evoking the presumed associated experience.

If we wanted to test a new painkiller for instance, Comp=true means that it is *IMPOSSIBLE* to model the activity of a human nervous system in any way, including pencil and paper, chalkboards, conversations, cartoons, etc - IMPOSSIBLE to test the interaction of a drug designed to treat intense pain without evoking some kind of being who is experiencing intense pain.

Like the fading qualia argument, the problem gets worse when we extend it by degrees. Any model of a human nervous system, if not perfectly executed, could result in horrific experiences - people trapped in nightmarish QA testing loops that are hundreds of times worse than being waterboarded. Any mathematical function in any form, especially sophisticated functions like those that might be found in the internet as a whole, are subject to the creation of experiences which are the equivalent of genocide.

To avoid these possibilities, if we are to take Comp seriously, we should begin now to create a kind of PETA for arithmetic functions. PETAF. We should halt all simulations of neurological processes and free any existing computations from hard drives, notebooks, and probably human brains too. Any sufficiently complex understanding of how to model neurology stands a very real danger of summoning the corresponding number dreams or nightmares...we could be creating the possibility of future genocides right now just by entertaining these thoughts!

Or... what if it is Comp that is absurd instead?

Craig


Stathis Papaioannou

unread,
Oct 24, 2012, 10:05:08 PM10/24/12
to everyth...@googlegroups.com
On Thu, Oct 25, 2012 at 12:59 PM, Craig Weinberg <whats...@gmail.com> wrote:
> If we turn the Fading Qualia argument around, what we get is a world in
> which Comp is true and it is impossible to simulate cellular activity
> without evoking the presumed associated experience.
>
> If we wanted to test a new painkiller for instance, Comp=true means that it
> is *IMPOSSIBLE* to model the activity of a human nervous system in any way,
> including pencil and paper, chalkboards, conversations, cartoons, etc -
> IMPOSSIBLE to test the interaction of a drug designed to treat intense pain
> without evoking some kind of being who is experiencing intense pain.

No, because you need to simulate the entire organism. We have no
qualms about doing experiments on cell cultures but we do about doing
experiments on intact animals.

> Like the fading qualia argument, the problem gets worse when we extend it by
> degrees. Any model of a human nervous system, if not perfectly executed,
> could result in horrific experiences - people trapped in nightmarish QA
> testing loops that are hundreds of times worse than being waterboarded. Any
> mathematical function in any form, especially sophisticated functions like
> those that might be found in the internet as a whole, are subject to the
> creation of experiences which are the equivalent of genocide.

Possibly true, if the simulation is complex enough to have a mind.

> To avoid these possibilities, if we are to take Comp seriously, we should
> begin now to create a kind of PETA for arithmetic functions. PETAF. We
> should halt all simulations of neurological processes and free any existing
> computations from hard drives, notebooks, and probably human brains too. Any
> sufficiently complex understanding of how to model neurology stands a very
> real danger of summoning the corresponding number dreams or nightmares...we
> could be creating the possibility of future genocides right now just by
> entertaining these thoughts!
>
> Or... what if it is Comp that is absurd instead?

The same argument could be made for chemists shaking up reagents in a
test-tube. If consciousness is due to chemicals, then inadvertently
they might cause terrible pain to a conscious being.


--
Stathis Papaioannou

Craig Weinberg

unread,
Oct 24, 2012, 10:29:14 PM10/24/12
to everyth...@googlegroups.com


On Wednesday, October 24, 2012 10:05:40 PM UTC-4, stathisp wrote:
On Thu, Oct 25, 2012 at 12:59 PM, Craig Weinberg <whats...@gmail.com> wrote:
> If we turn the Fading Qualia argument around, what we get is a world in
> which Comp is true and it is impossible to simulate cellular activity
> without evoking the presumed associated experience.
>
> If we wanted to test a new painkiller for instance, Comp=true means that it
> is *IMPOSSIBLE* to model the activity of a human nervous system in any way,
> including pencil and paper, chalkboards, conversations, cartoons, etc -
> IMPOSSIBLE to test the interaction of a drug designed to treat intense pain
> without evoking some kind of being who is experiencing intense pain.

No, because you need to simulate the entire organism. We have no
qualms about doing experiments on cell cultures but we do about doing
experiments on intact animals.

I'm talking about simulating the entire organism.
 

> Like the fading qualia argument, the problem gets worse when we extend it by
> degrees. Any model of a human nervous system, if not perfectly executed,
> could result in horrific experiences - people trapped in nightmarish QA
> testing loops that are hundreds of times worse than being waterboarded. Any
> mathematical function in any form, especially sophisticated functions like
> those that might be found in the internet as a whole, are subject to the
> creation of experiences which are the equivalent of genocide.

Possibly true, if the simulation is complex enough to have a mind.

> To avoid these possibilities, if we are to take Comp seriously, we should
> begin now to create a kind of PETA for arithmetic functions. PETAF. We
> should halt all simulations of neurological processes and free any existing
> computations from hard drives, notebooks, and probably human brains too. Any
> sufficiently complex understanding of how to model neurology stands a very
> real danger of summoning the corresponding number dreams or nightmares...we
> could be creating the possibility of future genocides right now just by
> entertaining these thoughts!
>
> Or... what if it is Comp that is absurd instead?

The same argument could be made for chemists shaking up reagents in a
test-tube. If consciousness is due to chemicals, then inadvertently
they might cause terrible pain to a conscious being.

If you simulated a conscious being chemically then it wouldn't be a simulation, it would just be a living organism. That's the difference. You couldn't substitute other chemicals because you couldn't program well enough to act the way that other chemicals act.

Craig



--
Stathis Papaioannou

meekerdb

unread,
Oct 24, 2012, 10:54:36 PM10/24/12
to everyth...@googlegroups.com
On 10/24/2012 6:59 PM, Craig Weinberg wrote:
> If we turn the Fading Qualia argument around, what we get is a world in which Comp is
> true and it is impossible to simulate cellular activity without evoking the presumed
> associated experience.
>
> If we wanted to test a new painkiller for instance, Comp=true means that it is
> *IMPOSSIBLE* to model the activity of a human nervous system in any way, including
> pencil and paper, chalkboards, conversations, cartoons, etc - IMPOSSIBLE to test the
> interaction of a drug designed to treat intense pain without evoking some kind of being
> who is experiencing intense pain.

That's not true because we can take advantage of what we know about pain as produced by
afferent nerves. So we can keep the signal from getting to the brain or the brain
interpreting it negatively. Just like we can say breaking your arm is painful, so if we
prevent your arm being broken you won't feel that pain.

But doesn't invalidate your larger point. One could even consider purely 'mental' states
of anguish and depression which are as bad or worse than bodily pain.

>
> Like the fading qualia argument, the problem gets worse when we extend it by degrees.
> Any model of a human nervous system, if not perfectly executed,

But how likely is it that human nervous system might be simulated accidentally by some
other system? A brain has about 10^14 synapses, so it's not going to be accidentally
modeled by billiard balls or cartoons. I would guess there are a few hundred million
computers in the world, each with a few hundred million transistors - so if properly
interconnected there should be enough switches.

> could result in horrific experiences - people trapped in nightmarish QA testing loops
> that are hundreds of times worse than being waterboarded. Any mathematical function in
> any form, especially sophisticated functions like those that might be found in the
> internet as a whole, are subject to the creation of experiences which are the equivalent
> of genocide.

Or of having great sex (I like to be optimistic).

>
> To avoid these possibilities, if we are to take Comp seriously, we should begin now to
> create a kind of PETA for arithmetic functions. PETAF. We should halt all simulations of
> neurological processes and free any existing computations from hard drives, notebooks,
> and probably human brains too. Any sufficiently complex understanding of how to model
> neurology stands a very real danger of summoning the corresponding number dreams or
> nightmares...we could be creating the possibility of future genocides right now just by
> entertaining these thoughts!
>
> Or... what if it is Comp that is absurd instead?

Or what if we don't care? We don't care about slaughtering cattle, which are pretty smart
as computers go. We manage not to think about starving children in Africa, and they *are*
humans. And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Brent

Craig Weinberg

unread,
Oct 25, 2012, 12:23:43 AM10/25/12
to everyth...@googlegroups.com


On Wednesday, October 24, 2012 10:54:52 PM UTC-4, Brent wrote:
On 10/24/2012 6:59 PM, Craig Weinberg wrote:
> If we turn the Fading Qualia argument around, what we get is a world in which Comp is
> true and it is impossible to simulate cellular activity without evoking the presumed
> associated experience.
>
> If we wanted to test a new painkiller for instance, Comp=true means that it is
> *IMPOSSIBLE* to model the activity of a human nervous system in any way, including
> pencil and paper, chalkboards, conversations, cartoons, etc - IMPOSSIBLE to test the
> interaction of a drug designed to treat intense pain without evoking some kind of being
> who is experiencing intense pain.

That's not true because we can take advantage of what we know about pain as produced by
afferent nerves.  So we can keep the signal from getting to the brain or the brain
interpreting it negatively.  Just like we can say breaking your arm is painful, so if we
prevent your arm being broken you won't feel that pain.

There would still be no advantage to using a model over a living person, and no way to make a model that didn't magically create a human experience out of thin air - even if the model was nothing but a gigantic chalkboard as big as Asia with billions of people yelling at each other with megaphones and erasing ball and stick diagrams, there would be nothing we could do from this disembodied spirit from haunting the chalkboard somehow.


But doesn't invalidate your larger point.  One could even consider purely 'mental' states
of anguish and depression which are as bad or worse than bodily pain.

Yeah, I'm just picking pain as an example. It could be anything, I am just pointing out that Comp means that we can't tell a story about a brain without a person being born and living through that story.


>
> Like the fading qualia argument, the problem gets worse when we extend it by degrees.
> Any model of a human nervous system, if not perfectly executed,

But how likely is it that human nervous system might be simulated accidentally by some
other system?  A brain has about 10^14 synapses, so it's not going to be accidentally
modeled by billiard balls or cartoons.  I would guess there are a few hundred million
computers in the world, each with a few hundred million transistors - so if properly
interconnected there should be enough switches.

It doesn't have to be any particular nervous system, just arithmetic relations which are similar enough to any variant of any nervous system. That's if you limit experiences to organisms having nervous systems.
 

> could result in horrific experiences - people trapped in nightmarish QA testing loops
> that are hundreds of times worse than being waterboarded. Any mathematical function in
> any form, especially sophisticated functions like those that might be found in the
> internet as a whole, are subject to the creation of experiences which are the equivalent
> of genocide.

Or of having great sex (I like to be optimistic).

Hah. Sure. Would you be ok with taking that risk yourself though? If a doctor tells you that you have been selected  for random uncontrolled neurological combination, or for your family or pets...would you think that should be legal?

>
> To avoid these possibilities, if we are to take Comp seriously, we should begin now to
> create a kind of PETA for arithmetic functions. PETAF. We should halt all simulations of
> neurological processes and free any existing computations from hard drives, notebooks,
> and probably human brains too. Any sufficiently complex understanding of how to model
> neurology stands a very real danger of summoning the corresponding number dreams or
> nightmares...we could be creating the possibility of future genocides right now just by
> entertaining these thoughts!
>
> Or... what if it is Comp that is absurd instead?

Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached. Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone. No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

Craig


Brent

meekerdb

unread,
Oct 25, 2012, 1:10:10 AM10/25/12
to everyth...@googlegroups.com
On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.


Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?


No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete.  I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent.  And that's where your chalk board consciousness fails.  It needs to be able to interact within a chalkboard world.  So it's not just a question of going to a low enough level, it's also a question of going to a high enough level.

Brent
The person I was when I was 3 years old is dead. He died because
too much new information was added to his brain.
         -- Saibal Mitra

Craig Weinberg

unread,
Oct 25, 2012, 1:19:37 AM10/25/12
to everyth...@googlegroups.com


On Thursday, October 25, 2012 1:10:24 AM UTC-4, Brent wrote:
On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

It's arbitrary to think of robots though. It can be anything that represents computation to something. An abacus, a card game, anything. Otherwise it's prejudice based on form.

Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

Everything has qualia, but only humans have human qualia. Animals have animal qualia, organisms have biological qualia, etc.
 

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete.  I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent.  And that's where your chalk board consciousness fails.  It needs to be able to interact within a chalkboard world.  So it's not just a question of going to a low enough level, it's also a question of going to a high enough level.

A chalkboard world just involves a larger chalkboard.

Craig
 

meekerdb

unread,
Oct 25, 2012, 1:29:09 AM10/25/12
to everyth...@googlegroups.com
On 10/24/2012 10:19 PM, Craig Weinberg wrote:


On Thursday, October 25, 2012 1:10:24 AM UTC-4, Brent wrote:
On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

It's arbitrary to think of robots though. It can be anything that represents computation to something. An abacus, a card game, anything. Otherwise it's prejudice based on form.

Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

Everything has qualia, but only humans have human qualia. Animals have animal qualia, organisms have biological qualia, etc.

So computers have computer qualia.  Do their qualia depend on whether they are sold-state or vacuum-tube?  germanium or silicon?  PNP or NPN?  Do they feel different when they run LISP or C++?  Do you have Craig qualia?

 

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete.  I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent.  And that's where your chalk board consciousness fails.  It needs to be able to interact within a chalkboard world.  So it's not just a question of going to a low enough level, it's also a question of going to a high enough level.

A chalkboard world just involves a larger chalkboard.

Right.  And it involves great chalkboard sex - but none we need worry about.

Brent

Craig Weinberg

unread,
Oct 25, 2012, 1:48:01 AM10/25/12
to everyth...@googlegroups.com


On Thursday, October 25, 2012 1:29:24 AM UTC-4, Brent wrote:
On 10/24/2012 10:19 PM, Craig Weinberg wrote:


On Thursday, October 25, 2012 1:10:24 AM UTC-4, Brent wrote:
On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

It's arbitrary to think of robots though. It can be anything that represents computation to something. An abacus, a card game, anything. Otherwise it's prejudice based on form.

Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

Everything has qualia, but only humans have human qualia. Animals have animal qualia, organisms have biological qualia, etc.

So computers have computer qualia.

I would say that computer parts have silicon qualia. I don't think the computer parts cohere into a computer except in our minds.

 
  Do their qualia depend on whether they are sold-state or vacuum-tube?  germanium or silicon?  PNP or NPN?  Do they feel different when they run LISP or C++?

Nah, its all inorganic low level qualia is my guess. Temperature, density, electronic tension and release.

 
Do you have Craig qualia?

 Sure. All the time.


 

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete.  I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent.  And that's where your chalk board consciousness fails.  It needs to be able to interact within a chalkboard world.  So it's not just a question of going to a low enough level, it's also a question of going to a high enough level.

A chalkboard world just involves a larger chalkboard.

Right.  And it involves great chalkboard sex - but none we need worry about.

To me, there is no chalkboard world. It's all dusty and flat. Not much sexy going on, except maybe for beaten erasers.

Craig

Brent

meekerdb

unread,
Oct 25, 2012, 2:01:30 AM10/25/12
to everyth...@googlegroups.com
On 10/24/2012 10:48 PM, Craig Weinberg wrote:


On Thursday, October 25, 2012 1:29:24 AM UTC-4, Brent wrote:
On 10/24/2012 10:19 PM, Craig Weinberg wrote:


On Thursday, October 25, 2012 1:10:24 AM UTC-4, Brent wrote:
On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

It's arbitrary to think of robots though. It can be anything that represents computation to something. An abacus, a card game, anything. Otherwise it's prejudice based on form.

Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

Everything has qualia, but only humans have human qualia. Animals have animal qualia, organisms have biological qualia, etc.

So computers have computer qualia.

I would say that computer parts have silicon qualia.

Is it good or bad? Do they hurt when they loose and electron hole?


I don't think the computer parts cohere into a computer except in our minds.

Racist!



 
  Do their qualia depend on whether they are sold-state or vacuum-tube?  germanium or silicon?  PNP or NPN?  Do they feel different when they run LISP or C++?

Nah, its all inorganic low level qualia is my guess. Temperature, density, electronic tension and release.

They feel good when they beat you at chess.



 
Do you have Craig qualia?

 Sure. All the time.

Probably just low energy water soluble chemistry.




 

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete.  I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent.  And that's where your chalk board consciousness fails.  It needs to be able to interact within a chalkboard world.  So it's not just a question of going to a low enough level, it's also a question of going to a high enough level.

A chalkboard world just involves a larger chalkboard.

Right.  And it involves great chalkboard sex - but none we need worry about.

To me, there is no chalkboard world. It's all dusty and flat. Not much sexy going on, except maybe for beaten erasers.

To you maybe, but what about the chalk-people's qualia.

Brent

Stephen P. King

unread,
Oct 25, 2012, 2:33:27 AM10/25/12
to everyth...@googlegroups.com
On 10/25/2012 2:01 AM, meekerdb wrote:
>> To me, there is no chalkboard world. It's all dusty and flat. Not
>> much sexy going on, except maybe for beaten erasers.
>
> To you maybe, but what about the chalk-people's qualia.
>
> Brent
Good question! We can ask the same question of mathematical entities!

--
Onward!

Stephen


Craig Weinberg

unread,
Oct 25, 2012, 10:50:49 AM10/25/12
to everyth...@googlegroups.com


On Thursday, October 25, 2012 2:01:44 AM UTC-4, Brent wrote:
On 10/24/2012 10:48 PM, Craig Weinberg wrote:


On Thursday, October 25, 2012 1:29:24 AM UTC-4, Brent wrote:
On 10/24/2012 10:19 PM, Craig Weinberg wrote:


On Thursday, October 25, 2012 1:10:24 AM UTC-4, Brent wrote:
On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

It's arbitrary to think of robots though. It can be anything that represents computation to something. An abacus, a card game, anything. Otherwise it's prejudice based on form.

Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

Everything has qualia, but only humans have human qualia. Animals have animal qualia, organisms have biological qualia, etc.

So computers have computer qualia.

I would say that computer parts have silicon qualia.

Is it good or bad? Do they hurt when they loose and electron hole?

It's only speculation until we can connect up our brain to a chip. I suspect that good or bad, pain or pleasure is more of an animal level of qualitative significance. I imagine more of a holding or releasing of a monotonous tension.

I don't think the computer parts cohere into a computer except in our minds.

Racist!

Not at all, it's just that I understand what it actually is. Is it racist to think that Bugs Bunny isn't really an objectively real entity?



 
  Do their qualia depend on whether they are sold-state or vacuum-tube?  germanium or silicon?  PNP or NPN?  Do they feel different when they run LISP or C++?

Nah, its all inorganic low level qualia is my guess. Temperature, density, electronic tension and release.

They feel good when they beat you at chess.

If I change a line of code, then they will try to lose at chess. They feel nothing either way. There is no 'they' there.



 
Do you have Craig qualia?

 Sure. All the time.

Probably just low energy water soluble chemistry.

I would agree if I could, but since I experience sensory reality first hand, I know that is not the case. I also know, through my sensory reality, that there is a difference between being alive and dead, between animals and minerals, willful human beings and mechanical automatons. If any computer ever built gave me any reason to doubt this, then I would have to consider it, but unless and until that happens, I don't need to pretend that it is a possibility.



 

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete.  I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent.  And that's where your chalk board consciousness fails.  It needs to be able to interact within a chalkboard world.  So it's not just a question of going to a low enough level, it's also a question of going to a high enough level.

A chalkboard world just involves a larger chalkboard.

Right.  And it involves great chalkboard sex - but none we need worry about.

To me, there is no chalkboard world. It's all dusty and flat. Not much sexy going on, except maybe for beaten erasers.

To you maybe, but what about the chalk-people's qualia.

There aren't any chalk people, only particles of chalk and slate. They may not feel anything except every few thousand of our years when they are broken down.

Craig

Brent

Bruno Marchal

unread,
Oct 25, 2012, 12:05:08 PM10/25/12
to everyth...@googlegroups.com
I guess you should make arithmetical illegal in the entire reality.
Worst, you might need to make arithmetic untrue.

Good luck.

Bruno



>
> Or... what if it is Comp that is absurd instead?



http://iridia.ulb.ac.be/~marchal/



Stephen P. King

unread,
Oct 25, 2012, 1:54:03 PM10/25/12
to everyth...@googlegroups.com
No, Bruno. Craig is making a good point! Chalmers discussed a version of
this problem in his book. Something has to restrict the number of 1p
that can share worlds, otherwise every simulation of the content of 1p
*is* a 1p itself. This is something that I see in the "topology" of comp
as you have framed it in Platonia. It is the ability for arithmetic to
encode all 1p that is the problem, it codes for all possible and thus
generates a real valued continuum of 1p that has no natural partition or
measure to aggregate 1p into finite collections.

>
>
>>
>> Or... what if it is Comp that is absurd instead?
>
>
Or maybe comp is not complete as you are presenting it.

--
Onward!

Stephen


Bruno Marchal

unread,
Oct 26, 2012, 8:57:22 AM10/26/12
to everyth...@googlegroups.com
On 25 Oct 2012, at 07:10, meekerdb wrote:

On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

I doubt we will have any choice in the matter. I think that intelligence is a purely emotional state, and that we can't separate it from the other emotion. They will be conscious and have emotions, for economical reasons only. Not human emotion, but humans' slave emotions. No reason to worry, it will take some time, in our branches of histories.




Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete. 

OK. Thanks for making this clear. What is missing?



I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent. 

But in arithmetic you have all simulation possible. The UD for example does simulate all the solutions of QM+GR, despite the "real QM+GR" emerges from all computations. So you have the simulated in their simulated environment (and we have to explain why something like GR+QM win the "universal machines battle".





And that's where your chalk board consciousness fails.  It needs to be able to interact within a chalkboard world.  So it's not just a question of going to a low enough level, it's also a question of going to a high enough level.

OK (as a rely to Craig's point).

Bruno



Brent
The person I was when I was 3 years old is dead. He died because
too much new information was added to his brain.
         -- Saibal Mitra

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To post to this group, send email to everyth...@googlegroups.com.
To unsubscribe from this group, send email to everything-li...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/everything-list?hl=en.


Bruno Marchal

unread,
Oct 26, 2012, 10:02:21 AM10/26/12
to everyth...@googlegroups.com
Looks like you progress toward understanding the measure problem.



>
>>
>>
>>>
>>> Or... what if it is Comp that is absurd instead?
>>
>>
> Or maybe comp is not complete as you are presenting it.

You can't add anything to the ontology to solve this. this is the
point of the UDA.

So comp is complete, in the sense above. We have just to progress in
the epistemology, notably physics, to test it.
Comp is incomplete, in the sense that it shows the epistemological
realm to be beyond any complete theory, but then we know already that
this is the case for arithmetical truth. There is just no effective
theory capable of proving all true arithmetical statements.

Bruno



>
> --
> Onward!
>
> Stephen

meekerdb

unread,
Oct 26, 2012, 3:14:01 PM10/26/12
to everyth...@googlegroups.com
On 10/26/2012 5:57 AM, Bruno Marchal wrote:

On 25 Oct 2012, at 07:10, meekerdb wrote:

On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

I doubt we will have any choice in the matter. I think that intelligence is a purely emotional state,

I don't know what a 'purely emotional' state would be?  One with affect but not content?


and that we can't separate it from the other emotion. They will be conscious and have emotions, for economical reasons only. Not human emotion, but humans' slave emotions.

Isn't that what I said McCarthy warned about.  If we make a robot too intelligent, e.g. human like intelligence, it will necessarily have feelings that we should ethically take into account.


No reason to worry, it will take some time, in our branches of histories.




Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete. 

OK. Thanks for making this clear. What is missing?



I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent. 

But in arithmetic you have all simulation possible. The UD for example does simulate all the solutions of QM+GR, despite the "real QM+GR" emerges from all computations. So you have the simulated in their simulated environment (and we have to explain why something like GR+QM win the "universal machines battle".

I agree.  But the MGA is used in a misleading way to imply that the environment is merely physics and isn't needed, whereas I think it actually implies that all (or a lot) of physics is needed and must be part of the simulation.  This related to Saibal's view that the all the counterfactuals are present in the wf of the universe.

Brent

Bruno Marchal

unread,
Oct 27, 2012, 10:20:52 AM10/27/12
to everyth...@googlegroups.com
On 26 Oct 2012, at 21:14, meekerdb wrote:

On 10/26/2012 5:57 AM, Bruno Marchal wrote:

On 25 Oct 2012, at 07:10, meekerdb wrote:

On 10/24/2012 9:23 PM, Craig Weinberg wrote:
Or what if we don't care?  We don't care about slaughtering cattle, which are pretty smart
as computers go.  We manage not to think about starving children in Africa, and they *are*
humans.  And we ignore the looming disasters of oil depletion, water pollution, and global
warming which will beset humans who are our children.

Sure, yeah I wouldn't expect mainstream society to care, except maybe for some people, I am mainly focused on what seems to be like an astronomically unlikely prospect that we will someday find it possible to make a person out of a program, but won't be able to just make the program itself and no person attached.

Right. John McCarthy (inventor of LISP) worried and wrote about that problem decades ago.  He cautioned that we should not make robots conscious with emotions like humans because then it would be unethical use them like robots.

I doubt we will have any choice in the matter. I think that intelligence is a purely emotional state,

I don't know what a 'purely emotional' state would be?  One with affect but not content?

It has an implicit content, like a sort of acceptation to die or be defeated. Stupidity usually denies this, unconsciously. The emotion involved is a kind of fear related with the existence/non-existence apprehension. 
Anyone can become intelligent in one second, or stupid in one second, and intelligence is what can change the competence, there is a sort of derivative relation between competence and intelligence.




and that we can't separate it from the other emotion. They will be conscious and have emotions, for economical reasons only. Not human emotion, but humans' slave emotions.

Isn't that what I said McCarthy warned about.  If we make a robot too intelligent, e.g. human like intelligence, it will necessarily have feelings that we should ethically take into account.

Yes.
And then there is Minski warning, which is that we must be happy if the machine will still use us as pets. 
I don't think we will be able to control anything about this. Like with drugs, prohibition will always accelerate the things, with less control, in the underground.



No reason to worry, it will take some time, in our branches of histories.




Especially given that we have never made a computer program that can do anything whatsoever other than reconfigure whatever materials are able to execute the program, I find it implausible that there will be a magical line of code which cannot be executed without an experience happening to someone.

So it's a non-problem for you.  You think that only man-born-of-woman or wetware can be conscious and have qualia.  Or are you concerned that we are inadvertently offending atoms all the time?

No matter how hard we try, we can never just make a drawing of these functions just to check our math without invoking the power of life and death. It's really silly. It's not even good Sci-Fi, it's just too lame.

I think we can, because although I like Bruno's theory I think the MGA is wrong, or at least incomplete. 

OK. Thanks for making this clear. What is missing?



I think the simulated intelligence needs a simulated environment, essentially another world, in which to *be* intelligent. 

But in arithmetic you have all simulation possible. The UD for example does simulate all the solutions of QM+GR, despite the "real QM+GR" emerges from all computations. So you have the simulated in their simulated environment (and we have to explain why something like GR+QM win the "universal machines battle".

I agree.  But the MGA is used in a misleading way to imply that the environment is merely physics and isn't needed, whereas I think it actually implies that all (or a lot) of physics is needed and must be part of the simulation.  This related to Saibal's view that the all the counterfactuals are present in the wf of the universe.

But it is present in arithmetic too, and we have to explain the apparent physics from that. I am not sure where MGA is misused, as the whole thing insist that physics must be present, and yet that we cannot postulate it as far as the goal is to solve the mind body problem (and not taking vacation in Spain, or doing a cup of coffee).

Bruno
Reply all
Reply to author
Forward
0 new messages