The code for AGI will be simple

[John Clark]

This is an interview of the great computer programmer John Carmack, he thinks the time when computers can do everything, not just some things, as good or better than humans is much closer than most people believe, he thinks there is a 60% chance it will happen by 2030. Like me Carmack is much more interested in intelligence than consciousness and has no interest in the "philosophical zombie" argument. As far as the future history of the human race is concerned the following quotation is particularly relevant:

"It seems to me this is the highest leverage moment for a single individual potentially in the history of the world. [...]   I am not a mad man in saying that the code for artificial General intelligence is going to be tens of thousands of lines of code, not millions of lines of code. This is code that conceivably one individual could write, unliker writing a new web browser or operating system."

The code for AGI will be simple

John K Clark    See what's on my new list at  Extropolis

b30

[Telmo Menezes]

Am Do, 8. Sep 2022, um 14:09, schrieb John Clark:

This is an interview of the great computer programmer John Carmack, he thinks the time when computers can do everything, not just some things, as good or better than humans is much closer than most people believe, he thinks there is a 60% chance it will happen by 2030.

I agree. I believe that the breakthrough towards AGI will be neurosymbolic. The current deep learning models will be the tissue / cartilage and formal logic will be the skeleton. I suspect that evolutionary computation will play a role too, probably for hyperparameter optimization and tasks of that nature.

 Like me Carmack is much more interested in intelligence than consciousness and has no interest in the "philosophical zombie" argument.

It is possible to be highly interested in both. Why not?

As far as the future history of the human race is concerned the following quotation is particularly relevant:

"It seems to me this is the highest leverage moment for a single individual potentially in the history of the world. [...]   I am not a mad man in saying that the code for artificial General intelligence is going to be tens of thousands of lines of code, not millions of lines of code. This is code that conceivably one individual could write, unliker writing a new web browser or operating system."

The code for AGI will be simple

In a sense, I agree. But remember that, even with code, we are sitting on the shoulders of giants. A few lines of code in contemporary Python mobilize decades upon decades of the blood sweat and tears of the programmers that came before, who built all of this amazing infrastructure. How many lines in the Linux kernel?

Telmo

John K Clark    See what's on my new list at  Extropolis

b30

--

You received this message because you are subscribed to the Google Groups "Everything List" group.

To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/CAJPayv3ZEbXXVjs803%3Dutjc2pvkCgpZGA%2Bad_OWBhue-5kxDJQ%40mail.gmail.com.

[John Clark]

On Thu, Sep 8, 2022 at 8:19 AM Telmo Menezes <te...@telmomenezes.net> wrote:

 >> Like me Carmack is much more interested in intelligence than consciousness and has no interest in the "philosophical zombie" argument.

> It is possible to be highly interested in both. Why not?

Because one is a useful activity and the other is not. Even if you have an IQ of 200 and spend your entire life studying consciousness you will advance the field precisely as much as the entire human race has in the last thousand years. And that would be precisely zero. Isaac Newton must've had an IQ of about 200 and unfortunately he spent much more time studying theology than physics and mathematics put together, but despite that colossal effort he advanced the field of theology not at all, and nobody else has managed to do any better. The same is true with consciousness.

>> "It seems to me this is the highest leverage moment for a single individual potentially in the history of the world. [...]   I am not a mad man in saying that the code for artificial General intelligence is going to be tens of thousands of lines of code, not millions of lines of code. This is code that conceivably one individual could write, unliker writing a new web browser or operating system."

> In a sense, I agree. But remember that, even with code, we are sitting on the shoulders of giants. A few lines of code in contemporary Python mobilize decades upon decades of the blood sweat and tears of the programmers that came before, who built all of this amazing infrastructure. How many lines in the Linux kernel?

That's why I disagree with those who say Moore's law only applies to hardware and not to software.  Imagine if there were no modern software tools and you had to program everything in machine language using nothing but 0 and 1. Fortunately we don't have to do that because machines have been able to help us write computer programs for many decades.

John K Clark    See what's on my new list at  Extropolis

stc

[Brent Meeker]

On 9/8/2022 8:00 AM, John Clark wrote:

On Thu, Sep 8, 2022 at 8:19 AM Telmo Menezes <te...@telmomenezes.net> wrote:

 >> Like me Carmack is much more interested in intelligence than consciousness and has no interest in the "philosophical zombie" argument.

> It is possible to be highly interested in both. Why not?

Because one is a useful activity and the other is not.

If one were unconscious to whom would it be useful?

Brent

Even if you have an IQ of 200 and spend your entire life studying consciousness you will advance the field precisely as much as the entire human race has in the last thousand years. And that would be precisely zero. Isaac Newton must've had an IQ of about 200 and unfortunately he spent much more time studying theology than physics and mathematics put together, but despite that colossal effort he advanced the field of theology not at all, and nobody else has managed to do any better. The same is true with consciousness.

>> "It seems to me this is the highest leverage moment for a single individual potentially in the history of the world. [...]   I am not a mad man in saying that the code for artificial General intelligence is going to be tens of thousands of lines of code, not millions of lines of code. This is code that conceivably one individual could write, unliker writing a new web browser or operating system."

> In a sense, I agree. But remember that, even with code, we are sitting on the shoulders of giants. A few lines of code in contemporary Python mobilize decades upon decades of the blood sweat and tears of the programmers that came before, who built all of this amazing infrastructure. How many lines in the Linux kernel?

That's why I disagree with those who say Moore's law only applies to hardware and not to software.  Imagine if there were no modern software tools and you had to program everything in machine language using nothing but 0 and 1. Fortunately we don't have to do that because machines have been able to help us write computer programs for many decades.

John K Clark    See what's on my new list at  Extropolis

stc

--
You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/CAJPayv3Oo3GJW%2BXU%2BgO0_h0cKXf0EwpEooOiEZwk%3DnqR_EmNdg%40mail.gmail.com.

[John Clark]

On Thu, Sep 8, 2022 at 6:31 PM Brent Meeker <meeke...@gmail.com> wrote:

 >>>>  Like me Carmack is much more interested in intelligence than consciousness and has no interest in the "philosophical zombie" argument.

>>>  It is possible to be highly interested in both. Why not?

>> If one were unconscious to whom would it be useful?

 

> Because one is a useful activity and the other is not. 

I meant that research into the nature of intelligence is useful, that is to say it leads somewhere, but consciousness research leads nowhere and accomplishes nothing.  

 John K Clark    See what's on my new list at  Extropolis

kkq

 

[Henrik Ohrstrom]

Could research into unconsciousness be somewhat easier?

When are a brain unconscious? 

During anesthesia you might say. 

Alas, as everything else, it is not that simple.

If we look at what's actually required for a satisfied patient, a truly unconscious brain ai'nt that.

Especially as long as both lawyers and patients agree that impaired memory are enough for satisfaction.

Absence of objections are not evidence of good enough.

A well and sufficiently (not over) medicated patient have a quite active brain during anesthesia. Conscious? Probably not. Memory retention? Hopefully not . Pain reactions to surgery? Actually possible to measure, se www.med-storm.com most patients with classic standard anesthesia are severly as in cognitive impairment dose level overmedicated during surgery. No Pain. 

So a patient who officially are the definition of unconscious, aint and should not be.

So what is unconscious then? I would wager that it is difficult to prove anything about a regular stone, so lets call that one unconscious.

/Henrik 

 

--

You received this message because you are subscribed to the Google Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/CAJPayv13bweRRmHnBk4BHB7ASwCgwTSWmgucqeCNevQrahB2EA%40mail.gmail.com.

[Telmo Menezes]

Am Do, 8. Sep 2022, um 17:00, schrieb John Clark:

On Thu, Sep 8, 2022 at 8:19 AM Telmo Menezes <te...@telmomenezes.net> wrote:

 >> Like me Carmack is much more interested in intelligence than consciousness and has no interest in the "philosophical zombie" argument.

> It is possible to be highly interested in both. Why not?

Because one is a useful activity and the other is not.

If there is one thing one cannot doubt, John, is that you are thoroughly American :)

Even if you have an IQ of 200 and spend your entire life studying consciousness you will advance the field precisely as much as the entire human race has in the last thousand years. And that would be precisely zero. Isaac Newton must've had an IQ of about 200 and unfortunately he spent much more time studying theology

Unfortunately for you maybe, but perhaps it gave him joy and I bet that was the main thing that mattered to Isaac Newton. Good for him, I would say. At some point we will all be dead, and nothing will matter or be useful to us by then.

Telmo

than physics and mathematics put together, but despite that colossal effort he advanced the field of theology not at all, and nobody else has managed to do any better. The same is true with consciousness.

>> "It seems to me this is the highest leverage moment for a single individual potentially in the history of the world. [...]   I am not a mad man in saying that the code for artificial General intelligence is going to be tens of thousands of lines of code, not millions of lines of code. This is code that conceivably one individual could write, unliker writing a new web browser or operating system."

> In a sense, I agree. But remember that, even with code, we are sitting on the shoulders of giants. A few lines of code in contemporary Python mobilize decades upon decades of the blood sweat and tears of the programmers that came before, who built all of this amazing infrastructure. How many lines in the Linux kernel?

That's why I disagree with those who say Moore's law only applies to hardware and not to software.  Imagine if there were no modern software tools and you had to program everything in machine language using nothing but 0 and 1. Fortunately we don't have to do that because machines have been able to help us write computer programs for many decades.

John K Clark    See what's on my new list at  Extropolis

stc

--

You received this message because you are subscribed to the Google Groups "Everything List" group.

To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/CAJPayv3Oo3GJW%2BXU%2BgO0_h0cKXf0EwpEooOiEZwk%3DnqR_EmNdg%40mail.gmail.com.

[John Clark]

On Fri, Sep 9, 2022 at 5:20 AM Telmo Menezes <te...@telmomenezes.net> wrote:

>> Even if you have an IQ of 200 and spend your entire life studying consciousness you will advance the field precisely as much as the entire human race has in the last thousand years. And that would be precisely zero. Isaac Newton must've had an IQ of about 200 and unfortunately he spent much more time studying theology than physics and mathematics put together, but despite that colossal effort he advanced the field of theology not at all,

> Unfortunately for you maybe, but perhaps it gave him joy and I bet that was the main thing that mattered to Isaac Newton. Good for him, I would say. At some point we will all be dead, and nothing will matter or be useful to us by then.

I can only speak for myself but all else being equal if given the choice between a task that has a chance of accomplishing something and a task that can only lead to a dead end then I would choose the one that might actually achieve something, even if that achievement is just an abstract understanding of something with no practical value.

John K Clark    See what's on my new list at  Extropolis

npv

[smitra]

This is something that fits in with what I wrote here some time ago
about insect-level AI taking over from us.

A system with AGI doesn't have to be all that intelligent for it to be
extremely useful. Today we cannot build a remotely controlled spider
that could survive in Nature. That little intelligence a spider has is
the GI it needs to take on the challenges of surviving. If we have
something similar, say spider level AGI then that's good enough to fully
automatize our entire economy with. The reason why you can't replace all
factory worker with machines is due to a lack of even a minimal amount
of AGI.

So, I think insect-level AGI will cause a rapid transition to a machine
civilization. This will lead to a new biology of machines with insect
level intelligence ending up wiping out all life on Earth due to
pollution, similar to the great oxygenation event:

https://en.wikipedia.org/wiki/Great_Oxidation_Event

And as I pointed out earlier, I think this is a universal phenomena that
all intelligent life is subject to. The whole point of being intelligent
is to let as much of the work be done for you by entities that are
dumber than you. But in that process that leads to faster and faster
economic growth, its inevitable that at some point you are going to
crate autonomous systems that will grow exponentially. The point where
the transition to artificial life starts is going to be close to the
minimum intelligence level needed for exponential growth.

If you make it hotter and hotter in some closed space, a fire will break
out, this is going to happen close to the minimum required temperature
for ignition, not at some extremely high value for the temperature.
Nature shows us that the minimum amount of intelligence required for
efficient self-maintenance and reproduction that yields exponential
growth is very low.

Saibal

On 08-09-2022 14:09, John Clark wrote:
> This is an interview of the great computer programmer John Carmack, he
> thinks the time when computers can do everything, not just some
> things, as good or better than humans is much closer than most people
> believe, he thinks there is a 60% chance it will happen by 2030. Like
> me Carmack is much more interested in intelligence than consciousness
> and has no interest in the "philosophical zombie" argument. As far as
> the future history of the human race is concerned the following
> quotation is particularly relevant:
>

> "___It seems to me this is the highest leverage moment for a single
> individual potentially_ _in the history of the world._ [...] _I am

> not a mad man in saying that the code for artificial General
> intelligence is going to be tens of thousands of lines of code, not
> millions of lines of code. This is code that conceivably one
> individual could write, unliker writing a new web browser or operating

> system._"
>
> The code for AGI will be simple [1]
>
> John K Clark See what's on my new list at Extropolis [2]
>
> b30

>
> --
> You received this message because you are subscribed to the Google
> Groups "Everything List" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to everything-li...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/everything-list/CAJPayv3ZEbXXVjs803%3Dutjc2pvkCgpZGA%2Bad_OWBhue-5kxDJQ%40mail.gmail.com
> [3].
>
>
> Links:
> ------
> [1] https://www.youtube.com/watch?v=xLi83prR5fg
> [2] https://groups.google.com/g/extropolis
> [3]
> https://groups.google.com/d/msgid/everything-list/CAJPayv3ZEbXXVjs803%3Dutjc2pvkCgpZGA%2Bad_OWBhue-5kxDJQ%40mail.gmail.com?utm_medium=email&utm_source=footer

[Jesse Mazer]

On Fri, Sep 9, 2022 at 8:26 AM smitra <smi...@zonnet.nl> wrote:

So, I think insect-level AGI will cause a rapid transition to a machine
civilization. This will lead to a new biology of machines with insect
level intelligence ending up wiping out all life on Earth due to
pollution, similar to the great oxygenation event:

Are you assuming insect-level AGI would also be small like insects and could self-replicate just as rapidly using commonly-found materials as "nutrients"? If we had insect-level AGI but they were larger and easier to spot, and also took much longer than an insect to self-replicate (and perhaps required external infrastructure or uncommon materials to do so), it seems hard to imagine a scenario in which humanity wouldn't be able to prevent them from going into runaway self-replication mode.

I think the possibility of relatively "dumb" self-replicating machines, even if large and relatively slow like Eric Drexler's concept of a "clanking replicator" (see http://wfmh.org.pl/enginesofcreation/EOC_Chapter_4.html ), could disrupt society for a different reason--they could spell the end of capitalism, or at least radically change its nature. If there were commercially available machines that could replicate themselves, those who owned them could make copies for just the cost of raw materials and energy, and if they were competing to sell them, competition would tend to drive the cost down to materials/energy cost or barely above it, basically destroying profits for any good that isn't forced into artificial scarcity by intellectual property laws. This would likewise go for any other goods the machines are capable of replicating. If self-replicating machines could also extract resources (fully automated mining facilities, say), then profit would still be possible if raw materials returned > raw materials invested (akin to 'energy return on energy invested' in energy economics), but if companies were making profits by just setting up mining machines and then sitting back and doing nothing, this would probably cause political instability, both in democracies and autocratic systems, where either the people or the politicians would likely prefer to be the ones reliably getting back more than their initial investment with no work needed. Perhaps instead of totally ending capitalism, we might end up with a hybrid system where some sort of intellectual property laws would still be in place so companies and individuals could still profit from those, but actual production machinery would mostly be publicly owned, and people (along with retail companies) could order up any good from a database of designs, receiving something like a basic income in raw materials and energy (funded by mining and energy generation facilities which could also be publicly owned).

Arthur C. Clarke, in his 1962 nonfiction book Profiles of the Future, commented about how a self-replicating machine which could also replicate other goods, which he just called a "Replicator", would disrupt our current economic system:

"The advent of the Replicator would mean the end of all factories, and perhaps all transportation of raw materials and all farming. The entire structure of industry and commerce, as it is now organized, would cease to exist. Every family would produce all that it needed on the spot — as, indeed, it has had to do throughout most of human history. The present machine era of mass-production would then be seen as a brief interregnum between two far longer periods of self-sufficiency, and the only valuable item of exchange would be matrices, or recordings, which had to be inserted into the Replicator to control its creations.

"No one who has read thus far will, I hope, argue that the Replicator would itself be so expensive that nobody could possibly afford it. The prototype, it is true, is hardly likely to cost less than £1,000,000,000,000 spread over a few centuries of time. The second model would cost nothing, because the Replicator's first job would be to produce other Replicators. It is perhaps relevant to point out that in 1951 the great mathematician, John von Neumann, established the important principle that a machine could always be designed to build any describable machine -- including itself. The human race has squalling proof of this several hundred thousand times a day. 

"A society based on the Replicator would be so completely different from ours that the present debate between Capitalism and Communism would become quite meaningless. All material possessions would be literally cheap as dirt. Soiled handkerchiefs, diamond tiaras, Mona Lisas totally indistinguishable from the original, once-worn mink stoles, half-consumed bottles of the most superb champagnes – all would go back into the hopper when they were no longer required. Even the furniture in the house of the future might cease to exist when it was not actually in use.”

Probably this book was a major influence on Gene Roddenberry's vision of a post-scarcity future in Star Trek, see his comments quoted at https://arthurcclarke.org/site/how-arthur-c-clarke-helped-save-star-trek/ where he specifically references Profiles of the Future. For a more cyberpunk depiction of how fully automated self-replicating machinery could lead to a transition to a new kind of economic system, I recommend Cory Doctorow's recent sci fi novel "Walkaway".

Jesse

 

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/d4b54074fe283e5c198ff6a6d709b143%40zonnet.nl.

[Telmo Menezes]

If you choose, try to lose

for the loss of remain come and start

Start the game

I chi-chi, chi-chi I

chi-chi-chi, ka-ta-ko, choose to choose

choose to choose, choose to go

John K Clark    See what's on my new list at  Extropolis

npv

--

You received this message because you are subscribed to the Google Groups "Everything List" group.

To unsubscribe from this group and stop receiving emails from it, send an email to everything-li...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/CAJPayv1KXcLo84RTCVSK50G1wUbWs0Mh2uC1WqgLtTFkYVYt%2BA%40mail.gmail.com.

[spudb...@aol.com]

This is only opinion here, but I agree with Telmo because we as a species require and roi, a return on investment, for a few good reasons. 

1) We all like money or free stuff, or both.

2) In case robots make everyone unemployed.

3) We need longer term goals to ensure human survival. So we require better longer term projects. 

The vagueness of Alice & BoB will eventuate in something tasty someday, But not today. 

https://groups.google.com/d/msgid/everything-list/151dd21a-ce7e-4fd2-9cc8-212b30c21c39%40www.fastmail.com

.

[John Clark]

On Fri, Sep 9, 2022 at 8:26 AM smitra <smi...@zonnet.nl> wrote:

> A system with AGI doesn't have to be all that intelligent 

AGI is supposed to stand for Artificial General Intelligence, if something can have AGI without being intelligent then I don't know what "AGI" means.

>  for it to be extremely useful. 

Well yeah, yeast cells are very stupid but they're extremely useful if you wanna bake bread or make beer.  

> Today we cannot build a remotely controlled spider that could survive in Nature. 

I think you mean autonomous spiders, we've been making remotely controlled robots for years that survive just fine in the natural world. 

> If we have something similar, say spider level AGI then that's good enough to fully automatize our entire economy 

You'll need to do a lot better than that if you wanna automatize our entire economy. You need something smarter than a spider to write good software or design a microchip or diagnose an illness or perform surgery or even understand enough English to take an order at Burger King. I have no doubt it will just be a matter of time before computers can do all of those things better than a human, and with nanotechnology such a computer could be much smaller than a spider.

> This will lead to a new biology of machines with insect level intelligence ending up wiping out all life on Earth due to pollution, similar to the great oxygenation event:

I agree the earth is approaching something as monumental as the great oxygenation event and biological humans don't have much of a future, but our replacement will not be something with just insect level intelligence.

John K Clark    See what's on my new list at  Extropolis

jil

[Lawrence Crowell]

The code will be that length in some high level coding language.

There seems to be little asking the question, "What is all of this for?" An AI or AGI, whether conscious or not, that reduces humans to indolence does not seem like a good idea. If the AGI is mindless, say a mindless golem of sorts, then the whole thing seems utterly purposeless. If the system has some intentionality, or say a type of sentience that makes it capable of self-reference and actualization, then this could be very disturbing.

A brother of mine retired early from working as a programmer. He wrote the codes involved with barcoding, such as what we see at the store etc, and other things. He cannot help talk about computers without f-bombs. He is convinced all of this is heading into a disaster.

LC

[Lawrence Crowell]

On Friday, September 9, 2022 at 2:10:21 PM UTC-5 jessem wrote:

On Fri, Sep 9, 2022 at 8:26 AM smitra <smi...@zonnet.nl> wrote:

So, I think insect-level AGI will cause a rapid transition to a machine
civilization. This will lead to a new biology of machines with insect
level intelligence ending up wiping out all life on Earth due to
pollution, similar to the great oxygenation event:

Are you assuming insect-level AGI would also be small like insects and could self-replicate just as rapidly using commonly-found materials as "nutrients"? If we had insect-level AGI but they were larger and easier to spot, and also took much longer than an insect to self-replicate (and perhaps required external infrastructure or uncommon materials to do so), it seems hard to imagine a scenario in which humanity wouldn't be able to prevent them from going into runaway self-replication mode.

I think the possibility of relatively "dumb" self-replicating machines, even if large and relatively slow like Eric Drexler's concept of a "clanking replicator" (see http://wfmh.org.pl/enginesofcreation/EOC_Chapter_4.html ), could disrupt society for a different reason--they could spell the end of capitalism, or at least radically change its nature. If there were commercially available machines that could replicate themselves, those who owned them could make copies for just the cost of raw materials and energy, and if they were competing to sell them, competition would tend to drive the cost down to materials/energy cost or barely above it, basically destroying profits for any good that isn't forced into artificial scarcity by intellectual property laws. This would likewise go for any other goods the machines are capable of replicating. If self-replicating machines could also extract resources (fully automated mining facilities, say), then profit would still be possible if raw materials returned > raw materials invested (akin to 'energy return on energy invested' in energy economics), but if companies were making profits by just setting up mining machines and then sitting back and doing nothing, this would probably cause political instability, both in democracies and autocratic systems, where either the people or the politicians would likely prefer to be the ones reliably getting back more than their initial investment with no work needed. Perhaps instead of totally ending capitalism, we might end up with a hybrid system where some sort of intellectual property laws would still be in place so companies and individuals could still profit from those, but actual production machinery would mostly be publicly owned, and people (along with retail companies) could order up any good from a database of designs, receiving something like a basic income in raw materials and energy (funded by mining and energy generation facilities which could also be publicly owned).

Arthur C. Clarke, in his 1962 nonfiction book Profiles of the Future, commented about how a self-replicating machine which could also replicate other goods, which he just called a "Replicator", would disrupt our current economic system:

"The advent of the Replicator would mean the end of all factories, and perhaps all transportation of raw materials and all farming. The entire structure of industry and commerce, as it is now organized, would cease to exist. Every family would produce all that it needed on the spot — as, indeed, it has had to do throughout most of human history. The present machine era of mass-production would then be seen as a brief interregnum between two far longer periods of self-sufficiency, and the only valuable item of exchange would be matrices, or recordings, which had to be inserted into the Replicator to control its creations.

"No one who has read thus far will, I hope, argue that the Replicator would itself be so expensive that nobody could possibly afford it. The prototype, it is true, is hardly likely to cost less than £1,000,000,000,000 spread over a few centuries of time. The second model would cost nothing, because the Replicator's first job would be to produce other Replicators. It is perhaps relevant to point out that in 1951 the great mathematician, John von Neumann, established the important principle that a machine could always be designed to build any describable machine -- including itself. The human race has squalling proof of this several hundred thousand times a day. 

"A society based on the Replicator would be so completely different from ours that the present debate between Capitalism and Communism would become quite meaningless. All material possessions would be literally cheap as dirt. Soiled handkerchiefs, diamond tiaras, Mona Lisas totally indistinguishable from the original, once-worn mink stoles, half-consumed bottles of the most superb champagnes – all would go back into the hopper when they were no longer required. Even the furniture in the house of the future might cease to exist when it was not actually in use.”

Probably this book was a major influence on Gene Roddenberry's vision of a post-scarcity future in Star Trek, see his comments quoted at https://arthurcclarke.org/site/how-arthur-c-clarke-helped-save-star-trek/ where he specifically references Profiles of the Future. For a more cyberpunk depiction of how fully automated self-replicating machinery could lead to a transition to a new kind of economic system, I recommend Cory Doctorow's recent sci fi novel "Walkaway".

Jesse

All of that would require an enormous amount of energy. That is one thing that would put a limit on this.

LC