Self-Replicating Robots and Galactic Domination

[John Clark]

Self-Replicating Robots and Galactic Domination

John K Clark

[smitra]

On 19-05-2022 12:54, John Clark wrote:
> Self-Replicating Robots and Galactic Domination [1]
>

We're now producing very smart AI systems that can replace humans, but
those AI systems have still insect-level intelligence. So, while a
technological civilization like ours will get transformed into a machine
civilization due to economic growth, the self-replicating machines that
will arise will be rather dumb machines. They are then likely to destroy
all biological life forms, analogous to what happened when oxygen
producing microbes arose about 2.4 billion years ago:

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

The self-replicating machines will produce whatever chemical waste
products they can live with and they will be too dumb to realize that
they are killing everything else.

Saibal

> John K Clark
>
> --
> 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/CAJPayv0cq-zu6wL4c0i60E%2B-xYz1qs6_%2B7rrRcXmiAXt26wgvA%40mail.gmail.com
> [2].
>
>
> Links:
> ------
> [1] https://www.youtube.com/watch?v=4H55wybU3rI
> [2]
> https://groups.google.com/d/msgid/everything-list/CAJPayv0cq-zu6wL4c0i60E%2B-xYz1qs6_%2B7rrRcXmiAXt26wgvA%40mail.gmail.com?utm_medium=email&utm_source=footer

[John Clark]

On Thu, May 19, 2022 at 11:41 AM smitra <smi...@zonnet.nl> wrote:

>  AI systems have still insect-level intelligence.

10 years ago that might've been true, maybe even five, but not today.  

> The self-replicating machines will produce whatever chemical wasteproducts they can live with and they will be too dumb to realize that

they are killing everything else.

If an AI decides to kill all biological life it won't be because it's stupid, and even if they do decide to do that it wouldn't explain why we don't see an engineered universe when we look at the night sky. Somebody has to be the first and I think it must be us. 

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

fal

 

 

[Tomasz Rola]

On Thu, May 19, 2022 at 02:22:04PM -0400, John Clark wrote:
> On Thu, May 19, 2022 at 11:41 AM smitra <smi...@zonnet.nl> wrote:
>

> *> AI systems have still insect-level intelligence.*

>
>
> 10 years ago that might've been true, maybe even five, but not
> today.

I think I can easily wait five more years and see if it all boils down
to the mix of hype, half hype and enthusiastic lies...

[...]

> If an AI decides to kill all biological life it won't be because it's stupid,

I believe AI will not kill us, but dumb idiot-rats in their race for
another big bucks, releasing untested and hard to debug black boxes,
they have the potential. Of course, the untested black box may contain
certain form of AI.

> and even if they do decide to do that it wouldn't explain why we don't see
> an engineered universe when we look at the night sky. Somebody has to be
> the first and I think it must be us.

I really wonder, what makes you think this Universe is not engineered?
How would one test it for signs of "engineeredness"?

--
Regards,
Tomasz Rola

--
** A C programmer asked whether computer had Buddha's nature. **
** As the answer, master did "rm -rif" on the programmer's home **
** directory. And then the C programmer became enlightened... **
** **
** Tomasz Rola mailto:tomas...@bigfoot.com **

[John Clark]

On Fri, May 20, 2022 at 2:48 AM Tomasz Rola <rto...@ceti.pl> wrote:

> I think I can easily wait five more years and see if it all boils down to the mix of hype, half hype and enthusiastic lies...

In five years people will be saying what people have always been saying about AI, yes the progress made five years ago was real but there will be no further progress in the next five years, in fact there will be no further progress in AI EVER; and then they'll get all misty eyed and start talking about some secret sauce that only biological life in general and humans in particular have.  

> I really wonder, what makes you think this Universe is not engineered? How would one test it for signs of "engineeredness"?

If the universe was engineered then when we look at the Andromeda galaxy we shouldn't see light and other high-quality photons radiating uselessly into infinite space, we should only see low energy infrared and radio waves because even civilizations that prefer to live in a virtual world would need power to run their computers. And if even one individual in one civilization in the Andromeda galaxy decided to make a Von Neumann probe that's exactly how their galaxy would look to us, except that we wouldn't be here to look at it because one of their probes would've arrived at our galaxy by now.  All you need is one probe, and once Drexler style Nanotechnology has been developed, which I remind you would require no scientific breakthroughs just better engineering, everything could be divided into jist 2 classes, things that are easy and cheap to make, and things that are impossible to make because they violate the fundamental laws of physics. Nothing would be possible but expensive and hard to make.

That's why I think we must be the only intelligent species in the observable universe, it is after all a finite space so somebody has to be first. Some may argue for other explanations for the Fermi paradox, such as civilizations destroying themselves in a nuclear war, but I don't think that's a viable answer. Humans developed nuclear weapons about five years before they discovered the transistor but that may be unusual, it only happened because the USA was under enormous pressure and it was the only country in the world that had the industrial capacity to develop nuclear bombs in a remarkably short time. And if there are thousands or millions of civilizations in our galaxy as some claim then there must be some that developed on planets similar to Earth except that it had very little of the element Uranium in the planet's crust; that lack would not hinder the development of life since Uranium plays no part in biology, but it would mean that nuclear weapons would be impossible to make until their technology (not their science) had advanced enough to make asteroid mining practical, and by that point they'd almost certainly have the ability to make Von Neumann machines. Hell, I think humans will almost certainly have the capability to make Von Neumann probes in less than 100 years, probably less than 50, maybe much less.

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

edn

[smi...@zeelandnet.nl]

John Clark schreef op 19 mei '22:

> On Thu, May 19, 2022 at 11:41 AM smitra <smi...@zonnet.nl> wrote:
>

> *> AI systems have still insect-level intelligence.*

>
> 10 years ago that might've been true, maybe even five, but not today.

There has been a lot of progress in recent years. The problem with AI is
that we cannot build the right sort of hardware, making the systems we
have rather inefficient. We may now have systems that rival or exceed
insect brain power, but only using very large and computationally
inefficient systems.

>
>> * > The self-replicating machines will produce whatever chemical

>> wasteproducts they can live with and they will be too dumb to realize
>> that

>> they are killing everything else.*

>
> If an AI decides to kill all biological life it won't be because it's
> stupid,
> and even if they do decide to do that it wouldn't explain why we don't
> see
> an engineered universe when we look at the night sky. Somebody has to
> be
> the first and I think it must be us.
>

Most living organisms on Earth today are not very intelligent at all.
So, AI systems under development may soon be good enough to build
self-replicating machinery for fully automated factories with. The AI
systems running these factories don't need to be much smarter than
insects. Such factories will outcompete conventional factories, so
within a very short time period, most of our economy will run on such AI
systems.

The fact that many people have the ambition to build AI systems with
human-level intelligence or even more, is also true, but that's not
going to stop the use of insect-level AI systems in fully automated
factories. What may then happen is that we gradually lose control over
the self-replicating systems running the economy. Waste products
produced by the system accumulate, so it becomes like a cancer killing
the entire biosphere.

If this sounds far-fetched, note that the current climate crisis is due
to our economy producing CO2 and our resistance to reduce CO2 emissions.
We find it difficult to reduce CO2 emissions because we need to take
measures that will reduce out wealth to a limited degree (on the short
term). If we have difficulties doing this for such a simple system
(fossil fuels taken from the ground, burned in power plants to generate
power) that can easily be modified without ruining the economy, think
about how difficult it would be for the next generations to intervene in
a system of self-replicating factories that starts to grow a bit out of
control, producing toxic chemicals.

They'll have the choice of adapting to live in a polluted world and
still enjoy their enormous wealth without having to work for it, or go
back to the sort of civilization we live in where you have to work hard
to get a bit more than your basic needs. Given that it took us more than
40 years to finally do something about our CO2 emissions, I don't think
we can be very confident that they'll promptly revert to the old way of
living. A big obstacle being also that just like in case of climate
change, this requires a global consensus. It's no good if China decides
that self-replicating machines are with the pollution they produce and
then refuses to put a stop to using them.

So, I see this as just another phase in the evolution of life where
technological lifeform arise and end up replacing all biological
lifeforms except microbes.

Saibal

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

> <https://groups.google.com/g/extropolis>
> fal

[Tomasz Rola]

On Fri, May 20, 2022 at 06:21:53AM -0400, John Clark wrote:
> On Fri, May 20, 2022 at 2:48 AM Tomasz Rola <rto...@ceti.pl> wrote:
>

> *> I think I can easily wait five more years and see if it all boils down
> > to the mix of hype, half hype and enthusiastic lies...*

>
>
> In five years people will be saying what people have always been saying
> about AI, yes the progress made five years ago was real but there will be
> no further progress in the next five years, in fact there will be no

> further progress in AI *EVER*; and then they'll get all misty eyed and

> start talking about some secret sauce that only biological life in general
> and humans in particular have.

Uh huh. Becoming misty eyed, this will be a great experience to have.

> *> I really wonder, what makes you think this Universe is not engineered?
> > How would one test it for signs of "engineeredness"?*

>
>
> If the universe was engineered then when we look at the Andromeda galaxy we
> shouldn't see light and other high-quality photons radiating uselessly into
> infinite space, we should only see low energy infrared and radio waves
> because even civilizations that prefer to live in a virtual world would
> need power to run their computers. And if even one individual in one
> civilization in the Andromeda galaxy decided to make a Von Neumann probe
> that's exactly how their galaxy would look to us, except that we wouldn't
> be here to look at it because one of their probes would've arrived at our
> galaxy by now. All you need is one probe, and once Drexler style
> Nanotechnology has been developed, which I remind you would require no
> scientific breakthroughs just better engineering, everything could be
> divided into jist 2 classes, things that are easy and cheap to make, and
> things that are impossible to make because they violate the fundamental
> laws of physics. Nothing would be possible but expensive and hard to make.

I meant a bit different engineered Universe, something on a scale
comparable to a bubble, say, fourty bilion ly in radius? On such
scale, I think your assumptions simply do not apply.

Also, you seem to think that a "normal" or "most probable" way to
proceed for technological civilisation is to go with such excessive
consumption that it would require sucking resources and energy of the
whole galaxy. Myself, I consider such way to be both abnormal and
abysmal. I suppose that if our civilisation is classified one day, it
will be something like "compulsive-psychotic" type.

> That's why I think we must be the only intelligent species in the
> observable universe, it is after all a finite space so somebody has to be
> first. Some may argue for other explanations for the Fermi paradox, such as
> civilizations destroying themselves in a nuclear war, but I don't think

Perhaps we do not see, do not hear from C-P civilisations, because for
them, there is a limit which they can hardly cross. Something like, I
would not expect a compulsive fried chicken eater to build a bridge
across Grand Canyon. Especially if all restaurants are on this side.

So, a C-P will go on for a while, and later probably
degenerates. Excessive consumption cannot go forever when resources
are limited (and guess what, a galaxy is limited too). But excessive
piety can go for a very long time with minimal consumption.

Once C-P, always C-P.

Now given that "a galaxy is limited, too", why some intelligent,
non-C-P folks (aliens, not from this planet, obviously), would want to
engage in sucking the galaxy? Knowing very well, because they
understand exponential functions better than humans, knowing very well
that galaxy has some limits and they will not grow forever. Since they
are going to stop anyway, they can as well stop now.

This is why we do not see, do not hear from non-C-P civilisations. But
they do not have to degenerate, because they still have choices and
can make them at will.

Since they can have choices, they can travel around. But if they see
someone who is stupid enough to make von Neumann probe, they will
execute such idiot in the name of their own safety. Clearly, the guy
cannot restrain himself, so they will restrain him for everybody's
good.

[...]

> practical, and by that point they'd almost certainly have the ability to
> make Von Neumann machines. Hell, I think humans will almost certainly have
> the capability to make Von Neumann probes in less than 100 years, probably
> less than 50, maybe much less.

Oh really. So we have maybe fifty, maybe hundred years to live.

[Lawrence Crowell]

The numbers are way off. I suspect there may only be a few thousand planets with complex life in any galaxy. I would then say that given life has been here in complex form for 600 million years and that Homo sapiens have been here 100,000 years this means at any given time using this as a guide  each planet with complex life has a 1.7x10^{-4} probability of having intelligent life. This is even smaller when one considers how long we have had advanced technology, or about 3x10^{-7}. So this might mean that at any given time a galaxy might have 1/1000 chance of having ETI. Maybe then on the Hubble frame, or on our past light cone, there is on average an ETI on every 1000 galaxies. Now consider there is probably some filter that selects out ETIs. Either they annihilate themselves with nuclear explosive, or they collapse the life support system of their planet or .   , a range of things we may do to ourselves that prevents them from ever developing a von Neumann probe. 

LC

[John Clark]

On Sun, May 22, 2022 at 6:59 PM Tomasz Rola <rto...@ceti.pl> wrote:

>  you seem to think that a "normal" or "most probable" way to
proceed for technological civilisation is to go with such excessive
consumption that it would require sucking resources and energy of the
whole galaxy. Myself, I consider such way to be both abnormal and abysmal.

Even if wanting to do such a thing is abnormal if there are millions of civilizations and billions or trillions of individuals in each civilization then it's inevitable that some of them, or at least one of them, will want to do something that is not "normal", hell even I would want to do it although i'll concede I may not be entirely normal.  As for "abysmal", well I can only speak for myself but I don't find anything particularly ignoble about wanting to reorganize dumb brainless matter into matter that can can think and solve some of the mysteries of the universe and even contemplate it's own existence, in fact I think such an activity would be about as noble as it's possible to be. 

> Perhaps we do not see, do not hear from C-P civilisations, because for them, there is a limit which they can hardly cross.

What limit are you referring to?  

>  Excessive consumption cannot go forever

No activity can go on forever but if you stop letting all those photons from trillions of suns radiate uselessly into infinite space and instead harness them to perform some useful work then things can go on for a hell of a lot longer than if you have some irrational moral scruple against increasing the amount of intelligent activity in the universe. 

> This is why we do not see, do not hear from non-C-P civilisations.

So you think the reason the night sky does show any indication of having been  engineered is that 100% of the intelligent beings in the observable universe have the exact same likes and dislikes as you do, not 99.99999999% but 100% . 

 > if they see someone who is stupid enough to make von Neumann probe, they will execute such idiot in the name of their own safety.

That would be impossible because if they have von Neumann probes they can perform vastly more computations than those like you  who refuse to make von Neumann probes, and so they will be much smarter than you. And it's just not possible to outsmart somebody who is smarter than you.   

 > Hell, I think humans will almost certainly have the capability to make Von Neumann probes in less than 100 years, probably less than 50, maybe much less.

> Oh really.

Yes really.

> So we have maybe fifty, maybe hundred years to live.

I would be very surprised if there are still biological humans around in 100 years.  

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

9u7

 

[spudb...@aol.com]

I am thinking for CO2 and methane, you are not giving renewables a chance,  and the possibility of CO2 and methane extraction from the atmosphere a fair try, as yet. The greens worldwide are keen on attacking the FF vendors, and low key on providing alternatives first. Most seem to be from areas outside of technology and seem foolish because they are determined to shut down FF before then install renewables. 

For automation, the test bed is the solar system, and we need to give machine intel it's chance to build space habitats out of local materials. Eventually, depending on "smartness," we should be prepared to share the solar output, 50/50. But this deal is obviously for a far future generation, rather than dwellers of the inner solar system. 

> fal

--
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-list+unsub...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/225b1a4ac581a2b1c7dd7e1c947c7dc1%40zeelandnet.nl.

[John Clark]

On Sun, May 22, 2022 at 7:18 PM Lawrence Crowell <goldenfield...@gmail.com> wrote:

>I suspect there may only be a few thousand planets with complex life in any galaxy. I would then say that given life has been here in complex form for 600 million years and that Homo sapiens have been here 100,000 years this means at any given time using this as a guide  each planet with complex life has a 1.7x10^{-4} probability of having intelligent life. This is even smaller when one considers how long we have had advanced technology, or about 3x10^{-7}. So this might mean that at any given time a galaxy might have 1/1000 chance of having ETI. Maybe then on the Hubble frame, or on our past light cone, there is on average an ETI on every 1000 galaxies.

The trouble is it's extremely difficult to obtain useful statistics if you only have one example. On Earth life occurred almost as soon as liquid water could exist on its surface, but does that mean life is inherently easy to make? Maybe not because the only example we have is a planet in which not only life developed but intelligent life did, so life had to have started quickly here, perhaps freakishly quickly, because if it had taken much longer the sun would be turning into a red giant about the time intelligent creatures started building a civilization, and that would make any sort of life on the planet impossible. And even after life started all you had was simple Prokaryotic cells, it took another 2 billion years for far more complex Eukaryotic cells to evolve, and then another 1.5 billion years for multicellular life to evolve, and 500 million years after that to get a technological civilization which only happened a few thousand years ago.

How typical is any of this? Nobody knows but we have hints; most major evolutionary developments, such as the invention of the eye and flight, evolved independently more than once, sometimes dozens of times but intelligence, defined as the ability to make a radio, occurred only once. But even when we look deep into space with enormous telescopes we don't see even the slightest hint of engineering.  

 

> Now consider there is probably some filter that selects out ETIs. Either they annihilate themselves with nuclear explosive, or they collapse the life support system of their planet or .   , a range of things we may do to ourselves that prevents them from ever developing a von Neumann probe. 

The human race only needs to hang together for a few more decades before a von Neumann probe gets made, I don't think it's reasonable to expect that every single civilization without exception destroys itself just an instant before it has the ability to make such a thing, especially when you consider that if intelligence is as common as some claim then some of those civilizations must have started on planets where the element Uranium is much rarer in its surface crust than it is on the Earth. On Earth's surface Uranium is not particularly rare, it's far more common than Gold, but if it was as rare as Osmium, Rhodium or Iridium we'd have no nuclear weapons today and wouldn't have any in the future until asteroid mining became practical, and by then making a von Neumann probe would be practical too.

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

swl

[Lawrence Crowell]

On Monday, May 23, 2022 at 7:22:47 AM UTC-5 johnk...@gmail.com wrote:

On Sun, May 22, 2022 at 7:18 PM Lawrence Crowell <goldenfield...@gmail.com> wrote:

>I suspect there may only be a few thousand planets with complex life in any galaxy. I would then say that given life has been here in complex form for 600 million years and that Homo sapiens have been here 100,000 years this means at any given time using this as a guide  each planet with complex life has a 1.7x10^{-4} probability of having intelligent life. This is even smaller when one considers how long we have had advanced technology, or about 3x10^{-7}. So this might mean that at any given time a galaxy might have 1/1000 chance of having ETI. Maybe then on the Hubble frame, or on our past light cone, there is on average an ETI on every 1000 galaxies.

The trouble is it's extremely difficult to obtain useful statistics if you only have one example. On Earth life occurred almost as soon as liquid water could exist on its surface, but does that mean life is inherently easy to make? Maybe not because the only example we have is a planet in which not only life developed but intelligent life did, so life had to have started quickly here, perhaps freakishly quickly, because if it had taken much longer the sun would be turning into a red giant about the time intelligent creatures started building a civilization, and that would make any sort of life on the planet impossible. And even after life started all you had was simple Prokaryotic cells, it took another 2 billion years for far more complex Eukaryotic cells to evolve, and then another 1.5 billion years for multicellular life to evolve, and 500 million years after that to get a technological civilization which only happened a few thousand years ago.

How typical is any of this? Nobody knows but we have hints; most major evolutionary developments, such as the invention of the eye and flight, evolved independently more than once, sometimes dozens of times but intelligence, defined as the ability to make a radio, occurred only once. But even when we look deep into space with enormous telescopes we don't see even the slightest hint of engineering.  

A galaxy dominated by von Neumann probes would like bare some signatures of such. So far nothing is evident. Of course galaxies we observer are in the past. We could just be one of the early IGUS or intelligent life forms on the Hubble frame. 

LC

[spudb...@aol.com]

Then, there's this! 

 Planets of binary stars as possible homes for alien life -- ScienceDaily

Tatooine!

--

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/8c336ffe-4f23-4c99-b16c-823e601da831n%40googlegroups.com

.

[spudb...@aol.com]

And, there is this article-

How advanced civilizations could leave us a message of their presence.

We Were Here - Nautilus | Science Connected

Seems like radio would be a lot easier? Maybe this would offer up aliens with a very large budget? 

https://groups.google.com/d/msgid/everything-list/963607691.1164396.1653362852700%40mail.yahoo.com

.

[spudb...@aol.com]

Do note l professori, that this idea even to a shmoop like myself occurred decades ago as a explanation over how or why no ETI signals! So I am real good with your comment and observation that yeah how well a radio signal propagates through the Milky Way may be highly and profoundly limited. So I'm not exactly biting through my lower lip at your reminder. And I say thanks! Signals no propagate, signals no get received. No ticket no shirty. And I say thank ee' master.

Sent from the all new AOL app for Android

On Tue, May 24, 2022 at 6:58 AM, Lawrence Crowell

<goldenfield...@gmail.com> wrote:

Sorry to say, but radio signals may not make it that far through the galaxy, Charged plasma in interstellar space may attenuate any EM signal.

LC