How close are we to producing a self-replicating machine

  • B
  • Thread starter bauhaus
  • Start date
  • Tags
    Machine
In summary, the conversation discussed the concept of self-replicating machines and their potential uses in various scenarios, such as space exploration and colonization. The idea of self-replicating machines is still theoretical, with only a few demonstration models, such as the rip rap 3d printer, being produced. The conversation also touched upon the issue of how a self-replicating machine would power itself without violating the Conservation of Energy, with the suggestion of using starlight near suns as a possible source of energy. It was also mentioned that a fully self-replicating machine would need to gather its own raw materials, similar to how yeast turns raw materials into copies of itself. The conversation also delved into the possibility of
  • #1
bauhaus
15
0
Are self-replicating machines still theoretical or have demonstration models (other than the rip rap 3d printer)been produced?
 
Physics news on Phys.org
  • #2
They are still theoretical.
 
  • #3
How could a self-replicating machine power itself without violating Conservation of Energy? From starlight near suns? Would it then just be drifting between stars?
 
  • #4
I guess a fully self replicating machine would also need to gather all the raw material required to make itself. In the human world we seem to be heading the other way. Not many companies make everything including their tools from raw materials. Even companies that do process raw materials buy-in tools, machinery, computers, trucks etc.

It's possible to imagine a biological "machine" like a yeast that when released into a pool of raw material turns it all into copies of itself. See Grey Goo..

https://en.wikipedia.org/wiki/Grey_goo
 
  • #5
samsara15 said:
How could a self-replicating machine power itself without violating Conservation of Energy?

Does it have to power itself? Humans self replicate without violating CoE.
 
  • #6
Humans obtain chemical energy from what they eat. ATP to ADP, etectera.

Ultimately, all the energy that powers biological organisms comes from sunlight.

Sunlight has very little energy between stars.
 
  • #7
What does 'between the stars' have to do with anything?
 
  • #8
I presumed that the self-replicating bots would be used to colonize other planets in other star systems, which apparently, in this case, was an incorrect assumption. That is the most common reason why I have heard people propose the creation of self-replicating bots.
 
  • #9
My guess is that an artificial virus will be the first. In fact, a google search for "artificial virus synthesis" returned more than 1,300,000 hits.
 
  • #10
anorlunda said:
My guess is that an artificial virus will be the first.

A virus is not self-replicating. I think an artificial bacterium will be the first and there already are almost full artificial bacteria.
 
  • #11
There are already self replicating machines.

You might with to use a far more restrictive description of what you mean.
 
  • #12
Thanks for the comments.
So from what I see so far only the rip rap is close and artificial bacteria. From what I've read so far they are theoretically possible, no question. The science and engineering aspects are sound, just never actually fabricated yet. How many months are we away from constructing a 100% or close to 100% device? I mean a riprap 3d printer can already print more than half its own parts to create a daughter machine. reprap.org I have been researching this issue but could not make any definitive determination.
 
  • #13
This is an ill-defined question. The part that you are missing is "from what"? We could certainly make a self-replacating machine if the input is self-replacating machine parts. We certainly could not do it if the input was a bunch if hydrogen gas. How close we are depends on where in between you are.
 
  • Like
Likes samsara15 and billy_joule
  • #14
bauhaus said:
Thanks for the comments.
So from what I see so far only the rip rap is close and artificial bacteria. From what I've read so far they are theoretically possible, no question. The science and engineering aspects are sound, just never actually fabricated yet. How many months are we away from constructing a 100% or close to 100% device? I mean a riprap 3d printer can already print more than half its own parts to create a daughter machine. reprap.org I have been researching this issue but could not make any definitive determination.
Every candidate, however, requires some sort of energy input. A riprap 3d printer requires electricity. Artificial bacteria eat things. Surely you are not visualizing a self-replicating machine as some sort of perpetual motion system?
 
  • #15
I don't think the OP is suggesting that no input of energy is required.
However, given that energy IS required, I think it's unlikely that we could do better than making modified versions of natural organisms.
 
  • #16
Um, that's "RepRap". https://en.wikipedia.org/wiki/RepRap_Project

Photograph of one RepRap that produced the elements of another RepRap.
First_replication.jpg

Note: The assembly robots, dressed in sky-blue and black shirts, were not produced by the Parent RepRap.
 
  • #17
What
 
  • #18
Ignore the above. Finger trouble and this app won't let me edit posts.

If nobody is trying to build a self replicating machine it will take a long time to make one. Who is and why?
 
  • #20
samsara15 said:
I presumed that the self-replicating bots would be used to colonize other planets in other star systems, which apparently, in this case, was an incorrect assumption. That is the most common reason why I have heard people propose the creation of self-replicating bots.
When imagined in the capacity of colonization, they don't replicate in deep space - there's no reason to, nor - as has been pointed out - any way to. What they typically do is seek out resources needed, for example finding planets or asteroids, and harvesting the required materials.

Self-replicating machines, thus, need to not be merely factories for machining and assembly, but also ore mining and processing plants.
 
  • #21
What is the point of self-replicating anyway? It's not the first time I've seen the question and I've never understood why bother or why focus on this when there are major/important related inventions that don't exist yet...like general purpose robots.
 
  • #22
bauhaus said:
Thanks for the comments.
So from what I see so far only the rip rap is close and artificial bacteria. From what I've read so far they are theoretically possible, no question. The science and engineering aspects are sound, just never actually fabricated yet. How many months are we away from constructing a 100% or close to 100% device? I mean a riprap 3d printer can already print more than half its own parts to create a daughter machine. reprap.org I have been researching this issue but could not make any definitive determination.
RepRap is hardly close.
It may make some of the parts, but, even if and when it does, it needs extra input to do the following:
- it does not assemble itself
- it does not produce the computer, nor the program, needed to command the manufacture of parts..
- it does not provide its own method of energy acquisition.

All three of those are outside inputs - assemblage, energy, and intelligence.
It is a long way from being self replicating, if the definition you are using is completely autonomous.
 
  • #23
russ_watters said:
What is the point of self-replicating anyway?

I guess you do not mean self-replication at all (because you wouldn't exist without it) but artificial self-replicating machines. A practical example would be building an industry on the Moon. Without self-replication this would be incredible expensive or even impossible due to the transport costs from Earth to Moon. These costs would be obsolete with robots that can build infrastructure and machines (including copies of themselves) using local resources.
 
  • #24
Thanks for the ideas.
Vanadium 50 said:
This is an ill-defined question.
2946460 My response is yes, i should have mentioned that m-type asteroids and possibly icy asteroids would be the exclusive source for raw materials. Difficulty, the "brains" part of the self-replication machine would be the most difficult to reproduce.
"RepRap is hardly close.
It may make some of the parts, but, even if and when it does, it needs extra input to do the following:
- it does not assemble itself
- it does not produce the computer, nor the program, needed to command the manufacture of parts.
- it does not provide its own method of energy acquisition.

All three of those are outside inputs - assemblage, energy, and intelligence.
It is a long way from being self replicating, if the definition you are using is completely autonomous." 256bits
I should have thought about some of the points the commenters have come up with before I posted.
Also, good point. 1. Trivial to add as a feature with off the shelf current technology.
2. "Non-Trivial task."This is the dense part, the hardest part of the process is reproducing a cpu/brain for the machine. I wonder what the current technology is regarding completely printed cpu and computing components. This would seem to be the only substantial barrier to producing a demonstration model of a Von Neumann Probe
3. Trivial task. The program can be transferred, no need to reinvent the wheel, especially when the purpose is simply to make Von Neumann probes.
4. Trivial task. Add a energy production system, Perovskite cells can be made by simply printing a layer of perovskite on to a plastic backing. If not photovoltaics then heat produced energy production, Stirling Engine.
 
  • #25
DrStupid said:
russ_watters said:
What is the point of self-replicating anyway?
A practical example would be building an industry on the Moon. Without self-replication this would be incredible expensive or even impossible due to the transport costs from Earth to Moon. These costs would be obsolete with robots that can build infrastructure and machines (including copies of themselves) using local resources.
At the risk of putting words in Russ' mouth, I'm fairly certain he is well-aware what we might do with it. His question is more along the lines of there are much larger and nearer hurdles to overcome.

Consider:
"How close are we to an intergalactic drive, for colonizing other galaxies?"
"What is the point in looking at an intergalactic drive? We haven't even made it off our own planet yet. Never mind to other stars in our own galaxy."A self-replicating robot is useless if it does not have the intelligence to operate autonomously in the circumstances it will encounter without human supervision. How would a robot prospect for minerals? We're many, many decades away from a robot that could do any general task under direct human supervision never mind on its own, never mind also able to replicate.
 
  • Like
Likes russ_watters
  • #26
DaveC426913 said:
His question is more along the lines of there are much larger and nearer hurdles to overcome.

Consider:
"How close are we to an intergalactic drive, for colonizing other galaxies?"
"What is the point in looking at an intergalactic drive? We haven't even made it off our own planet yet. Never mind to other stars in our own galaxy."

.
1. That concern is trivial, within only a few solar cycles a von neumann system could include a em drive, a canne drive. Building a intergalactice drive is unlikely to have any exponential effect that from which a von neumann system benefits.
2. Also a trivial concern see above. "How close are we to an intergalactic drive, for colonizing other galaxies?"A self-replicating robot can make enough reconfigurable parts that it can eventually create megastructures for power production and mega factory systems to create cheap endless supply of em drives, cannae drives, in other words RF resonant cavity thrusters to endlessly explore every part of the universe. Not much is needed beyond a metallic asteroid or maybe an additional ice asteroid. Humans have had the knowledge to make real life physical von neumann devices since the 80's, no?
See: the NASA publications "Advanced Automation for Space Missions" which was the title for the final report of a NASA summer study, conducted in 1980 requested by President Jimmy Carter. Just the study cost 11.7 million dollars. Commentators have stated, "The result of the study was a realistic proposal for a self-replicating automated lunar factory system, capable of exponentially increasing productive capacity and, in the long run, exploration of the entire galaxy within a reasonable timeframe."
 
  • #27
bauhaus said:
1. That concern is trivial,
You completely and utterly missed the point of my post.

The reference to intergalactic drive was purely analogy.I might as well have used an analogy with a learning baby.

"Are we near a point where a one year old human baby can solve differential equations?"
"What's the point in that? We don't even have a one year old baby that can tie its own shoes."

To reiterate. von Neumanesque capabilities are completely useless for a device that cannot yet perform general unsupervised activity such as prospecting for minerals, mining those minerals, or picking lint out of its own navel.

(OK, we might have the technology to allow it to pick lint out of its own navel.)

We are nowhere near a device that can - even under direct human supervision - perform the multitude of general tasks one might need to warrant a device also being able to reproduce itself.

The very premise of a machine that can reproduce itself is that it is beyond human influence and control.

And if you disagree with this, you can take it up with Russ! :wink:
 
Last edited:
  • Like
Likes russ_watters
  • #28
Hmm
[QUOTE="DaveC426913, post: 5188856, member: 15808"
To reiterate. von Neumanesque capabilities are completely useless for a device that cannot yet perform general unsupervised activity such as prospecting for minerals, mining those minerals. The very premise of a machine that can reproduce itself is that it is beyond human influence and control. :wink:[/QUOTE]

I respectfully disagree.

The Large Hadron Collider cost $13.25 billion. After doing some cursory research I am confident that humanity can create a machine that can reproduce itself for much less than the cost of the LHC. It appears that its not true that we are no where near a device when you consider the following, completely closed known designs. : Project Daedalus
Examples from the book Kinematic Self-Replicating Machines.
[Mentor's Note: Post edited to condense links]
 
Last edited by a moderator:
  • #29
Sorry. That is one giant link, linking to Project Daedalus. Can you edit? And perhaps, elaborate how it is salient to your counterargument?
 
  • #30
bauhaus said:
I respectfully disagree.

The Large Hadron Collider cost $13.25 billion. After doing some cursory research I am confident that humanity can create a machine that can reproduce itself for much less than the cost of the LHC. It appears that its not true that we are no where near a device when you consider the following, completely closed known designs.

I respectfully disagree with you. Neither of your links have working designs or designs that we are close to being able to create. In fact, I wouldn't even call them designs at all, more like concepts.
 

1. How close are we to producing a self-replicating machine?

Currently, we have made significant progress in creating self-replicating machines in the field of nanotechnology. However, we are still far from producing a fully functional and efficient self-replicating machine that can replicate itself without human intervention.

2. What are the challenges in creating a self-replicating machine?

One of the biggest challenges in creating a self-replicating machine is designing a system that can accurately and efficiently replicate itself without any errors. Additionally, ensuring that the machine has the ability to acquire and utilize energy and materials from its environment is also a major hurdle.

3. Can self-replicating machines be used for beneficial purposes?

Yes, self-replicating machines have the potential to revolutionize various industries such as manufacturing, medicine, and space exploration. They can be programmed to perform tasks that are too dangerous or tedious for humans, leading to increased efficiency and safety.

4. Are there any ethical concerns surrounding the development of self-replicating machines?

As with any emerging technology, there are ethical concerns that need to be addressed when it comes to self-replicating machines. These include the potential for these machines to replicate uncontrollably and cause harm to the environment or humans, as well as the possibility of creating a new form of life.

5. What are the potential risks associated with self-replicating machines?

Some of the potential risks associated with self-replicating machines include the loss of control over their replication process, leading to unintended consequences. There is also a concern that these machines could outcompete and replace human labor, leading to job loss. Additionally, there is a risk of these machines being used for malicious purposes if they fall into the wrong hands.

Similar threads

  • Other Physics Topics
Replies
7
Views
3K
Replies
2
Views
949
  • Programming and Computer Science
Replies
23
Views
1K
  • Materials and Chemical Engineering
Replies
17
Views
2K
  • Biology and Medical
Replies
13
Views
767
  • Classical Physics
Replies
5
Views
3K
Replies
15
Views
2K
  • Other Physics Topics
Replies
27
Views
1K
  • Differential Equations
Replies
1
Views
646
  • General Engineering
Replies
22
Views
916
Back
Top