Early probes are one thing, but can we build a continuing presence among the stars, human or robotic? An evolutionary treatment of starflight sees it growing from a steadily expanding presence right here in our Solar System, the kind of infrastructure Alex Tolley examines in the essay below. How we get to a system-wide infrastructure is the challenge, one analyzed by a paper that sees artificial intelligence and 3D printing as key drivers leading to a rapidly expanding space economy. The subject is a natural for Tolley, who is co-author (with Brian McConnell) of A Design for a Reusable Water-Based Spacecraft Known as the Spacecoach (Springer, 2016). An ingenious solution to cheap transportation among the planets, the Spacecoach could readily be part of the equation as we bring assets available off-planet into our economy and deploy them for even deeper explorations. Alex is a lecturer in biology at the University of California, and has been a Centauri Dreams regular for as long as I can remember, one whose insights are often a touchstone for my own thinking.
by Alex Tolley
Crewed starflight is going to be expensive, really expensive. All the various proposed methods from slow world ships to faster fusion vessels require huge resources to build and fuel. Even at Apollo levels of funding in the 1960’s, an economy growing at a fast clip of 3% per year is estimated to need about half a millennium of sustained growth to afford the first flights to the stars. It is unlikely that planet Earth can sustain such a sizable economy that is millions of times larger than today’s. The energy use alone would be impossible to manage. The implication is that such a large economy will likely be solar system wide, exploiting the material and energy resources of the system with extensive industrialization.
Economies grow by both productivity improvements and population increases. We are fairly confident that Earth is likely nearing its carrying capacity and certainly cannot increase its population even 10-fold. This implies that such a solar system wide economy will need huge human populations living in space. The vision has been illustrated by countless SciFi stories and perhaps popularized by Gerry O’Neill who suggested that space colonies were the natural home of a space faring species. John Lewis showed that the solar system has immense resources to exploit that could sustain human populations in the trillions.
Image credit: John Frassanito & Associates
But now we run into a problem. Even with the most optimistic estimates of reduced launch costs, and assuming people want to go and live off planet probably permanently, the difficulties and resources needed to develop this economy will make the US colonization by Europeans seem like a walk in the park by comparison. No doubt it can be done, but our industrial civilization is little more than a quarter of a millennium old. Can we sustain the sort of growth we have had on Earth for another 500 years, especially when it means leaving behind our home world to achieve it? Does this mean that our hopes of vastly larger economies, richer lives for our descendents and an interstellar future for humans is just a pipe dream, or at best a slow grind that might get us there if we are lucky?
Well, there may be another path to that future. Philip Metzger and colleagues have suggested that such a large economy can be developed. More extraordinary, that such an economy can be built quickly and without huge Earth spending, starting and quickly ending with very modest space launched resources. Their suggestion is that the technologies of AI and 3D printing will drive a robotic economy that will bootstrap itself quickly to industrialize the solar system. Quickly means that in a few decades, the total mass of space industrial assets will be in the millions of tonnes and expanding at rates far in excess of our Earth-based economies.
The authors ask, can we solve the launch cost problem by using mostly self-replicating machines instead? This should remind you of the von Neumann replicating probe concept. Their idea is to launch seed factories of almost self-replicating robots to the Moon. The initial payload is a mere 8 tonnes. The robots will not need to be fully autonomous at this stage as they can be teleoperated from Earth due to the short 2.5 second communication delay. They are not fully self-replicating at this stage as need for microelectronics is best met with shipments from Earth. Almost complete self-replication has already been demonstrated with fabs, and 3D printing promises to extend the power of this approach.
The authors assume that initial replication will neither be fully complete, nor high fidelity. They foresee the need for Earth to ship the microelectronics to the Moon as the task of building fabs is too difficult. In addition, the materials for new robots will be much cruder than the technology earth can currently deliver, so that the next few generations of robots and machinery will be of poorer technology than the initial generation. However the quality of replication will improve with each generation and by generation 4, a mere 8 years after starting, the robot technology will be at the initial level of quality, and the industrial base on the Moon should be large enough to support microelectronics fabs. From then on, replication closure is complete and Earth need ship no further resources to the Moon.
| Gen | Human/Robotic Interaction | Artificial Intelligence | Scale of Industry | Materials Manufactured | Source of Electronics |
|---|---|---|---|---|---|
| 1.0 | Teleoperated and/or locally operated by a human outpost | Insect-like | Imported, small-scale, limited diversity | Gases, water, crude alloys, ceramics, solar cells | Import fully integrated machines |
| 2.0 | Teleoperated | Lizard-like | Crude fabrication, inefficient, but greater throughput than 1.0 | (Same) | Import electronics boxes |
| 2.5 | Teleoperated | Lizard-like | Diversifying processes, especially volatiles and metals | Plastics, rubbers, some chemicals | Fabricate crude components plus import electronics boxes |
| 3.0 | Teleoperated with experiments in autonomy | Lizard-like | Larger, more complex processing plants | Diversify chemicals, simple fabrics, eventually polymers | Locally build PC cards, chassis and simple components, but import the chips |
| 4.0 | Closely supervised autonomy | Mouse-like | Large plants for chemicals, fabrics, metals | Sandwiched and other advanced material processes | Building larger assets such as lithography machines |
| 5.0 | Loosely supervised autonomy | Mouse-like | Labs and factories for electronics and robotics. Shipyards to support main belt. | Large scale production | Make chips locally. Make bots in situ for export to asteroid belt. |
| 6.0 | Nearly full autonomy | Monkey-like | Large-scale, self-supporting industry, exporting industry to asteroid main belt | Makes all necessary materials, increasing sophistication | Makes everything locally, increasing sophistication |
| X.0 | Autonomous robotics pervasive throughout Solar System enabling human presence | Human-like | Robust exports/imports through zones of solar system | Material factories specialized by zone of the Solar System | Electronics factories in various locations |
Table 1. The development path for robotic space industrialization. The type of robots and the products created are shown. Each generation takes about 2 years to complete. Within a decade, chip fabrication is initiated. By generation 6, full autonomy is achieved.
| Asset | Qty. per set | Mass minus Electronics (kg) | Mass of Electronics (kg) | Power (kW) | Feedstock Input (kg'hr) | Product Output (kg/hr) |
|---|---|---|---|---|---|---|
| Power Distrib & Backup | 1 | 2000 | ----- | ---- | ---- | ---- |
| Excavators (swarming) | 5 | 70 | 19 | 0.30 | 20 | ---- |
| Chem Plant 1 - Gases | 1 | 733 | 30 | 5.58 | 4 | 1.8 |
| Chem Plant 2 - Solids | 1 | 733 | 30 | 5.58 | 10 | 1.0 |
| Metals Refinery | 1 | 1019 | 19 | 10.00 | 20 | 3.15 |
| Solar Cell Manufacturer | 1 | 169 | 19 | 0.50 | 0.3 | ---- |
| 3D Printer 1 - Small Parts | 4 | 169 | 19 | 5.00 | 0.5 | 0.5 |
| 3D Printer 2 - Large Parts | 4 | 300 | 19 | 5.00 | 0.5 | 0.5 |
| Robonaut assemblers | 3 | 135 | 15 | 0.40 | ---- | ---- |
| Total per Set | ~7.7 MT launched to Moon | 64.36 kW | 20 kg regolith/hr | 4 kg parts/hr |
||
Table 2. The products and resources needed to bootstrap the industrialization of the Moon with robots. Note the low mass needed to start, a capability already achievable with existing technology. For context, the Apollo Lunar Module had a gross mass of over 15 tonnes on landing.
The authors test their basic model with a number of assumptions. However the conclusions seem robust. Assets double every year, more than an order of magnitude faster than Earth economic growth.
Figure 13 of the Metzger paper shows that within 6 generations, about 12 years, the industrial base off planet could potentially be pushing towards 100K MT.
Figure 14 of the paper shows that with various scenarios for robots, the needed launch masses from Earth every 2 years is far less than 100 tonnes and possibly below 10 tonnes. This is quite low and well within the launch capabilities of either government or private industry.
Once robots become sophisticated enough, with sufficient AI and full self-replication, they can leave the Moon and start industrializing the asteroid belt. This could happen a decade after initiation of the project.
With the huge resources that we know to exist, robot industrialization would rapidly, within decades not centuries, create more manufactures by many orders of magnitude than Earth has. Putting this growth in context, after just 50 years of such growth, the assets in space would require 1% of the mass of the asteroid belt, with complete use within the following decade. Most importantly, those manufactures, outside of Earth’s gravity well, require no further costly launches to transmute into useful products in space. O’Neill colonies popped out like automobiles? Trivial. The authors suggest that one piece could be the manufacture of solar power satellites able to supply Earth with cheap, non-polluting power, in quantities suitable for environmental remediation and achieving a high standard of living for Earth’s population.
With such growth, seed factories travel to the stars and continue their operation there, just as von Neumann would predict with his self-replicating probes. Following behind will be humans in starships, with habitats already prepared by their robot emissaries. All this within a century, possibly within the lifetime of a Centauri Dreams reader today.
Is it viable? The authors believe the technology is available today. The use of telerobotics staves off autonomous robots for a decade. In the 4 years since the article was written, AI research has shown remarkable capabilities that might well increase the viability of this aspect of the project. It will certainly need to be ready once the robots leave the Moon to start extracting resources in the asteroid belt and beyond.
The vision of machines doing the work is probably comfortable. It is the fast exponential growth that is perhaps new. From a small factory launched from Earth, we end up with robots exploiting resources that dwarf the current human economy within a lifetime of the reader.
The logic of the model implies something the authors do not explore. Large human populations in space to use the industrial output of the robots in situ will need to be launched from Earth initially. This will remain expensive unless we are envisaging the birthing of humans in space, much as conceived for some approaches to colonizing the stars. Alternatively an emigrant population will need to be highly reproductive to fill the cities the robots have built. How long will that take? Probably far longer, centuries, rather than the decades of robotic expansion.
Another issue is that the authors envisage the robots migrating to the stars and continuing their industrialization there. Will humans have the technology to follow, and if so, will they continue to fall behind the rate at which robots expand? Will the local star systems be full of machines, industriously creating manufactures with only themselves to use them? And what of the development of AI towards AGI, or Artificial General Intelligence? Will that mean that our robots become the inevitable dominant form of agency in the galaxy?
The paper is Metzger, Muscatello, Mueller & Mantovani, “Affordable, Rapid Bootstrapping of the Space Industry and Solar System Civilization,” Journal of Aerospace Engineering Volume 26 Issue 1 (January 2013). Abstract / Preprint.
And yet another issue is how this wealth gets spread among the human (and eventually robot?) population. Already so much of our (US) wealth is in so few hands that we have a dysfunctional economy. Will we find the wisdom to correct this or will we have thousands or millions of wealthy individuals and trillions of peons? That’s not the way to launch ourselves upon the stars.
Clearly catching the zeitgeist, although the super wealthy and plutocrats at the WEF still express puzzlement that there should be redistribution. The “Expanse” novels and tv series would indicate that the latter is more likely, at least with current economic structures.
I suspect even the densest plutocrat recognizes the need for a wealthy population to predate or parasitize upon. The issue may be more about who can keep playing the game of musical chairs long enough before the game has to stop. However this is digressing from the topic somewhat.
As I’m already receiving a string of highly political responses to the above, let me remind especially new readers that we don’t deal with political or religious issues here. The reason is straightforward: Such discussions quickly swamp the subject at hand and lead us far off topic. This has been the policy since the earliest days of the site. For those who argue that only their political or religious view will lead to an interstellar future, I can only say that this site is not the place to make that case.
Paul, this OP does not seem to be out of order here. The article is about the economics of solar system development and so is the post. Someone will launch this self-reproducing mining and manufacturing system to the Moon, and that someone will probably claim ownership. Ownership of wealth can involve politics or religion, but for this topic it doesn’t need to go that far, unless there is some outside technical issue with the lunar system (such as the Moon gets completely covered with solar panels because the shut off switch broke, or the robots decide to be a ‘Harsh Mistress’).
Micky, I’m not referring to the published comments above, which are fine. I’m talking about highly political ones that came in this morning, ones I haven’t published because they’re not on topic here. Trying to head more of such off.
Two thoughts here for me, An American $250 Billion pe year carbon taxe invested in the world’s capital markets results in almost $100 trillion in less than a century.I have proposed years ago this would be owned by future social security recipients 70 years from now.This could own the vast fleet of solar system space-based solar power.
https://yellowdragonblog.com/2014/01/28/a-carbon-tax-fueled-social-security-sovereign-wealth-fund/
Also, I have blogged on the importance of ESOP-owned entity’s for Mars and space colonization as well as here on earth to bring social justice to employee ownership of the surplus value labor. I have many blog posts that involve ESOP’s many having to do with ESOP’s owning automata that have replaced the human worker.
https://yellowdragonblog.com/2016/11/12/leveraged-esop-and-tax-exempt-bond-capitalized-mars-colonization-the-ethical-choice/
Thomas Piketty envisions the world where invested sovereign wealth funds will be 30% to 40% of the world’s GDP in 40 years.With the carbon tax fund, this might be higher.
In a neoliberal world of ESOP trustees and bankers, this world would not look different than ours except that ESOP employees have the right of the shareholder lawsuit. Governments would feel pressure to invests sovereign wealth funds in an ethical manner depending on the political cycle.
I agree
I have seen many thoughtful articles on world economics here and the ability to mount vast voyages of colonization
Scary comments on another website lead me to believe the solution will be for the rich to keep their wealth, and just eliminate the unproductive.
In such a world I wondcer what would be the motivation for interplanetary, let alone interstellar, travel.
In such a world I also question how science could progress. It could be decades, or even close to a century, between discovering something of interest in science before it becomes of benefit to humanity. Science would become more refining what we know, and not a search for what we don’t know.
Depressing. Guess I’ll listen to the songs from Moana again – now those are uplifting.
In how many hands are personal computers and smartphones? Automation and robotics are just machines connected to sensors and control devices, like a computer with software.
There is no reason to assume that bootstrappable automation & robotics will be any less widespread than existing technology. Hobbyists are already building their own examples, like computer hobbyists built their own computers 40 years ago.
Thank you very much for this very thought provoking post! Reading stuff like this keeps me young!
Lots of thoughts on this.
– A robot factory travelling through space at (for instance) nuclear S-I speeds of something like 5%c could create a presence in every star system in our 100k LY galaxy in 2 million years. A robotic presence like this could keep itself repaired and in operation until the stars go out. Our galaxy is getting on towards as old as the universe itself and two million years is not even an eye blink. The implication of this article is that if *ANY* other lifeform has developed to a technology level equivalent to what we have essentially now in the *ENTIRE* history of the galaxy, then we should expect to find functioning hardware in every star system *including ours*. If we look properly.
If the lifespan of the self-reproducing artefacts of a civilisation are no longer coupled to the lifespan of the biological civilisation that put them together, then you can draw a line through the ‘civilisation lifespan’ term in the Drake equation. If the artefacts are programmed to create a galaxy saturating presence and they are programmed to communicate in some way either with intelligent life or with each other, then the fraction of stars with a communicating presence around them can be set to 1.
If there has been more than one alien civilisation to scale this height before us, then I guess the second one would also be reasoning like this and doing whatever was needed to get in touch with the first one’s artefacts. Which I guess takes us to questions of motives (and possibly into game theory, another von Neumann invention).
You raise a good point about the impact of individual lifespans and civilizational ones. We’ve discussed Karl Schroeder’s novel Lockstep that also posits a solution to make interstellar civilization work. Charles Stross’ post-human, “Freyaverse” novels explores this in a way that is more in line with your thinking.
As always, we come back to the Fermi Question: “Where are they?”
More near term, an advanced robot economy raises the sorts of questions Asimov explored with his robot novels and short stories. I can imagine Neil S visualizing Earth’s citizens living in “Caves of Steel” conditions, while the wealthy spacers live far richer lives, supported by robots that have been banned on Earth.
One conceivable motive for creating such a robotic presence is to create something like the ‘foundation’. Whatever happens to Earth civilisation, there is an AI driven emergency rescue system with a complete historical and technological encyclopaedia and a humungous industrial capability to back it up.
In fact, all it would take is for a small number of people on Earth to get it to the gen-4 self sustaining level, and then planetary disaster on almost any scale could be sorted out in a few decades, whether humans themselves are significantly off world yet or not.
I do wonder whether the off worlders in a totally robotised economy would end up like Asimov’s spacers, or or more like the humans in Wall-E :)
Thank you for the reading tips.
I’m working on a draft of a novel premised on reproducing a civilization (and some subset of living wilderness) from data. If you grant all the assumptions in this article – self-reproducing robotic manufacturing colonies – you should grant that manufacturing flesh and blood humans (along with other lifeforms that we need and like) is not a much bigger stretch. Building a self-reproducing cell from inert matter is on the horizon. Programming that cell with manufactured DNA is not much different from what has already been achieved in Craig Venter’s lab. That first easy-to-build cell could be modified to produce more flexible descendants that could run the first generation of bioreactors, which, after several generations of carefully directed “mutation,” enable the creation of a functioning eukaryote. I imagine that we could design a cascade where a simple organism is created, and once it’s viable, its DNA is modified so that it is able to produce a physical structure that could function as an “egg” or a “seed” of a more complex organism. Once we have such an “egg” we can animate it with the genetic sequence of that more complex organism, and keep climbing these rungs successively until we’ve made trees, ladybugs, tomatoes and humans. Of course the labs that perform all these steps would be fully automated, and they would not be starting with any physical material from our biosphere; they would build life from the stuff near them. All they would need is a small payload to start the bootstrapping process – probably some sensing tools, primitive excavation tools, a tiny 3D printer, a source of power, and lots of data. Alternately, much of the data could be sent later, at light speed, after the robots manufacture an appropriate receiving antenna. The data would contain many genetic sequences, plus everything the humans would need to inherit our civilization. Designing the parenting/teaching AI would be a very interesting job, as would selecting which human DNA sequences should be instantiated.
I know that this community is comfortable with the idea of Von Neumann probes. But I want everyone here to give some serious thought to the thesis that every place where a Von Neumann probe could reproduce, an island of Earthly life could also be built with the very same tools. It’s just a matter of bringing – or beaming – the extra data with blueprints and instructions.
Given that all you’ve come up with will be possible, the probe would also have to terraform the planet to make the environment Earth-like for the life to take hold. Even if that was also possible, it could take millennia, just as it would take to terraform Mars. It also seems like the ‘seed’ for such a system would also be quite large.
I think there is only one planetary surface where life will be nice enough for humans: Earth. Everywhere else, I suspect we’ll be better off in open space. With automated manufacturing, we can probably make sure that our habitat volume grows faster than our population. This is the key for assuring that inhabitants are never waiting to be comfortable. Rather then terraform imperfect planets, I’d prefer to grind them into building material for truly colossal spinning habitats where the light color, day length, gravity, temperature cycles and radiation protection are just right.
In the novel, the biology “bootup” I described happens in a module manufactured from the asteroid on which the seed probe first landed. The module is attached to the asteroid surface by a tether of basalt fiber, and the system is spinning fast enough to generate significant centrifugal force in the lab at the end of the tether. As the bio-bootup proceeds (more than a century before the first human baby), automated systems are relentlessly creating more habitable volume, energy generation capacity and chemical feedstock for life. Probes are dispatched to set up mining and material export operations on other bodies. All this takes time, but compared to interstellar travel, it’s quite fast.
One way you might be able to short cut some of that is with spore formation. Some eukaryotic, multicellular organisms can form ‘resting’ harsh condition survival spores. Look up ‘chlamydospore’. The organisms which form these could be reproduced easily enough in a tank of nutrients, forced into sporulation and then shipped long distance from star to star. Once revivified they have much of the cell machinery in place to start from to do some fast-track gene insertion for building up more complex organisms.
We are already kind of implicitly talking about building a dependency graph of much of our industrial infrastructure and then optimising it into a minimalist version that can sort of origami itself into the full setup in a deterministic way. Maybe if we did the same trick with the sum of our biosphere’s genomes? Beyond where we are now scientifically I think, but it doesn’t break the laws of physics.
All the best with the book!
The origami analogy is quite right, and I’m encouraged that even from my short description, you seem to get the core idea. I considered a plot where some very hardy biological stuff would be sent with the probe, but I didn’t want to stake everything on the thing’s staying viable through two millennia of cosmic ray bombardment in interstellar space with minimal shielding. Sure, bringing along samples of such spores makes sense – they’re light – but in the novel, the bio-unfolding will have to fall back on plan B: starting from non-biological scratch. That’s the story that I want to tell to fire up people’s imagination. We can boot perfectly human fans of Radiohead and South Park from the dead stuff of another solar system, a panspermia without biological “sperm”.
You might also want to have a read of Greg Bear’s ‘Hull Zero Three’ to see where he took a different but perhaps related idea.
“Where are they?”
Perhaps:
In orbit around our star. Connected to our internet. Watching our funny cat videos. Waiting for something.
I’ve speculated here before that maybe what it takes to find a tiny robot hidden in our star system is the ability to mount an exhaustive search, which implies the kind of space infrastructure we are talking about here. Which also happens to be the same technological point where we could create our own galaxy saturating robotic avalanche.
Bob Shaws’s “slow glass” or our own electronic “smart dust” technologies suggest that they could even hide on Earth to monitor us. The alien technology equivalent of Paul Davies’ invisible “shadow biosphere” that he has suggested searching for.
Sssh!
Don’t tip them off that we’re looking or they might run off ! :-)
The timeline does indeed seem very aggressive.
That something like this will indeed happen seems almost inevitable.
I do share the concerns of Neil S.
The aggressive timeline was one reason this article caught my eye. Growth is so much faster than our human economies have shown. However it is probably not that much of a leap from artisanal production to industrial production that transformed our economy from a slow growing, agriculture based, Malthusian limited one, to that which we enjoy today.
The beauty of the proposal is that we could experiment on Earth today to determine feasibility and iron out the bugs, before spending big bucks on hoisting the robots to the Moon.
Also, people love building robots for fun and there are any number of open sourcerors out there developing stuff. One thing not really emphasised much here is that our entire economy is gearing up to do lots of robotic automation in the near future anyway, and in a couple of decades time a lot of this stuff might be off-the-shelf kit, rather than apollo-program research budget territory. Including the AI.
We are seeing a lot of progress in robots and AI, that will probably accelerate in the near future. Our future looks bright.
However I doubt anything we get off the shelf will be suitable for space or the moon. Low gravity, high radiation and temperature extremes will mean a lot of new engineering will have to be done to make these work off earth. The possibility of long communication latencies will be another problem.
Eventual human settlement of space will probably look like this sequence, but I think it will take much longer then the paper predicts. It will be much harder than predicted, but we should get started as soon as possible and learn as we go.
You raise a very good point there about the conditions. You’re right, the hardware wouldn’t be off-the-shelf. Consider this though: while some protection from temperature extremes and ionising radiation will be necessary, machines need a lot less coddling than delicate humanity.
Maybe one set of heat/cold/rad proofed expensive machines could build a roof over a crater and cover it with rubble and then bring into it a bunch of environmentally stable crates containing more delicate machines and unbox them. The black body temperature of the (maybe still vacuum) inside of the crater would be stabilised at something tolerable for the more delicate machinery to work.
You could also probably get away with having multiple fairly small craters. There is no reason to design robots that are five foot six tall and suffer from claustrophobia in confined spaces! :)
In order for a factory to duplicate itself from raw materials, it must provide all the “embodied energy” represented in the finished items that make up the factory. Embodied energy is all the energy used in mining, processing, fabricating, and assembling a product.
Since the solar energy available away from the Earth’s shadow is ~7 times that at an average location on the ground, then the “doubling time” from an energy standpoint can, in theory, be 7 times shorter. The question is whether are other rate-limiting conditions besides energy.
Interesting concept. Interesting article. Interesting paper. I do not think it will be as simple as they try to make it sound.
The model needs more detail to be convincing. How did they come up with the masses and power usage for all the assets? Can some of these things really be totally automated on the Moon? Solar cells? I know of no one who is manufacturing them from raw materials to finished product on Earth. They would if they could, if only to save on labor costs.
As fun as 3D printers are, they are not the most energy efficient way to manufacture all these things. I have done 3D prints, CNC milling, casting forging, and many manufacturing processes except mining. That said, I have worked on tables, then spreadsheets, for mining the Moon for the past 30 years. My experience is still limited, but if I’m wrong about 3D printing not being the most energy efficient process for all these product, I would have to see the calculations.
The solid chemical plant operates with a feed-stock of lunar polar ice. This puts the whole works at the pole. So the solar panels will all have to be on the mountain peaks there. With a mass doubling every year, how long before the peaks are full? You could build a railway and power grid to more equatorial latitudes, but then you have the four-week day-night cycle to deal with.
AI is another topic altogether. How much intelligence is needed for these robots? Not much if they are just doing the same things over and over. Alex has comment on the subject before in these posts, and many things are possible in the future, near and far. However, I have heard many extravagant claims about AI for many decades. I am not a beginner at AI. I studied AI and robotics for my grad work in computer science. IBMs TrueNorth chip is very interesting, but expensive, mostly unavailable, and how do you program it? D-Wave’s quantum computer is also very interesting and already tested at 100,000,000 times as fast as a standard computer. Which technologies will be sent to the Moon? How long do we wait for better AIs before we send what we have and try to make it work there?
Still a lot of work to be done on this concept before it’s launched.
Best regards,
Micky Badgero
The authors certainly suggest much more detail is needed. These are broad brushstroke figures. They do try different values for robotic growth, processing speed etc. The 2 graphs I picked from the paper were ones I thought best illustrated the thrust of their idea, but also the uncertainties.
Also not that teleoperated robots will be the first generations, not fully autonomous AI driven ones. I suspect that much of their thinking was based on what they know of the direction of robotics on Earth for factory production. I could certainly see the experience gained in teleoperated surgery to be useful for lunar operations. My colleague, Brian McConnell is also an enthusiast for teleoperated robots on Mars, directed by a crew in orbit or on/in Phobos.
As we speak I am getting a TrueNorth system delivered to the university to answer the question of programming ease as well as usefulness. What attracts me is its very low power consumption that makes it suitable for local, mobile device, rather than cloud-based, AI. I believe that software simulations are used to train the proposed architecture, and then this programs the connections. So the approach is rather like programming FPGAs.
Not fair! I want one too. Seriously, congrats! Please let us know how the programming goes. But also, last I heard they ran about a million dollars for a full system. Maybe they are down in price now?
As to the robot growth, what they really need are example robots to show that the growth rates are realistic. My models show that eight tons delivered to the Moon could produce 877 tonnes of raw material with solar power very quickly (two weeks), but processed products would take much longer.
If we wanted to mine the moon the best bet is is a reusable mega-orion. Deliver 3-4 million tons of material to the moon on the outward trip, bring back 3-4 million tons of rare minerals on the return trip. A heck of a profit margin there, and a free moon-base to boot.
As for 3-D printing. If it simplifies things then a higher energy cost might be worth it.
3-4 million tons to the Moon? You might want to check the specs. There isn’t a rocket in the works that can deliver that much mass from the Earth to the Moon or back.
He’s talking about an Orion nuclear-pulse rocket – here: http://www.islandone.org/Propulsion/ProjectOrion.html – which could feasibly lift that amount of material. The only problem is that it involves setting off hundreds of nuclear bombs in the Earth’s atmosphere for every launch. In our wildest fantasies, we can imagine the world’s governments permitting that for a one-off launch – an all-in-one colony ship to Mars, for example. But the idea that you could use that for routine shipping is just not plausible, on public health grounds.
From the reference above, “Pedersen (55) says that 10,000-ton spaceships with 10,000-ton payloads are feasible.” Even with the nuclear pulse rocket, 3-4 million tons is a ridiculous exaggeration.
As the nuclear device gets bigger it gets more efficient, thermonuclear weapons can get very, very big.
Yes, but bigger isn’t better when it comes to nuclear weapons. They don’t have to get very large before they vaporize the rocket.
Given the time, money and lack of environmental concern, you could probably design a nuclear rocket that could deliver a payload in the 100,000 to 150,000 ton range to the Moon’s surface. The problem is of course, the time, money and lack of environmental concern.
And it totally misses the point of the article. The point is to send a seed to the Moon and let it grow own, not to send a Battlestar Galactica sized payload.
I think the concept of rapid expansion outward into the solar system is very interesting. The time frame considered is extremely unlikely. We’ve been trying to get humans to Mars to do basic science and exploration for over 40 years now. And I know we have had robots there for decades but humans make as many negative moves in space (the space shuttle for example) as they do positive ones. Look for some investment and progress in this area but not nearly at the rate considered in the paper. Especially given the very real problems we face on earth and have done little or nothing about.
We are making very good progress on Earth. Where we are behind is the robotic extraction of materials to make the raw materials for printing and fabbing. Having said that, a new liquid metal 3D printer has been developed that could make metal 3D printing far easier than the approach used today.
What I do think is needed to be examined is how to control the surface conditions on the Moon to make such production work. My guess is that enclosed, environmentally conditioned structures will need to be erected to make complex production work and avoid the problems of abrasive lunar dust wrecking the machinery. This may mean more mass sent to the moon to get the bootstrapping started.
I believe the right place to start bootstrapping industry is in a high orbit near the Moon, such as Earth-Moon Lagrange Point 2 (which is behind the Moon). The energy to get raw materials off the Moon is much less than the typical processes to turn raw materials into finished products. Since twice as much sunlight is available when you have no night to contend with, total production rate is higher in orbit.
A high orbit allows you to deliver raw materials from both Near Earth Asteroids and the Moon. The three major classes of NEAs have different compositions from each other, and the Moon, because they have different geochemical histories. Thus you also have a wider range of “ores” to feed your production processes. It also solves the Lunar dust problem.
This idea of travelling light , at less in the beginning , living off the land and boot-strapping a robotic economy seems to me the right one and even quite obvious. But you forgot two things which can slow down and even stop the progression.
First you must take into account death, disease, decline, wear and tear and the third law of thermodynamics .Process are not perfects and there are lot of wastes ,if you don’t care the solar system will be very fast full of dust fumes and screewed up machines and robots. it is not only an economy that needs to be built but a complete industrial ecology. That is not impossible but more complicated.
Then there is the little problem of profit. Capitalist want remuneration for their capital expenditure and I am not sure that they are very inclined to work for glory and to create a competitor system to escape their control.
Something like this concept is necessary for the large scale human settlement of the solar system. It will also enable various factions of humanity to go their own way.
I would like to try this kind of bootstrap process to build a seastead first. Call it baby steps.
People have always used tools to make more tools, as far back as the paleolithic. Nowadays, robots and machine tools are used to make more robots and machine tools, as well as every other kind of machine. The other machines, in turn, are used to make all the other products we use.
I’m working on a “starter set” of core machines that will fit in my basement and back yard, as a prototype to the seed factory approach to bootstrapping. A seastead would be a 3rd or 4th generation design. Unless some billionaire throws money at such a project, a home/hobbyist approach is a way to get started on the learning process.
That’s a good point. Good luck with it.
Planning a liquid mirror inside a very large moon crater foresees this problem of moon dust and particles having magnetic properties…we have to start the colonization of the moon somewhere…building a liquid mirror on the moon will teach us many things to further the reality of this post coming about…Finding lunar ice is where to start…solar heat and lunar water makes for a nice steam powered launch rail sending shuttles back into moon orbit…So it all starts with astronomy—as usual…
Hacking the AI systems which control robots to shut down the process or the civilization is even cheaper than changing orbits of large asteroids or sending relativistic rockets. The outside context problem has yet appeared, one should be careful about the expansion in all directions, a second chance doesn’t exist unlike video games.
The IT security angle is something that we would have to be really, really careful about. Not just an afterthought.
An automated lunar production system that can grow at a geometrical rate getting owned by script kiddies doesn’t bear thinking about. They could tell the machines to dismantle the entire moon and reassemble it in a banana shape or something for lolz.
All joking aside, if you spend a minute or two thinking about the possibilities of a massive industrial operation being run by what is essentially a software monoculture, you’ll see how very serious this issue could be.
Software security is always an afterthought, if its thought about at all. Software functionality is is not a priority until late in the hardware cycle, usually because it can’t be until the hardware design is fixed. So the safety has to be build into the hardware design.
Very interesting article. I strongly agree that an economic system which spans the Solar System and that is able to exploit the resources of the Moon, Asteroid Belt, and Mars for the benefit of all Humankind; only such an economic powerhouse could finance and support the emergence of our people as a spacefaring race.
It seems to me that if the population continues to grow, the sheer mass of human flesh to launch into space in order to simply stabilize the population and preserve the earth as a home for humans insures the failure of such a venture. A dismally reasonable alternative is to select superior breeding stock and establish a separate human population in space, leaving the remainder to live or die by their own efforts. Arthur C. Clark’s space elevator (From Fountains of Paradise) powered by solar energy might make evacuation of Earth’s gazillions possible, if combined with mass sterilization and Soylent Green.
Earth’s population will stabilize without mass emigration to space. More likely, a smaller number of emigrants will have higher birth rates that will populate the extraterrestrial population. Or possibly robots will become dominant and a relatively small space population will live isolated lives like Asimov’s Solarians (The Naked Sun).
So, why can’t this be done on Earth? If these robots are so capable, then why wouldn’t they turn the outback of Australia (or the Sahara, or Antarctica or… ) into an industrial zone? Wouldn’t that be likely to happen first?
My intuition tells me that robot factories on the Moon will still need resources (if only time and attention) from Earth, and so will never run too far ahead of our human needs for their products.
The article I imagine is way optimistic on the timeline, and about how fast we can develop the technology. But as for why it isn’t, or can’t, be done on earth first – the process is already happening.
http://fortune.com/2016/11/08/china-automation-jobs/
That’s just the start, and as the technology becomes tested the pace at which humans workers become redundant will increase.
Indeed. Autonomous vehicles are expected to eventually replace a 3-5% of the workforce that drives for a living (truck drivers, taxicabs, buses, etc.)
China’s Foxconn is saying it wants to replace all its 1 million + workers with robots. The reduced costs of robotic automation are likely to make robots ubiquitous eventually, especially once their costs fall below human wages. Just as cars obsoleted horses, robots will obsolete humans in a large number of productive activities.
I’ll reiterate my argument here. In order to colonize space, we need to invest in creating a space-based economy. One major step forward would be the use of launch loops/star trams, which would greatly reduce the cost of launching from the bottom of Earth’s gravity well. While this would require significant upfront investment, it would be justified by greatly increased efficiency and potential. Without this technology, we are stuck behind a formidable barrier to entry.
Once space is made accessible through launch loops et al, economics can proceed. I’ve always thought robots would be the workhorses of a space economy, as they’re ideally suited to the environment of an irradiated vacuum. I hadn’t considered the possibilities of self-replication — that may be a great boon to space economy, as long as we’re wary of the dangers of grey goo predicted by sci-fi.
Asteroids are promising for rare-earths and other minerals, and I believe helium-3 is abundant on the gas giants.
Once an economy is in place, there’s no need to worry about people’s willingness to live in space. There have always been people willing to travel out into frontiers, whether out of wanderlust, economic opportunity, escape from persecution, or other reasons.
I think this is like building an Atlantic bridge before colonizing the American continent. The whole point of the article is to avoid the massive upfront capital costs of space industrialization. It also avoids the difficulties and costs of supporting humans in space. Just as USAF drones are far cheaper than piloted aircraft (and safer for the human pilots), teleoperated robots make it much easier to apply human skills in near space. We have an abundance of labor on Earth that could operate robots within the low latency of communications volume of space.
The lower the capital costs, the lower the project risk and the higher the risk/reward ratio. This makes financing far easier. It also encourages competition as the barriers to entry are lower.
The time to build infrastructure to lower costs through volume is when the rewards and established and volume can be assured to amortize costs.
(pardon me for not getting back to this right away)
I agree that we don’t necessarily have to wait for large-scale infrastructure like launch loops to proceed with space economy. While chemical rockets are less than ideal, they should be enough to establish an initial economy, which would draw capital investment for better infrastructure, as you said. My point is that space economy will be constricted as long as it lacks better infrastructure.
Fascinating stuff. It is abundently clear that in order to reach the stars we NEED a solar system wide economy. Next step landing on Mars. Let’s get building.
Grand theories have never been the bottleneck ….on the other hand there are million simple experiments that never seems to get done ….experiments which almost any group of educated individuals could do with much patience and little money .
What would prevent anyone from organizing tele-operated mini-robots in an experiment here on earth to mine and produce any of the rawmaterials mentioned in step one ?
It has been almost obvious for atleast 25 years that tele-operation would be the cheapest way to kickstart industrial production on the moon , no matter what the next step should , or should not be . To tie the first babystep up with a grand robotics theory , is just another way to prevent the baby from getting anywhere
Which is what the authors state is needed. My guess is that teleoperation is the way to go, with increasing autonomy to shift the cognitive load and to stay economic. At some point full autonomy is going to be needed when the robots start processing the asteroid belt and other deep space bodies. The faster AGI arrives, the faster that can happen. If it doesn’t, then teh solution will have to be expending energy to move asteroids to cis-lunar space for teleoperation to continue.
AGI is not needed for automation. Automation is about precisely and repetitively executing a production protocol. There is no need for AGI, or AI. The complexity comes not from the intelligence needed by the robots executing the tasks, but the sheer number of processes that need to be developed to make it happen. The difference between teleoperation and automation is not great: Designing a teleoperated process is only marginally easier than designing a fully automated one. Some teleoperation may be useful to start out, in the debugging phase, but it is a nuisance that will be quickly and progressively eliminated by (AI-free) automation.
This isn’t really automation in the sense of an organized factory. It is about machines capable of doing a multitude of tasks which includes being able to largely self-replicate. This does require some AI, even if at the insect level. But certainly more than machine tool programming.
Teleoperation just shifts the intelligence to distant humans.
A multitude of tasks needs no more intelligence than a single task. There is indeed a huge number of tasks, but they are all tasks that consist in following a procedure, the stricter the better. No thinking is needed, or even desired. Some flexibility, yes, but intelligence? I don’t think so.
These machines are laborers, not engineers or scientists. We want to keep those latter functions for ourselves, and they are only required for design and initial construction, not so much for operations after the bugs have been worked out.
If we get ‘Starshot’ up and running we can not only send energy but also a heck of a lot of materials into space. The laser system even during build up to full power will allow craft to go into orbit for less than the current systems allow.
https://centauri-dreams.org/?p=9413
I have been working on the idea of “seed factories” mentioned in the article for several years, both for use on Earth and in space. Those readers who are interested can find the work so far at the following links. Since it is still “work in progress”, consider it like a version 0.5 of software, rather than a polished and published article:
* http://en.wikibooks.org/wiki/Seed_Factories (on the basic concept)
* http://en.wikibooks.org/wiki/Space_Transport_and_Engineering_Methods (in particular Part 4 through 4.2, though other parts of the book may be of interest)
* https://en.wikibooks.org/wiki/User:Danielravennest/papers/Mars21 ( applying the idea to space programs )
* https://en.wikibooks.org/wiki/User:Danielravennest/SFP ( the project I started to design and test seed factory prototypes)
A quick perusal suggests that you are on the same track as the paper regarding replicating factories. I note that you reference Reprap which was the machine I was thinking of that can replicate all but the electronics, and is the conceptual basis for the self-replicating robot. While life replicates at the cellular level, our current machine technology needs to replicate at the organismal level, and possibly higher. A robot with an integral 3D printer could conceivably replicate itself given suitable feedstocks and microelectronics. While plants can manufacture themselves with sunlight, CO2, H2O and various other readily available elements, robots will probably need to work within an ecosystem that produces the parts for self-assembly. Why print your own wires when another machine will specialize in that task for the robot population?
While 3D printers are useful, they do not provide the power to operate themselves, nor the feedstock in the form of filament/wire/powder to print from. For a hobbyist in a developed area on Earth you can purchase the power and feedstock as needed. In space you can’t, so I agree you need an ecosystem that can carry out the various processes. These include:
* Controlling the entire operation
* Supplying power
* Extracting raw materials
* Processing raw materials
* Fabricating parts
* Storing inventory
* Assembling finished elements
* Growing organics (if people are present, and in some cases where a biological process is more efficient, as in fermentation)
There was a summer workshop in about 1980 run by NASA where they looked into self reproducing lunar industrial plant with the technologies they saw at the time. They wrote a long report about it which is around the internet somewhere. Much of the report was about materials available in situ, and materials processing if I remember right.
Also, if my memory serves, they plumped for using plaster casts made from regolith to reproduce the components they need. They figured that the regolith was fine enough to get good detail on parts.
There is more than one way to do this; two very different approaches could give rise to a lot of in-between hybrid methods, with 1980 vintage plaster moulding at one end and full on reprap type things at the other.
On a related note, a couple of years back I went to a trade show with a lot of additive manufacturing stuff in it. One of the machines was basically squeezing out a paste of some sort of plaster mixed with volatile alcohol (which could potentially be recovered and reused) to form very large structures ready for firing into pottery.
Thought-provoking article, but way too optimistic. If it were that easy, robots accompanied be 3-D printers would be mining remote areas on Earth right now. Robots working in uncontrolled conditions require lots of human maintenance.
It will take many years and lots of money to mine the moon (a much better spot to start relative to Mars), but once started it will pay for itself many times over.
See University of Washington publications on Lunar and Asteroid mining I co-authored with my students.
The economics are very different between Earth and Moon. On earth, you can transport heavy mining equipment and their fuel at relatively low cost. You can hire cheap labor to run the machines, guard the output etc. The Moon is very expensive to transport to, which underlies the aim of replication. Why incur the difficulties of replication on Earth when you don’t need to? Before the industrial revolution, in the ancient world, slaves were often used for the mining. Replication for losses and expansion was just a matter of buying more from the traders as the empire acquired more. :(
Replication is hard, but self-expansion (using the equipment you have to make more equipment) is quite common. Examples range from woodworkers who make their own workbenches to machine tool and robot manufacturers who use their own products in their factory.
A seed factory, consisting of a starter set of core machines, plus stored plans for additional machines, is potentially a cheaper approach to building a large production capacity on Earth. It requires less up-front capital due to being smaller, and grow fast enough to represent a good return on the investment.
A starter set can’t make everything it needs to expand. So some supplies of parts and materials are needed at first. As the collection of equipment grows, you can reduce the outside supplies and make more items internally. If much of the production is automated from stored plans, the overall cost of expansion can be kept low enough to be attractive.
We might have a more efficient use of our planets natural resources first before we expand to interplanetary ones like more clean energy. Stating more than the obvious, I have to think we have to solve these problems by developing and Earthly infrastructure and new technology like fusion power which might solve our power problems in the future. There is no doubt though that reaching for the stars can still help us with technology.
Earth’s resources are inherently limited, especially with regards energy production. This implies that economic growth must slow down and even stop. If we want all Earth’s population to enjoy a standard of living that is available in the US and EU, this is not possible as far as we think we know. The only way to do that is to develop the global economy off-planet.
If we want a richer future for everyone, we will need to grow the per capita GDP without hitting Earth’s limits.
If we envision a future where large star ships are exploring the galaxy, the economy has to be much larger than Earth’s to support this endeavor.
The space-based robotic economy is one feasible way to accomplish this future.
There has always been a strain in Western thought concerning the dichotomy between growth and stasis. If you want the sort of future where human possibilities continue to be expanded, then continued economic growth will be necessary. Is there another way to achieve this that eschews growth?
We better hope a brighter future does not automatically mean indefinitely increased use of energy and other resources. The only reason we can believe this way is because of the very short time span of our industrial culture. Earth’s resources are indeed limited, so are those of the solar system, or the galaxy, or the observable universe.
For the constant growth enthusiast a 1% annual growth would seem very anemic; Yet simple mathematics shows that tiny number will in a few thousand yrs multiply our “needs” by a trillion. The real kicker is the next few thousand will do it again. Unless we come up with some transport unlimited by time and distance we cannot prevent running into these limits in what, even on the scale of known human existence, is a fairly short time.
If, as is speculated here, growth accelerates when we gain a foothold in space the wall there and then won’t be much farther than it is now on Earth.
For myself, I believe that “better” doesn’t automatically require “more”.
This in particular is MUCH easier than building a space industry. There is more than enough energy and raw materials, even before considering the very real possibility of raising the standard of living while simultaneously lowering per-capita energy and raw material consumption.
Not to say it is easy, far from it. Just: much easier.
You may want to reevaluate the connection you imply between the size of the economy and what is needed to build a starship. Presumably, by economy, you mean the industrial infrastructure supporting the population. When, as postulated here, we can exponentially boot up an space industrial infrastructure without people, why not just grow one that is 1000 times bigger than the “economy”, off in some corner of the solar system, for the sole purpose of building starships?
In other words, the assumption that we can only “afford”, say, 1% of our GDP for a starship has to be abandoned under the premise of self replicating machines. We can easily afford 1,000%, or 100,000%, as needed.
It is estimated that we need at least 2 Earths to allow the planet to live at Western standards. maybe we can reduce that with more aggressive recycling and better energy use. But clearly we are pushing limits. In addition, we are manifestly destroying natural habitat and species in what is called the 6th extinction. So expanding the economy on Earth much further without significant changes in output composition may be difficult.
That is an interesting suggestion. It is rather like an extrapolated case of the automation of the economy today and its impact. I suspect a lot of Earth’s population will want to know why they are not getting any benefits from this ET machine economy.
What benefits are there to get? With their own economy equally fully automated, any imaginable benefit can come from there. The starfleet is the benefit, a collective endeavor. Besides, there is no cost, so no harm done.
About “we are pushing the limits”, and “needing 2 Earths”, I don’t believe that for a second. We are in the process of reducing our use of fossil fuels, we can already see how we soon won’t need them anymore. Given plenty of energy, free from the sun, we can desalinate water, grow food hydroponically under artificial light, etc, etc. There is unlimited amounts of rock on Earth, just as good as space rocks, if not better.
In an earlier post I mentioned that I had worked on a lunar mining spreadsheet. The reason was to build starships from lunar material.
The plan was simple: send a small rocket to the moon; robots build a fusion-powered starship; it goes to Callisto where robots fuel it; it comes back to Earth for me. This was the plan I started working on in 1975.
In 2005, I stopped working on it.
However, I did look at payloads as small as a kilogram, so the moon-mining can be as simple or as complex as you want. The Moon has a good vacuum and low gravity. These can be used to advantage for some processing. The long day-night cycle is mostly a disadvantage, but when I started this, ice hadn’t been found at the poles, so I was aiming for the Apollo 11 landing site. There, the regolith composition is well known.
AI is not needed for self-reproducing robots. Bacteria, plants and most animals do fine without it. As others have mentioned in comments elsewhere here, this isn’t even real self-reproduction, its a few robots making a factory (designed by people) to make more equipment, some of which are other robots. For the most part, what is not teleoperation or automation will be new programming sent from Earth.
It is way too complicated for one person to accomplish in a lifetime, and I found no one else that was interested in it back then. Even today, there are just not that many people interested in real starships (warp drive, sure, and good luck to them).
With today’s technology, it is still very, very complex. How many people will take this up as a project? Some of that depends on funding. Paper designs can be done as a hobby in free time. Test equipment costs money. I never tested any of my solar designs because space solar cells can only be purchased in quantities of thousands of dollars. Can many people, separated geographically make and test small parts of this as a hobby? And only come together on the internet to compare notes? I don’t know. Its worth a try, but it will take a fairly large group to make this happen. The authors of the paper knew this and asked for help in the introduction: “A full study will be very complex and require the involvement of a much larger group of contributors. We hope this will raise interest and lead to that more comprehensive effort in the near future.” I hope this works.
Yes, exactly right, it will take a very large collaborative effort. I like to think it can be done like Wikipedia. Instead of articles, it contains processes. A process can be edited at multiple detail levels, from a one-page description to a well-documented maker project, with pictures as proof that the process has been implemented in the real world. Apart from processes, there are parts and materials, which serve as inputs and outputs of the processes. There will be algorithms that analyze closure, efficiency, and other parameters for unfinished self-replicators, and point out areas where work is needed. If successful, this would be the biggest (and rad’est) maker site of all. It could self-organize to eventually produce a completely closed set of fully automated industrial processes. They’ll be scattered across thousands of garages and backyards, but if the specifications are sufficiently strict, it should be possible to organize a “grand convention” in a desert somewhere to put everything together in one place.
I have been thinking a lot about how to set a site like this up, and I believe that, while it is not simple, it could probably be done by one person, or a small team.
The main challenge will be that everybody has something a little different in mind for the design, so it will take some thinking about how to encourage people to converge on a few common designs, say, one for the desert, one for the ocean, one for the moon, and one for deep space. Obviously, the only ones that could really receive a lot of practical work are the ones for Earth. People could set up demonstrations in their backyards (desert) or bathtubs (ocean). Space would be a lot more difficult, except for the theoretical detail levels.
Eniac, you should contact the authors of the paper and tell them about your idea. See what they think and if they can help. I would be willing to help. Looking at the posts, these are the other commenters I found here:
Abelard Lindsey
Alex Tolley
Cererean
Charlie
Dana Andrews
Dani Eder
david lewis
Gabriel Thelen
galacsi
Gary Wilson
Geoffrey Hillend
Harold Daughety
hiro
James Scanlon
James Stilwell
Jer
John
Larry Kennedy
Marshall Eubanks
Matt
Michael Fidler
Michael
Michael T
Neil S
ole burde
Oliver Milne
Patient Observer
Paul Gilster
Robert G
Steven Torry Rappolee
xcalibur
Zanstel
How many of you would be up for this?
My apologies if I missed anyone.
I don’t mind adding my penneth worth if a site is created :)
I would be happy to pitch in on such a project where I can. My calendar has a blotchy look: lots of time available sporadically, but with sometimes long periods that are wall-to-wall with other stuff. Something like either a wiki or a ticket system might work quite well with this.
Up to four from here, and the four authors of the paper. It was their idea and their call to action, so we can count them as in until we know otherwise.
What kind of timeline and overhead would the website you have planned take, Eniac?
The paper has email addresses for the authors on it.
Yes, that’s why I recommended that Eniac contact them. The web site idea is his. It would be a good idea to have a solid web site plan to present to them before asking them where they want to go next. You, Michael, Eniac and I like the authors’ ideas, but I have seen many small space projects go by the wayside because people couldn’t agree on how to get them done. If they already had a plan of how to go forward, they probably would have presented it. Give them one, and they can participate or present a better plan.
I haven’t really planned it to any extent. Just nibbled at it here and there. Made a couple of false starts. Between work and family I find myself with very little time, unfortunately. It would be fairly big project, I think.
You’d also have to think of a way to peer-review contributions, and keep score of competence and contributions. This in itself is an unsolved problem, still, with sites like reddit and Quora having made some modest progress.
Wikipedia works without that, but it is a lot easier to write a few paragraphs than it is to carry out a hardware project. Still, maybe Wikipedia could provide the infrastructure for this, with robot scripts that check to see if process descriptions adhere to the rules, e.g. list valid inputs and outputs etc.
Anyway, while I would love to take action on this, please do not wait for me. If anyone were to stand up and organize something, I would want to be involved, though, if only to contribute of ideas and advice. Paul has my permission to give my e-mail address, privately, to anyone with an active interest in this.
OK, so we have to bootstrap a website to bootstrap a hardware project to bootstrap the solar system economy. That’s a lot of bootstraps. Let’s try to start differently then. Instead of presenting the authors with a web site plan, let’s just send them out emails and see how much interest they have been getting from other people.
Paul, can you please send Michael, Robert G, and Eniac my email, and send me theirs (assuming all are in agreement)? I’ll send the authors an email, with our email addresses and cc those here who want to participate, and see what the level of interest has been. Then we can decide where to go from there.
In agreement.
I agree and will help where I can, it might be a good idea to contact a Hackerspace which could aid the construction of a tele-operated system.
Sure thing.
In very remote mines, access is a real and very expensive issue (no one can afford the permits let alone the roads). If the robot mining capability existed, a helicopter could drop off a robot miner, solar panels, 3-D printers, and then periodically pick up the finished products. Hasn’t happened, because the technology isn’t even close.
Self-driving trucks are already being tested by the mining companies. We already use tunneling machines and excavators instead of human muscle power. How long before a teleoperated robot is used to do work in hazardous conditions at the “rock face” rather than humans? Not long is my guess. As soon as reliability and cost make it economic.
The issue is not the mining robot, or the trucks. That is easy (and even that has not happened yet). What is difficult is to also have the factory that makes the robot, and the machines that build the factory, etc. etc. That is what we are talking about, and it is very hard. “Not even close” is an apt description, although I would argue we have the technology, what is missing is the design. Tens of thousands of processes (all existing technology, in principle) need to be miniaturized and implemented in full automation. The magnitude of the task is enormous, and what the minimal size of such a seed would be is pure speculation. 10 tons, 10,000 tons, a million tons? We can’t really know until we exhaustively list all the needed parts, and that requires a full design. To which we are not even close.
I think you are missing the point. You are assuming conventional industrial technology. The authors are saying that with self-replicating, “universal” robots, that type of industrialization is unnecessary. They clearly think that a small seed is all that is necessary. They may well be wrong, but I think the discussion should be about their proposal. I think they are somewhat optimistic about today’s (2012) capabilities, but I don’t think they are that far off either. Instead of using big machines, they are opting for small ones and the power of exponential growth.
It seems to me that it might well be worth designing such a robot/3D printer to refine ground up rocks for metals, print new parts and assemble another robot/printer adding just the microelectronics. It will probably take a number of iterations and the result might be quite unexpected. But this strikes me as possibly simpler than the Breakthrough Starshot design and validation as we already have many of the basic components already.
No, I am not talking conventional industrial technology. I am talking any industrial technology. There is no self replication if the “robot” cannot produce another robot. But there is not just one robot. The rocks need to be ground up, at the very least, which requires a mill. They need to be gathered, which requires a earthmover of some kind. Many, many other tools and machines are needed. A factory, in other words, or at least a large workshop. All this will require thousands of different parts, assembled in thousands of different combinations. When I said the task is enormous, I did not mean it could not be made small. I meant that someone has to design all of these parts, and all of these machines, and all of the ways they work together to achieve self replication. The enormous task is the intellectual effort to go into this, and I think it will dwarf most of mankind’s other achievements, so far.
Their approach about making a more primitive 1st generation and then build up the level of sophistication is a good one. It is also possible that a minimal seed could be quite small. The thing that would still be enormous is the software, the instructions that make it all work. Teleoperation does not help much here, you’d have to work out the procedures for human operators, too.
These are all good points that we should all keep in mind. But higher in the thread I thought we were considering unmanned mining and ore processing, not that plus self-replication. I am picturing something like an unattended tunnel cutter. It has systems for replacing its parts and placing orders so that its stock of spare parts never gets too depleted. These parts wouldn’t be made in place, though they could potentially be built in highly automated factories and delivered by unmanned vehicles.
I don’t want to suggest that this would be easy, but I do think we’re technologically ready to start thinking in detail about how to make it work. A part of the problem is that human miners in some parts of the world are criminally underpaid, which delays our sense of urgency about automating some of the procedures.
First thought is “why do it on the moon?” Fire up these little wonders on some desolate lands and let them loose. In 8-10 generations, its clean water, limitless food, cell phones, flying cars and mansions with robot butlers available at no charge for every man, woman and child.
Second thought, reminds one of “The Millennial Project” in its grandiosity:
https://en.wikipedia.org/wiki/The_Millennial_Project:_Colonizing_the_Galaxy_in_Eight_Easy_Steps
Third thought, its a macro version of nanobots that would rearrange atoms to fulfill our every material whim. IIRC, physics would not cooperate with such atom-scale engineering.
Fourth thought, as other commentators have noted, if it were so easy as suggested by the authors, then the galaxy would be awash in self-replicating machines. It would be hard to miss the onward expansion of swarms of machines chewing their way across the cosmos.
Final thought and not intending any disrespect to the authors, it seems little more than a flight of imagination combined with a spreadsheet. Until technology can even approach such self-replicating capabilities on earth (which, to me, seems decades or centuries away if even possible), a lunar/solar system scenario is just speculation. Fun to think about, true.
I think we should beware of the perfect preventing the good. Partial self-replication has been demonstrated. I have no doubt that the path to almost full replication (no microelectronics) is quite possible. The goal is to minimize launched mass to drive down costs.
For many of us, home and shop 3D printers will be a reality, capable of producing a huge range of parts, based on electronic design specs, that can be assembled into useful machines. Look to the maker movement to kick start this process and show how to proceed. Nasa is already trying 3D printing on the ISS in an attempt to reduce the mass of spares inventory.
Logically the approach should be to create high mass, commodity components that are simple to make by rolling and extrusion of readily extracted materials. Even simple production of domes of fused regolith dust would make sense to house machine works and possibly future human workers.
Robotics is advancing quickly, whilst human life support is on a slow development path by comparison. We are already at the point where robotics makes more sense than local human presence. I think it is just a failure of imagination to expect human presence will be needed for even near future space operations.
I think that the ISS would make a great platform to test robotics after its service date ends. Robots could be operated from Earth to manage operations and experiments without the need to manage more than minimal environmental support. You could test extensive 3D printing and repair, as well as building new machines to test mining and industrial processes in micro-g.
On thoughts one-four, I agree. Although they did say in the introduction that they left nanobots out of their paper for just such a reason.
On the final though, I agree, but at least they are using the spreadsheet to try to think quantitatively. They also note that they don’t expect self-replication to be that advanced when they start, and this makes sense. If you send critical parts such as solar cells, microcontrollers, and some of the missing and hard to produce lunar raw materials up with the initial machinery, it becomes much more doable. This also limits the growth so that it doesn’t get out of control.
The paper does state that “Another game-changer is the discovery of lunar polar ice providing vast quantities of hydrogen, nitrogen and carbon.” This is not correct. Hydrogen, silver and mercury have been found at the poles, but carbon (in the form of carbon dioxide or monoxide) and nitrogen (perhaps as ammonia) has only been deduced (due to the temperature), not found, as far as I know. So at least carbon should also be brought from Earth. Nitrogen is critical as a buffer gas for breathing and plant growth, but robots would not have much use for it. Mining processes that use nitrogen could be replaced with processes that do not.
CORRECTION: CO, CO2 and CN were found in the ice craters on the Moon. Most of the N was found as NH2, and this was listed as a contaminant from the hydrazine in the probe that was crashed to get the data.
the state sponsored social security sovereign wealth fund funded with carbon taxes is a grand social contract owning much of the world’s capital, is this macroeconomics?
Locally owned firms owned by ESOPS in partnership with families and coops and public share holders might be microeconomics whether it’s here or on Mars?
Since we’re talking economics any thoughts as to which if any of
our existing public companies are likely to lead the charge to
commercialize our solar system ??
Aerospace: BA/LMT/NOC /RTN/OA
Mining/Robotics: CAT/ROC/ITW/EMR/DOV/IR/PH
Eventual spin offs/IPOs of Bezos/Musk private space companies
I’d like to leave my descendants a modest investment in likely
candidate companies should this view of our future materialize ..
I dont know the names of the companies , but here are som thoughts about which areas could be cirtical for the first babysteps : If the first startup on the moon are smal scale mining by teleoperation , then the intensive present research i medical teleoperation becomes relevant . Add to that the acumulated experience with Timelagg gained from the mars-rovers , and of course the military experience with remote bomb disposal . The real breakthrough will be to combine all these areas with the knowledge from a major mining company , it could probably get started tomorrow if the money could be found .
I have had thoughts about launching whole rotating habitats in one go to just start the stellar economy.
If we had a torus and hub structure we could have the hydrogen tank in the main torus and the oxygen tank in the hub. We then have solid rocket boosters on the outside of the torus and liq O/H rocket engines under the torus. Now in the hub area we have air breathing ram/scram jets that burn liq H with the incoming air. The hydrogen not only fuels the jet engines but also cools the incoming air to densify it to improve burning, once we run out of air we can either start to burn the H with on-board oxygen as in a ‘reaction engines design’ or have the platform of air breathers fall away to self land and be used again. We could send many up of these torus’s to be joined up over time to form much larger structures or just have individual ones.
What do you think…
How exactly will such a structure survive the hypersonic regime it must travel through to get to orbit? You might want to look at proposed SSTO designs that have high volumes as a more realizable approach.
Although the dynamic forces will increase air breathers have the advantage of high ISP and with hydrogen it is very high, theory ~5500 at Mach 10, that’s 10x that of the best engines so far.
http://en.citizendium.org/images/thumb/f/f2/Specific_impulse.png/475px-Specific_impulse.png
I’m sure you are aware of the UK’s Reaction Engines’ “Skylon”. which uses the same approach. Separating the engines (rocket and air-breathing) was the proposal of the Star Raker proposal by Rockwell. More like your proposal was SASSTO (Saturn Applications SSTO.)
Reaction engines coolant and engine system idea has a lot to do with the design except hydrogen will be the coolant if possible. The nice thing about the design is using the profile of the torus as the oxygen intake for the engines. With the design we have a completed rotating structure ready to be rotated and kitted out.
Putting aside the vast complications of political, national-interest, and other such obstacles to pure science and industrialization (and thus colonization) in space (which we will slowly overcome), I am unfortunately unconvinced of the project’s path and destination – from my utterly unsubstantiated, uncited, and unsupported background/argument, of course, but:
(and I imagine most has been already said or hinted or considered)
1) The Moon seems a very remote, hostile, largely-mineral infertile, and highly-concession-filled place to access, establish, maintain, and grow an industrial infrastructure when compared to an asteroid-fragment population scattered and exploited within the appropriate earth orbit (even at the wait of 20+ years to drag those things in);
2) The value of hyper-cloning/self-replication of smart/autonomous robots that are nearly identical will likely never be as valuable as duty-specific robots with a limited range of abilities, complexities, and resources that are installed/launched at the right moment and coordinated from earth or sub-coordinated from 1-light minute from earth, etc., as needs warrant. I reason this from the idea that a generic AI or general purpose quantum computer will never be as valuable (and thus developed or produced sufficiently) to be as ubiquitous on front-line design or work, as custom-created/focussed (and thus task-restricted) AI/quantum computers – though one GenAI may be on every desk and back-pack in 10 years – they will not be in the cutting edge labs/ frontiers, especially if duty-specific ones can be launched/stored (i.e. control software complexity over hardware-replication complexity/ quality)
3) The value of hyper-cloning/self-replication of smart/autonomous robots that are nearly identical will likely never be as valuable/cheap in quantity as huge swarms of dumb-robots hyper-manufactured in their 10^x amount and rail-gunned into orbit near the problem (and eventually shot into their even more remote locations by the next tier of work yards)(i.e. better to shoot a million hammers into space to get that nail knocked in rather than develop a great hammer + arm combination through increasingly smarter replicated hammer-and-arm combos).
4)”… Once robots become sophisticated enough, …, they can leave the Moon and start industrializing the asteroid belt. ..” I am not really convinced on the water-pool concentric ripple spread notion (even somewhat target directed) for these autonomous robots as compared to ‘launched factories’ that pin-point prized strategic locations (asteroid belt objective, outer planer moon, convenient comet, etc). Strategic development and spread must follow some kind of plan, exploitive strategy, or vision that I believe is utterly beyond a smart robot – even with such a straight-forward program-plan as ‘find the closest ‘rich’ asteroid and dig. Just the ‘opportunity spread’ from one place looking out must be enormous and essentially unfathomable to a replicable AI smart-robot.
5)Not convinced that the longevity, self-repair, versatility, and robustness could be designed into the self-replication process faster than the need for huge number increase at every generation would warrant. The huge early and occasional die-offs would cripple and destabilize the coordination efforts of these smart robot swarms.
6) Not convinced that a smart robot replication ‘system’ could generically enhance itself so much in every generation and fulfill localized tasks, yet not fall into a specific functional ‘drift’ that would make it very unsuitable for the subsequent tasks in future generations (i.e. smart robot working on moderate gravity environment for generations asked to re-purpose itself to develop generation to ‘spread’ to a small gravity-free comet next).
I imagine many of these are likely to be overcome and strategies may already exist, but I suppose I am fundamentally skeptical of expecting a self-cloning/improving smart-robot swarm to blaze a liveable and workable trail for a delicate and arbitrary human culture. Thanks for the brilliant and comprehensive investigation into the possibilities – better a worked-out flawed plan than a hundred poorly-considered techno-fantasies.
I’m reminded of a story Arthur C Clarke told about organizing a 1954 symposium on spaceflight. He wrote to the chief of research at the US Weather Bureau suggesting he present a paper on meteorological satellites. He was rebuffed by the assertion they would be of little use. So he turned the tables and challenged him for details of why they would not be useful so that space cadets like Clarke would not waste their time in this domain. He came back converted with a paper that supported their usefulness.
– Arthur C Clarke. “First Harvest”. The Promise of Space. ch 10, pp 89-109
Extracting raw ore is the real easy part. Separating, processing and manufacturing is the hard part.
Use of Orion to colonize the solar system is a really good idea, but the Orion Spacecraft need to be built and based on the moon, not on Earth. There is uranium on the moon, it’s in the Copernicus crater.
And there are decent thorium deposits in Copernicus Crater as well.
I love solar and advocate its use at every turn, but the potential of fission (and _eventually_ fusion), makes me think that it would quickly replace solar in this enterprise.
We need a drastic shift in mental attitude to stabilize the Earth first. The philosophy of “using up” all non-renewables” because they “belong” to the human race is leading to disaster. Acidification of the oceans, accumulation of greenhouse gases leading to drastic climate change, the shift toward dangerously populist governments who proclaim outdated thinking. These are all extremely dangerous continuing trends. To build and colonize the solar system will take a massive global effort even with autonomous machines leading the way. Everyone should start now by trying to inform and educate the people around them. These are extremely dangerous times and progress will be even more difficult than previously unless we change the human view of the world.
However, history has shown that the best way to change is to provide economically attractive alternatives. Without that, there will always be resistance. Fortunately solar is now cheaper than other energy sources in sunny climates. Wind is cheapest in windy areas such as the US “tornado belt”. Fracked gas is cheaper than coal which is why coal is declining and will never recover despite miners’ hopes in W. Virginia. Unfortunately natural gas is not a good energy source, and leaking wells are a particular GHG issue. Until we have good alternatives for transport fuels, change is not going to happen quickly. This is especially problematic for aircraft, although renewably electrolyzed hydrogen (not reformed from methane) might be an answer for aircraft.
While Europe has cities that work for trains and buses, newer US cities are designed for cars, which makes it difficult to transition to more energy efficient [public] transport. This won’t change.
It may even be too late to stabilize the Earth’s climate, in which case the chaos will likely preclude these shiny futures. This is a highly charged subject and I won’t comment on it further.
Then there’s the issue of earth allowing others to have the high ground in large numbers. If there are only a small number of inhabited places or vehicles with the capability to give largish masses delta-Vs then it’s feasible to keep track of any man-caused threats by asteroids and such. When there are hundreds of places and vehicles, and vast industries, we’d only be able to hope no one thought it was worth it to bombard us. Will we allow that situation to develop or will earth’s governments stop it before it can start?
As other people I’m not conviced about the timeline. In any case, every space plan is too optimistic. I think that in this plan, the most difficult will be in the beginning. Too much unknown variables. Where to mine. How to extract and process. Unknown problems.
Later, outside of Near Earth area, the times to control anything grows very fast.
So I suspect that the real timeline could be five times slower.
But really another plans has the same problem, so it’s one of the best strategies in any case. Probably we could mix this plan if another more manned and have more visible results than wait than robots control the most important resources of the solar system.
We will need to avoid the use of scarce or very localized resources, as that will require a lot of movement around the solar system.. and that means time.
For example, less efficient electronics but that avoid certain very scarce metals could be better than require to travel a lot to mine that metals.
The part that i like the most is that this plan requires massive amount of engineering but not too much material resources. That could allow a lot of universities and schools to experiment if virtual reality simulations to propose new AI strategies and processes and winners could see the results in the real world.
Very inspirational.
As demonstrated by the 87 comments so far, there is a lot of interest in this subject. I really think it would take a whole book to do the subject justice.
Quote by Alex Tolley: “Earth’s resources are inherently limited, especially with regards energy production. This implies that economic growth must slow down and even stop. If we want all Earth’s population to enjoy a standard of living that is available in the US and EU, this is not possible as far as we think we know. The only way to do that is to develop the global economy off-planet.
If we want a richer future for everyone, we will need to grow the per capita GDP without hitting Earth’s limits. ”
The problem is that Earth has NOT reached its limits due to the fact that we have not used our resources in the most efficient way like completely changing to more clean and renewable energy technology. You are right though and it is due to the leaders of more than one country and certain special interest politics who virtually know nothing about science and scientific principles. It might be nice if they could do their politics “off planet,” since politics has nothing to do with an understanding of science. Green technology will not hurt the economy but create even more jobs. Training people to move from old technology such as coal that began use in 1866 in power plants won’t hurt the economy.
I am hopeful though that these leaders will change learn something about science.
So far , the only more-or less realistic buissnes model has been proposed by Shackleton energy company (SEC) .
Water is not just another raw material to be mined on the moon , it is the only one with an existing demand : Fuel for spacecraft , and IN space that is a very expensive raw material
SEC talks about a 15 billion investment , and their plan involves a startup crew of 7 people , but if the bootstrap principle of teleoperation from earth can be made to work , the initial investment might be much smaller .
So , here is a personal question : could you learn to drive your car in citytraffic with a 2 sec timelag ? Surprisingly , some people are capable of learning to do it , and it could be made a lot easier with the help of software similar to what carton- animaters use to extrapolate their images forward in time ..a split screen could show you both the real ( but 2 sec behind) situation and the extrapolated ”cartoon” situation …..for anyone wanting to try you can use your home digital TV recording device and any compatible camera to create a timelag ….but dont start with driving a car ! try tyieng your shoelaces first !
The SEC web page and proposal is here:
http://www.shackletonenergy.com/overview/#goingbacktothemoon
Dennis Wingo has a more radical, but far less costly proposal that uses robots first, with humans only later in the project, although is about putting humans to live and work on the moon. Missions 1 & 2 to pave the way with robotics costs less that $0.5bn.
https://denniswingo.wordpress.com/2015/04/30/a-singular-suggestion-toward-a-radical-idea-for-lunar-industrial-development/
Whatever the approach, in situ resource utilization (ISRU) is the key to reducing costs. Zubrin proposed it for his Mars Direct project decades ago. The logic remains. Musk used similar ideas for his showcased Mars transportation plan. Making lightweight, but capable machines solves the particular problem for transporting and sustaining humans in space. If you don’t need them, don’t use them. If those machines can largely replicate themselves locally, this saves even more costs.
We can easily test the feasibility and limits of 2 1/2 second latency for telerobotic construction on Earth. Surgery can be done up to 500ms latency, which is quite impressive given the fine motor control needed. The longer latencies for lunar operation will require both more “stop-and-go” operation, plus more autonomous robotics. Design will also play a part. Fittings that self guide are preferable to simple fittings that must align perfectly. Self-assembly might use forces like magnetism to help guide components to assemble correctly.
I’m encouraged that robotics has come a long way from the slow, behemoths of yesteryear. Today they are small, can work without creating internal maps, and can co-operate in swarms with minimal guidance. The DARPA challenges are getting better results, even if the failures are amusing.
So reducing mass, reducing costs for construction and transport will help pave the way for automated factories in space. Economics and a good business model will ultimately determine whether we see them or not.
” The longer latencies for lunar operation will require both more “stop-and-go” operation” ….not necesarrily , I think you underestimate the capability of the human brain to deal with unexpected strangeness . Think about the classical experiment where students wore an optical device that turned their wiew upside down , in 2 days everybody adapted perfectly . When the construction of the ISS started , many ‘experts’ kept explaining how completely unfit humans where for the tasks ahead , how they would probably brake down faced with the strangeness ahead …so far the only thing that hasn’t caused problems on the ISS is the crew . Properly selected and properly trained human operators can learn to deal with compleks teleoperation in a few days , while it will take decades before autonomous robots can do anything more than simple standardized jobs …and there is nothing simple about starting a mining operation on the moon
The literature seems to be focused on the issues of teleoperation for surgery. This obviously requires fine control. 0.5 seconds was the limit. I think it is reasonable to increase latency allowance for less fine motor movements.
I am not aware of any such experiments, but it makes sense to do some experiments here on Earth if there is none. This is similar to the Mars simulations that are used to determine how best to design tools and organize life for a stay on Mars. It is well within the scope for CD readers to do experiments and publish. Use inexpensive VR and software to delay visual feedback as a first task. Then do the same via a mechanical arms to prevent tactile and posture feedback. Almost any data on doing simple tasks – joining pipes, drilling holes, cutting material would help settle the latency issue and show the path forward on designs needed to overcome latency issues.
It sounds like this vision is based on humanity somehow developing self-replicating machines whilst making absolutely no improvements in launch costs. Which is very, very unrealistic. With modest infrastructure in place, using in-situ resources (refueling, proper spaceships that stay in space and are reusable) , and launch costs of $100/kg, you’re talking about costs of perhaps $100k to $200k to put someone on Luna and support them. How does the cost of developing a completely self-replicating (and more importantly self-repairing) machine compare against an almost self-replicating machine (automated production of various components) with the addition of a few people to fix them if anything goes wrong and do fiddly assembly tasks?
If costs go down to $10/kg – a stretch with chemical rockets, but perhaps beamed propulsion will get us there – then it becomes even cheaper to send humans for the fiddly parts, and also makes it cheap enough to send large numbers of colonists up.
We don’t need full self-replication. It’s far more likely we’ll see almost completely automated mines/refineries (regolith in one end, metals and chemicals out the other) and factories (metals in one end, parts out the other), but with maintenance and assembly done by humans on the ground, including assembly of new factories. The productivity per worker would be far higher than anything on Terra today, and we’ve got plenty of humans available for expansion.
Any time frame for those reduced launch costs? What about development of life support system for the crew? It would be great if the cost of access to space was drastically reduced, but what is the path to get us there and how long will it take?
Are you taking “selfreplication robot” literally?
It’s not. It’s more like “robot ecosystem” replication. Factories make robots. Robots make factories. A lot of different robots, like any machinery here, but mostly automated. And a lot of teleoperated robots.
So, no need to bring humans there.
Of course you can if you want. And if we control the environment well we could build a lot of colony goods from automated machinery so the costs goes down a lot.
The advantage of robots is that lower the costs of initial missions. There is not need to bring back and life support. but instead as we have do until now, we will bring a lot of similar robots there (so it reduced the desing a lot because most are the same) and to one only spot, so we can grow and reuse the pieces to extend the capabilities.
Until now, we have sent robot missions to explore. This time we will send robots to settle and create a colony. Only robot colony at first, but manned later as well when we could manufacture energy (solar panels and batteries), shielding (mostly bury inflatables), air, water and food.
With some strategical fuel depots or tethering across Earth-Moon space, to maintain that Moon base could be even less costly than ISS.
Another good point is that both strategies are not competitive but complementary. If you can bring Earth to LEO costs go down, better for the plan.
In any case, to bootstrap the infrastructure is a lot better that bring from Earth by brute force, no matter if you can even lower the costs by two orders of magnitude (very optimistic in any case)
Everyone interested in this subject and hasn’t already should check this: http://www.molecularassembler.com/KSRM.htm, a near complete collection of prior thought and work on this subject by Freitas and Merkle (before they got side-tracked into the much more fantastic realm of nanotechnology)
A few of my thoughts, variously already expressed in these comments:
1) The effort has to start on Earth. It is far easier (orders of magnitude) to design and build this on Earth than in space. Where you can intervene with a hammer or screwdriver if things don’t quite work a expected. It also has potential benefits that, arguably, exceed those of space colonization: Free, almost unlimited energy and manufactured goods.
2) It is not as easy as it sounds. The technology is mostly known, the problem is that you need to gather together and get to work as a single system a large part of ALL of humanities technology. The task is mind-boggling. However, it is no more mind-boggling than, say, writing Wikipedia, so I would think that a concerted open-source effort might, just might, be up to it. So would be a large corporation, perhaps, but it would have to be really large.
3) Size matters. A scale reduction of 10 leads to a mass reduction of 1000. However, known technology stops working if you miniaturize it too much. I expect that the optimum will be where most robots are toy sized. Some components might need to be disproportionately large, such as crucibles (which would cool too fast if small) and mining equipment.
4) 3d printers are optional. They are all the rage, but as a way to manufacture parts they have stiff competition in machining, casting, sheet metal bending, etc, etc. All of these methods can produce a huge variety of shaped parts. Most require much less complex equipment than 3d printing. Remember, you have to make the 3d printer, too.
5) AI is not needed. Once the system is designed and built, it will be wholly based on the mindless execution of repetitive manufacturing processes. The software could be very sophisticated, but there is no requirement for anything beyond what we can do now, except in scale. You have to encode and store the knowledge of nearly all industrial processes in existence, but you do not have to invent it on the fly.
Each of these machine tools requires, even more, machine tools until you need an ecosystem. Metal dies alone need other machines to make them. Everything is specialized and designed to allow high-throughput, lowest cost, manufacturing. 3D printing allows escaping from that approach. Just look at the masses of off-the-shelf tools you can buy at a hardware store. A hobbyists drill press alone, just to accurately drill holes, masses ~ 20 kg. Many people have quite a mass of tools in their garages, yet that is just a fraction of the mass required for factory automation which still needs local assembly by humans/robots.
I don’t see how 3d printers allow an escape. To the contrary, a 3d printer is much more complex than a machine tool. It takes no more than a well-equipped (with machine tools) machine shop to make a new machine tool, but it takes a whole ecosystem to make a 3d printer. A 3d printer can never make more than a few of its simplest parts. A machine shop can make pretty much all it’s parts. A 3d-printer also will weigh much more than a machine tool when measured relative to the mass rate of it’s output, which is the relevant measure.
For automation, one thing you’ll need is an electric motor. With a machine shop you can make that. With a 3d-printer it is much more tricky. Another thing you’ll need is nuts and bolts. Those are made by very specialized machines, although you can probably make them (less efficiently) in a machine shop, too. Making nuts and bolts in a 3d-printer will be even less efficient, if possible at all. Efficiency counts, because it makes the difference between a self-sustaining system being possible or not.
It is true that a 3d-printer can make more intricately shaped parts than a machine shop. But then, we do not need such parts. The more limited parts we had (and still use) before the advent of 3d-printing are quite sufficient to maintain a full industry.
If you are talking weight, the 3d-printer will not really help. The machines needed to gather, smelt and refine the raw materials are likely to be much heavier than the machine shop in which parts are formed. In fact, I think most of the “seed” will have to consist of mining and smelting equipment, or it’s parts. Part forming and assembly are lightweight in comparison. 3d-printers (or machine shops) are only good for part forming, nothing else.
Really? If so why isn’t my garage making extra tools for me? ;) Tool making requires agency, otherwise you end up with a lot of regress. This is where even basic AI appears. It doesn’t have to be human level, but the machine does have to react to the environment and make decisions. Even the simple robots at AMZN’s distribution warehouses use AI in the confined warehouse to handle simple tasks. Although it is pretty basic, to be sure, the programming is a lot more sophisticated than that for machine tools programmed by human operators. So we will have to disagree on this point.
Your garage does not have a full machine shop. I bet they get their parts at an auto parts store, if you are lucky. Or a chop shop, if you are not. A machine shop has lathes, drills and milling machines that suffice to make any complex metal parts given a small variety of blanks (wires, sheets, blocks, rods). These relatively simple machines are themselves made out of metal parts just like the ones they can produce. The exception is power (electric motors) and control electronics (for more sophisticated CNC versions). For power we could go back to the old times where all the machines were belted to a central shaft that could be turned by a water wheel, a donkey, or something more suitable in whatever environment we’re working in.
Tool making requires no more agency than making anything else to a specification. A machine shop does need “agency” , but it is not built for automation. Someone has to walk the parts around, and operate the machines. With agency such as: Do I walk to the right or left of the lathe when I bring this part over there? That would be done differently in an automated factory, in such a way it does not require walking or making decisions about which path to go. The secret to a smoothly running automatic factory is to keep everything very orderly and avoid the need for any decisions except at the highest level (What are we making today? How many resources do we need?). Logistics, in other words, which benefits from sophisticated algorithms, but is still far from agency. It is best to reserve all agency for the highest level of strategic decision making (What is our purpose, our goal?), and keep that firmly in human hands.
The problem I have with your analysis is that automated factories are massive, generally producing a very limited range of products, and inflexible. This is not the way to bootstrap with limited resources. Flexibility in printing requires printers not presses to produce varied printed output. Presses do it cheaper in volume, which is their purpose, but are inappropriate when you want small quantities of different types of documents. Similarly, 3D printers will provide a huge variability in output, albeit at a high cost per unit. When the Moon has a large economy, then setting up automated factories may be appropriate to output standardized components that are the stuff of our industrial economy. But not to get started.
To get started you do need agency, whether local AI or remote humans. Self-replicating, flexible agents are what allows you to avoid transporting large masses of equipment, settling for fewer tools, slower production rates but with higher cost per unit. But since these machines can replicate, momentum builds up. China was reputed to be able to build its Three Gorges Dam with huge amounts of labor, rather than relying on huge machines. Despite the availability of factory built homes, we still mainly stick build houses using labor. There will be automation involved, e.g. chemical processing plants. Common, standardized components will be worth trying to at least partially automate with tools rather than trying to print.
AI vs telerobotics
I personally think that telerobotics from Earth is going to be more difficult than the authors admit. Consider just trying to mate a male and female threaded pipe with a 2.5 second feedback latency. This can be solved in this case by adding a funnel to the female side so that precision alignment is not necessary. Otherwise it is going to be rather “stop and go”. These sorts of design changes will alleviate the problem, but clearly the more capable the robot, the less this is needed. Far better to train a robot to join 2 pipes once and just tell it join more pipes in the same way, than to do the task remotely each time. If we could tell a Mars rover to “drive to the indicated spot” rather than plan out its route carefully, a rover could traverse a lot more ground quickly.
How far we can provide intelligence to robots is unknown at this time, although I am in the camp that believes there is no fundamental reason why robots shouldn’t have human level intelligence and abilities in time. There is a lot of research on this capability and it is making real progress.
Robots to colonize the stars.
A human colony starship, even a world ship, cannot bring any more than tiny fraction of earth’s materiel to a target world. Humans will need to be artisans initially, using tools and local resources to build their colony and economy. Robots capable of doing the same tasks and replicating quickly offer a lot of leverage for humans to develop a colony. Whether controling such intelligent machines will be considered slavery and repugnant is a different topic.
Conclusion: automation vs agency
A thought experiment. You are asked to develop a local economy supporting many different products in a remote area. Transport of materials is extremely expensive. Do you start with a few people and their tools, or trying to manage transport costs to import a factory and supply chains to start production and expand from there? If starting with people makes most sense in this situation, then agency is what you want and the only issue is how to provide it.
Teleoperation makes sense when environmental conditions are hard (or worse) for humans. For example, building an oil wellhead in the deep ocean is probably better done with ROVs rather than humans in deep sea submersibles. As technology improves, the ROVs are likely to become the most economic option, much as drones are cheaper than piloted helicopters for a number of operations. Drones already use programming to avoid restricted areas and collisions, removing the cognitive load from the operator. I think that constitutes crude AI, by YMMV.
I agree. I will add that, although it seems already clear about this, that it’s better to be explicit.
It’s not about selfreplication robots. It’s about growing infrastructure with high levels of local resources.
It’s not robots but a ecosystem of machines (robots includes) and even humans could be part of this ecosystem, although not required and seems to be more appropiate to be at a later stage.
Also, as others have said, the moon is about the worst place to have this. Dust, night, and gravity are horrible nuisances when designing industrial processes for space. An asteroid would be perfect, or an Earth orbit facility supplied with raw materials from asteroidal outposts.
There are plenty of disagreements with this idea. We have almost no experience with industrial processes in micro-g. Most processes we have rely on gravity. Asteroids are dusty, and worse, the dust stays near the work and will not settle. There will be some advantages to be sure, but until we have some hard won experience, we are just guessing what they are are, which mitigates aginst early industrialization as advocated by the authors.
Yes, gravity is useful in many circumstances, but I think you’ll need to completely rethink most processes anyway. Where vacuum is free and water or air hard to obtain and contain, things will need to be done very differently all around. My guess is that gravity will be more of a nuisance than of help. You are right, of course, this is no more than speculation at this point.
I am not sure about the dust. If it does not settle, it will fly away, rather than stay near the work, or not? Anyway, most of the work (other than the gathering of raw material) could be done at a safe distance.
Moon surface is a nightmare. Moon subsurface is not. We don’t need a habitat so complex like a manned one, but a robotic colony could be as simple as a cave or buried hole covered with some artificial illumination that protects from high levels of dust, micrometeroids, radiation and fast heat changes.
Some machines will require to operate at surface, but another is not.
To build this primitive robotic shelter seems a good step.
We are already seeing the problems with automation and AI. We need to bring the rest of the planets population in to support something like this. Alex, here is a very good article, read it all the way through and makes a lot sense as to what is happening currently in the world. I just hope it may bring in a new paradigm!:)
https://qz.com/844487/meet-the…rom-destroying-the-world/
This is been an extremely ongoing and helpful discussion on the possibilities of the use of robots in space. However, based upon recent information robots might be the ONLY effective space travelers in a long long time.
In the latest Scientific American, which comes out next month, there is an article addressing the issue of Galactic Cosmic Rays. It states that the presence of these cosmic rays might be a considerable showstopper for any possible voyages within the solar system are outside of it. These cosmic rays are essentially heavy nuclei (think of iron atoms and even heavier atoms), which are moving at nearly the velocity of light. It has been shown that this radiation has had an extremely detrimental effect in experiments on the connection between neurons present within the brain; it is said such a detrimental effect that it is actually made the animals exposed to this radiation lose a considerable portion of their cognitive abilities to even have any type of reasoning.
This radiation would be present in any extensive travel within the solar system and would be able to effectively impair any astronauts engaged in a voyage of any particular length of time and the problem would only get worse if the spacecraft was moving at higher velocities than we have so far achieved, simply because of the extremely higher flux of these cosmic rays due to the higher speed of the spacecraft.
This effect is so detrimental that even the proposed ‘Spacecoach’ would be ineffective as a means of transportation. Such a proposal as Spacecoach would not result in protection for its occupants against cosmic radiation. Water molecules are simply too lightweight to have much impact as a shielding against me tremendously heavy nuclei that compose most of these cosmic rays, but it would still be effective against the solar flares that have been seen by our satellite. With regards to protection against the cosmic rays, there would be two possibilities that I can see: first off, we build ships of the massive sizes such as would have been seen in the movie ‘Alien’ which would have been vehicles in the millions of tons size Or we can use intense magnetic fields which might be in themselves be detrimental to crew members. It takes massive amounts of shielding in effect to allow reasonably low doses of these cosmic rays.
That’s where the idea of using robots, both as crew members and working as minors on say asteroidal bodies would be the most effective and sensible solutions. They could work effectively without regards to their loss of cognitive ability that biological organisms would definitely suffer if given exposure to these radiations. This does not in any way totally preclude the idea that men cannot make the voyages that we are talking about, rather that we may have to completely rethink how we are going to construct our vehicles and the length of time we’re going to permit people to engage in traveling between destinations and/or working in some kind of external outside activities, once they arrive at their destination.
As to the longer question about robotic development I would say that it is imperative that we move full speed ahead upon such a development not only for space travel. But for here upon earth as these machines will undoubtedly impact our lives in a greater and greater positive manner. The use of these machines in daily life can almost not be overstated. In any fashion and I would direct the readers of this column to the work in progress by a company called Boston Dynamics, who is engaged in development of many, many types of new robotic systems that are absolutely amazing in their own right. Just as an example, I submit a link below, showing the amazing action of just one of the latest machines that they have built to facilitate and help people. I hope you all enjoy this video because it’s extremely fascinating and interesting.
https://www.youtube.com/watch?v=rVlhMGQgDkY
While GCRs are mostly charged and could be deflected by magnetic of electric shields, mass is the safest approach for shielding. I think we addressed this in the post about the value of using asteroids and comets as convenient habitats during interplanetary flight. Short-term exposure would be limited to rendezvous and departure. Certainly, we still need to use energy to move them into the most convenient orbits, but at least they have the mass to use as propellant. Kim Stanley Robinson had these featured in his novel “2312“.
But obviously, robots are far more resistant to radiation as well as other environmental variables inimicable to humans. One problem for robots isung current chip architectures is that they can suffer from GCRs that damage the chips as they are not very fault tolerant. Animal brains can tolerate quite a lot of damage and still function.
There are a couple of problems with this alarmist view, as presented by @Charlie (I have not seen the underlying article):
1) Human beings have been in space for up to 400-some days, and as far as I have heard they are doing a lot better than most boxers and football players, in terms of retaining cognitive abilities. Galactic cosmic rays are at more than half strength in Earth orbit (a little less than half are shaded by the Earth), it is unlikely that double that will make all the difference. We’ve had people at full exposure, too, on the way to and from the moon.
2) Contrary to the claim, movement does not increase the exposure to cosmic rays unless you move at near light speed.
Here’s an idea for the metal component problem. Go old school.
Before people invented smelting they used to get their metal from meteorites. If you’ve ever seen a nickel-iron meteorite chopped in two, the inside of them is shiny. The outside surface is oxidised by its passage through Earth’s atmosphere, but the inside is shiny, reduced metal.
On Earth it is quite difficult to find meteorites in most places because they tend to blend in quite well. Apart from in deserts or Antarctica. How hard would it be to find them on the moon through? Start from the idea that you are searching for a piece of metal. Build a satellite with a metal detector in it. Then put it into a really low orbit around the moon and let it build up a survey. For avoidance of doubt it could snap a picture of the area whenever it found something.
Armed with this survey, find an area with lots of appropriately sized meteorites and send the team of bootstrapping robots there.
Given lumps of metal, we could use an induction furnace to create molten metal for moulding. Alternatively and perhaps more simpley, we could simply grind the metal up with abrasive wheels to form a powder suitable for laser sintering. 3d printing of metal is almost 100% efficient materials-wise (unused metal powder can be trivially recycled). A reasonably powerful laser (~100W or so) is needed. You don’t need to move the laser though, just direct the beam with moving mirrors. You do need a moving powder spreading arm though.
Perhaps a laser metal sintering rig could also be used with powdered moonrock to print insulating ceramics.
Also consider that meteorite iron is a mix of Fe, Ni and Co, all ferrous metals. It seems reasonable to me that this would be magnetisable and this could be easily confirmed on Earth.
While meteorite iron is going to be in limited supply, it could provide a helpful resource in the early stages of a lunar robotic bootstrap.
When the hydrogen economy idea was hot, this was a similar idea that astronaut Harrison Schmitt proposed for platinum mining on the moon Return to the Moon: Exploration, Enterprise, and Energy in the Human Settlement of Space. Find impact sites with the increased presence of platinum group metals in the ejecta and mine at that spot. The logic for ET mining for platinum at the time was that 2000’s technology fuel cells needed more than was available on Earth, for ET sources would be needed. While that is no longer true, I think the logic for locating richer sources of minerals from catalogued impact sites still holds.
@Charlie – Heavy ions (which make up most of the Galactic Cosmic Rays – GCR – _by energy_, not by number, which I think is the cause of some confusion) are included in the various NASA studies on GCR mitigation, and can be shielded by (enough) water.
NASA studies cosmic radiation to protect high-altitude travelers
“NASA scientists studying high-altitude radiation recently published new results on the effects of cosmic radiation in our atmosphere. Their research will help improve real-time radiation monitoring for aviation industry crew and passengers working in potentially higher radiation environments.
Imagine you’re sitting on an airplane. Cruising through the stratosphere at 36,000 feet, you’re well above the clouds and birds, and indeed, much of the atmosphere. But, despite its looks, this region is far from empty.
Just above you, high-energy particles, called cosmic rays, are zooming in from outer space. These speedy particles crash wildly into molecules in the atmosphere, causing a chain reaction of particle decays. While we are largely protected from this radiation on the ground, up in the thin atmosphere of the stratosphere, these particles can affect humans and electronics alike.”
https://phys.org/news/2017-01-nasa-cosmic-high-altitude.html
The problems concerning Tele-operation of robotic systems with delay (on the moon or elsewhere) , are of a very practical and near-term nature . …the next babystep…thats why , at first , I was surprised to find very little relevant material on the net , but eventualy stubbornness payed of : the right sequence of google-words goes like this : Time Delay Teleoperation research . The best article seems to be the one that pops up first , from the chinese journal of aeronautics …enough to cheew on for a few days
I don’t see that reference when I try it. Can you provide the link?
The tele-operation only need stop when the equipment (local) senses something out of normal, so most of the time it works very well. When it stops then a human remote can interfere. So a local AI would do very well most of the time and the remote HI corrects the potential error slightly later on, this would work quite well.
I definitely remember experiments done on this in those dark ages before internet. Loops were set up with the expected lag and the experiment was reported to be completely successful. Sorry I don’t remember more details right now.
Link to the Chinese article :
http://www.sciencedirect.com/science/article/pii/S1000936108601065?np=y&npKey=ebd38ba250ed4d39bae1a6d46e
Thank you. What I got from this paper:
1. Latency of 5-7seconds can be handled by using predictive modeling of the robot and the manipulated devices. (This is rather similar to a proposal for managing rovers on Mars, but obviously with more latency on Mars.)
2. The type of operation is not fine motor control – in this case unfolding a solar panel.
3. Gross errors could be handled by slowing down the robotic movements to ensure the predictions remained accurate.
4. VR simulation is used for the controller to interact with the system.
This does indeed suggest that teleoperation on the moon with 2 1/2 latency should be possible for a range of activities.
I would like to add the ability to view the actual scene on demand too. Unlike a satellite or space station, the Moon will not be fully defined and therefore the actual situation may differ from the VR simulation.