Lets say I want to send a generation ship to colonize a distant planet. Because of the ships mass it's going to take a long time to get up to speed. I can send smaller probes on ahead since they are smaller and can accelerate quicker. Some of these have machines to start terraforming the planet. Others have machines to mine resources. The last type of probe is a city seed, so that when the colonists arrive they'll have a place to live.

The probe is too small to hold a bunch of pre-fabricated buildings. Instead I'd like to grow a city out of what is available on the planet: crystal, or biological, or use your imagination.

You can assume some level of nanotech, the city seed can use materials brought to it by the other probes, and handwavium is ok within reason.

Edit: Also, the colonists won't be getting there for a while, so long term projects are ok!

Edit 2: We don't know much about the planet since it's an exo-planet several light years away. We can tell it's roughly earth sized, maybe slightly smaller. It's in the Goldilocks zone. Spectroscopy shows the atmosphere has some oxygen, though the mix isn't quite right for humans, yet, and there is some water. Spectroscopy does not show any indicators for life.

  • $\begingroup$ So to clarify: you want to know how such a probe (which has to be small) could grow a city like this? Does the city have to be metal buildings like we know, or could it be underground caverns? $\endgroup$ Commented Apr 10, 2015 at 20:07
  • $\begingroup$ @TommyMyron The city can be made out of anything. Just looking for ideas. If you have one, make a case for it! There isn't a timeframe, as the colonists will be inroute for a long time. $\endgroup$
    – AndyD273
    Commented Apr 10, 2015 at 20:20
  • $\begingroup$ Please note that the text of your question do not actually contain a question mark. But I gather that you want to ask "how do the city-building machines look?", right? $\endgroup$
    – Abulafia
    Commented Apr 13, 2015 at 12:17
  • $\begingroup$ @Abulafia How they look, how they work, what they use for materials. If you were to design something that could land on a planet and grow a city that people could live in when they arrive in 100 years, how would you do it? I'm trying to keep it vague, because I want to see what other people come up with. I have an idea how I would do it, and I'll probably post my own answer at some point, but I am really interested in what you (and others) think would work. $\endgroup$
    – AndyD273
    Commented Apr 13, 2015 at 14:13

7 Answers 7


Do you know about Von Neumann Probes? It's a concept of an unmanned spacecraft that can build a copy of itself from planetary resources, and launch more of "itself" towards other planets.

Well, let's modify its software a bit.

First, it needs to survey the land for metal deposits during the orbital fly-bys. It can be done through numerous nanoprobes or just detecting magnetic anomalies of the planet's magnetic field, or many other methods of surveying planets.

After landing, it deploys the small "factory" and a few mining robots that start gathering resources - sand for silicon, iron for structural components. It contains some of less common (but needed in lesser amounts) materials and some more difficult to manufacture components (like electronics) as well.

Its first order of work is creating more surveyor/miner robots and enough solar batteries to keep the plant and the robots operating. The hulls and mechanics of the robots is manufactured from the local metal; the electronics is provided from the supplies.

Once more robots are built, they scatter to search for more scarce resources. Fetching their finds they enable the factory to produce more electronics - more, bigger robots. The bigger robots build a bigger factory that creates prefabricates of the city, more, bigger, more efficient mining robots, automated trucks to bring a plenty of more scarce minerals from afar, and finally construction robots that build the city with all the needed infrastructure.

One more thing: instead of sending more probes that would take up terraforming and other tasks, send just this one early enough. It will be capable of building the terraforming reactors too.

  • $\begingroup$ That's the answer I envisioned too. Upvoted. $\endgroup$ Commented Apr 10, 2015 at 22:36
  • $\begingroup$ If the probe contains "complete" robots ready for digging and manufacturing, it would be larger and would have issues coming to a stop at the destination. I'm thinking nanotechnology would look more like vegetables and fungi, not like steamshovels. "mining" is done, at least initially, with vines and hyphae. Von Neumann and contemporaries that tried to envision the tech predated Feynman (cf 1959) and Drexler, even the discovery of DNA coding! $\endgroup$
    – JDługosz
    Commented Apr 11, 2015 at 0:04
  • $\begingroup$ @JDługosz: I haven't thought of that, but it's a viable idea. Still, it enters area we believe possible but don't really know how to approach. Standard mechanical robots are possible today, the budget is the only barrier. Above all, DNA coding is an extremely efficient, lossy fractal compression algorithm and while it is possible to decode the genome and determine effects of separate genes, creating organisms this different from common biology and this specialized (and capable of survival on a non-terraformed planet!) is probably an NP-hard problem. $\endgroup$
    – SF.
    Commented Apr 11, 2015 at 12:02
  • $\begingroup$ I don't understand what you mean by DNA is a lossy fractal compression algorithm. Advanced nanotechnology would not simply be taylored life, but it will appear more like livimg things we know than like machines we have today. $\endgroup$
    – JDługosz
    Commented Apr 11, 2015 at 17:01
  • $\begingroup$ @JDługosz: You have about 2 gigabytes of data (DNA) that "decompress" to 37 trillion cells of human body, often highly specialized and distinct, although similar to neighbor cells (forming various tissues) - they all contain the same DNA and they aren't "aware" of their general location in the body, but by general "neighborhood" they "know" if they should develop into bone, muscle, taste bud or a blood cell - which parts of the genome are to be used in duplication of the cell: it's a fractal where different parts of the genome are used as cells split and differentiate. $\endgroup$
    – SF.
    Commented Apr 11, 2015 at 22:39

I think the easiest would be some kind of a little digging robot to excavate a city from pretty much anywhere it lands. It needs geological survey tools to locate the optimal spot (whatever terrain feature you set it to find - like water, AND a ground component that it can work with , like a bedrock near the surface), then it gets to work, digging out the materials it does not want and then digging out buildings and infrastructures. Even some of the plumbing/waterways can be done "dug".

Depending on how much time you have, sending more of these can speed up the process.

Dig dig dig. When you get there you have shelter and infrastructures and use your tech brought by the spaceship to add whatever else you need (wiring/electricity - I'm supposing by then any com is done wireless)


If terra-forming began recently, you can't use trees to build - they are priceless. Your best building material is stone.

You need water for industrial processes.

You need some rare materials, and sending out probes all over the planet for mining surveys is expensive. Your best bet is concentrating dissolved minerals from sea water. So select a site where you can quarry good stone that is by the sea. Use carbon nano-fiber-based mesh to filter water. This gives you pure water to put in a reservoir and you extract the dissolved minerals for use in manufacturing electronics or making fertilizer.

Energy is tough. Assume we have developed compact fusion power by then. The bottom of our oceans has lots of methane clathrate. Maybe the ocean floor of this world has something useful for fuel?


You write

We don't know much about the planet since it's an exo-planet several light years away.

I find that contradicts the main premise.

With the same technology, you could send seeds to distant asteroids and grow huge multi-mirror telescopes. Besides being limitless in size (no cost! Just let them grow limited by solar energy input) the signals from distant telecopes can be combined: distant in this case means Neptune orbit, or billions of miles.

You would be able to see the exoplanet quite clearly, mapping the continants and getting detailed spectra of the planet as a whole and of individual features.

It will not be an utter mystery. We would know the chemestry of the plant life and the oceans, and will have watched the passing of the seasons, before even planning the colonization meeting.

  • $\begingroup$ I think you misunderstood the question. The OP is asking for robots that will build cities on the new planet, so that when the humans arrive they have something to live in. $\endgroup$ Commented Apr 10, 2015 at 22:34
  • $\begingroup$ I was addressing Edit 2. The planning stage. $\endgroup$
    – JDługosz
    Commented Apr 10, 2015 at 22:50
  • $\begingroup$ This is a fair point. I guess I was trying to keep it a bit vague so as not to limit the ideas. It could be mostly water, it could be mostly desert. I haven't really decided, thought I may do something like this later. Thanks for the idea $\endgroup$
    – AndyD273
    Commented Apr 10, 2015 at 23:33
  • $\begingroup$ Don't forget too that you'll have the entire system to use, not just what's on one planet. You could bring material from comets and asteroids. Or, in the case of this novel, making the body to terraform. $\endgroup$
    – JDługosz
    Commented Apr 10, 2015 at 23:51

Go Underground

Not knowing much about the planet, my preference would be to start civilization entirely underground. I'm not sure if this is what @spacemonkey was implying.

The entire city should be underground, and with the materials you've removed, features can be installed.

Atmosphere/airlock - You have a potentially lethal environment on the surface, you can easily airlock your city.

Protection - Unlike dome-cities, 10m of hard rock offers a lot better protection from any small asteroids: you don't know what might be raining down on this distant planet. Also, I'm not sure what the affects from radiation are at the surface, so this provides a modicum of cheap protection.

Technology - you must have developed plants that do not require sunlight?

"Sun"light can provide energy to surface collection, and anybody sick of no sunlight (most will have lived their lives in darkness or artificial light, no?) can hop up to the surface, protected, to get their tan on.

  • $\begingroup$ Good point about the artificial light thing. You aren't going to get natural sunlight in deep space between stars, so they would have ways to deal with that until teraforming has progress far enough to allow them to live on the surface $\endgroup$
    – AndyD273
    Commented Apr 13, 2015 at 15:04
  • $\begingroup$ @AndyD273 I suspect, since several generations have gone by living in artificial light, the arriving settlers might not even desire sunlight. $\endgroup$
    – Mikey
    Commented Apr 13, 2015 at 19:22

That kind of nanotechnology (Von Neumann Probe style) is beyond any limits of the kind you allude to. It can remake a whole biosphere, change the chemistry of the atmosphere, and (eventually) not only change the oceans but change the presence of oceans.

Actual limiting factors:

  • atoms of the right types readily available. C, H, O, N, P should be no problem. Even without metals you would rather grow buildings out of organic molecules anyway. Small quantities of various elements are crucial for catalysts and special organo-metal complexes that do interesting things.

  • energy. The natural biosphere uses solar power. Advanced power could be used in local areas for specific industry. Fusion would require a lot of infrastructure before being possible.

  • executive intelligence and memory. Storage might be tight in the "seed" probes, but that could be supplemented with transmissions. After setting up a receiving station and growing lots of memory, it can download extensive plans that follow via laser.

Note that without nanotechnology, detailed plans for self-reproducing probes have been analyzed in detail and published (and this one) (etc.). Nanotechnology changes nearly everything, but you can get an idea of the kinds of planning needed.

Some miscellaneous thoughts: the seeds would be traveling quite fast, so how do they stop? Final landing can use aerobreaking and a large craft break up in to individual small seeds. But some executive control, archiving, and even heavy industry could be on an airless moon. If that's the first stop how do you stop? If the atmosphere is the first stop, how do tiny dumb seeds boot up enough intelligence to plan enough industrial complex to build rockets?

Tiny seeds will be dumb. They can start mining and collecting materials through vine-like growth, but what about planning? Maybe a larger seed unit would be necessary, with very dense and durable memory on board.

Basically the more bootstrapping, the more difficult. Keeping a functioning command unit in orbit from day 1 and sending down seeds with directed landing sites is far easier.

Such a "large" probe is still smaller than a colony ship, and not limited to acceleration levels. Slowing down can use a magnetic parachute, but is that enough? Retro-rocket reaction mass can double as shielding during transit.

Now if technology can do that, why send bodies? The original slow-boat colonists might arrive to find the world inhabited by those who beamed themselves over, or by intelligent machines who built the place.

You could probably have a whole series of stories about different seed destinations and how things went wrong on each one!

  • $\begingroup$ Very interesting! I like these ideas. I hadn't considered a seed pod, but if it has a way to reload without landing on the surface there's not much reason for it to stop... $\endgroup$
    – AndyD273
    Commented Apr 11, 2015 at 1:38
  • $\begingroup$ Reload from what? $\endgroup$
    – JDługosz
    Commented Apr 11, 2015 at 3:55
  • $\begingroup$ It's just an idea. The pod takes a cometary orbit, and seeds asteroids on the way through. These would grow into seed factories. Then on the next orbit it could pick them up. Getting in and out of a gravity well is tricky unless you can keep a lot of momentum, so if it can seed the planet at a distance and avoid losing momentum, then that would be an advantage. Get the colony started, and head toward the next star. $\endgroup$
    – AndyD273
    Commented Apr 11, 2015 at 5:33
  • $\begingroup$ Ah, so it has multiple systems worth of seeds on board and only drops some of them. What's onnthe vehicle other than seeds? But with that technology, they are not expensive! So just send them individually tomeach star. If the mothercraft did not slow down, the seeds woud have to, and what's the purpose of the mothership? $\endgroup$
    – JDługosz
    Commented Apr 11, 2015 at 5:59
  • $\begingroup$ The seeds are cheap, but small. No reaction mass, very basic processing power, relatively limited memory. They can carry a few complex instructions, but things like deciding where to land would be to difficult. The seed pod has something like an AI aboard, plus it's got a few other tricks. it decides where the seeds go, and as the seeds do their thing they increase their memory capacity until there is enough room for the seedpod to beam down the main payload of data, maybe a copy of the AI. At that point the seedpod doesn't need to stick around. $\endgroup$
    – AndyD273
    Commented Apr 11, 2015 at 6:07

Use the idea of a replicating virus. On a computer, a replicating virus creates copies of itself whenever you try to dispose of it. Think about using this same method. The probe goes to the planet; replicates itself, possibly out of the minerals from other probes; connects together with other probes like an atom; and finally does the replicating until a substantial building is built. After one building is built, it can replicate the building until there are no more minerals.

Furthermore, if this is a magical world, you could alternatively use seeds that grows cities.

How you want to go upon tech is up to you, but the Von Neumann Probes concept may work with that.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .