Self replicating robots do not exist

Or, there is no shortcut to industrial transcendence.

Part of my series on countering popular misconceptions in space journalism. This blog expands point three of Unpopular Opinions in Space.

Image result for self-replicating machine

In previous posts I’ve highlighted the difficulty of profitably selling raw materials in space, whether mined ore, water, or beamed solar power. Without such bulk beachhead products it is difficult to imagine the large scale investment needed to gradually move industry into space.

Part of the reason space is so difficult and expensive is that it requires a lot of technically sophisticated hardware to keep humans alive. There are few inhabited environments more adversarial, and the requisite technology is expensive.

For many decades it’s been recognized that refueling spacecraft at the destination is a way to potentially reduce cost and complexity. What if we could also produce some of the machines in space?

Unfortunately this thought experiment quickly goes awry when the limiting factor, labor, is brought into the picture. Mature space tech requires about a thousand sub-specialties to execute, representing a workforce at least this large, and increasing life support overhead by the same factor.

Once again, a scrappy bootstrapping approach has run into trouble. No matter, the story goes, we can replace nearly all human labor with robots. In fact, we’ll have the robots assemble themselves and each other, ultimately from raw materials.

This is an attractively simple idea, but it soon runs into fairly severe issues, all of which are summarized by the fact that: self-replicating robots don’t exist.

Self replicating robots are also called universal constructors, matter compilers, generic 3D printers, robotic eggs, and so on, depending on where the difficulty is magically handwaved away. As an idea they’re very attractive, since we’d only have to build one rudimentary one to get started, and it would take it from there.

Once we have a bunch, each relatively compact, we only have to launch one at each asteroid with a set of plans, wait a while, then move into a fully converted luxury space hotel.

Additionally, the existence of biology shows us that mechanical self-replicating machinery is physically possible. Bacteria reproduce themselves and yeasts make our alcohol with very simple raw materials in a warm, wet environment.

Yet it’s a huge leap from genetically-tweaked bacteria producing insulin for diabetics to generic industrial products, such as lumps of titanium or integrated circuits. It’s also pretty clear that even heavily modified bacteria would struggle to be much use in the vacuum of space. While biology is inspirational, deep space self replicating robots will have to be mechanical in nature.

Fundamentally there’s a deeper reason that it turns out to be really hard to keep a city in space alive with very minimal supplies of materials and people. While some sectors of society venerate the self sufficient and practical know how, there is a limit to how harsh an environment humans can survive alone, without cooperation.

Heavy industry needs specialists, millions of specialists, and that makes it big. That is why so few countries, all of them rich and highly populated, have comprehensive industrial stacks capable of producing anything from a cell phone to a fighter jet. Without a diverse industrial sector crammed with specialists it is not possible to mass produce complex machinery and propagate domain knowledge. Without mass production, parts cannot be replaced more rapidly than they break down, resulting in collapse.

The balance of human and robotic labor is a fascinating rich topic, beyond the scope of this blog. For more detail, check out my blog, talk, and book on the topic. In brief, improvements in technology are leading to a gradual shift toward forms of mechanical and computerized labor, and away from human muscles and brains. Transporting robots and humans to Mars is so fabulously expensive that it is easy to overlook the fact that humans will be proportionately more expensive still, and so in-space work will favor automation much more than on Earth. Still, there is a big difference between a cybernetic human-machine hybrid (essentially the status quo) and a 100% automated, self-reproducing industrial stack, let alone a miniaturized universal constructor robot.

The bottom line is that, cool as it might be, there is no nifty shortcut to industrial self sufficiency in space. Building space hardware in space will require thousands or millions of people. Our challenge is to determine the best way to do this with the best tools we currently have.

2 thoughts on “Self replicating robots do not exist

  1. Doesn’t this just put a timer on physical human involvement in the construction of the space industry? Supposing no development in space happens for whatever reason- spacex fails, nobody does more than some communications constellations for the next 100 years.
    Eventually, self-replicating robots/factories/machines with adequate AI will be developed (with all the economic/social/political/ecological mayhem they imply), and it’ll then become practical (for whoever’s left over) to have them build fully automated luxury space hotels and a dyson swarm for them.
    Or maybe someone develops a microbe than can live off cosmic rays and regolith, and it infests the surface of the moon.


  2. Let’s think about that last one. By “microbe” I’ll assume a reasonable extension of existing biotechnology. That means you need water inside a cell – and probably really interesting extracellular proteins to keep the cell wall away from the super-dry regolith.

    I dunno about cosmic rays, sunlight is probably more than sufficient.

    To reproduce, these cells will need to find carbon and water, in a form that can be liberated with sufficiently low energy investment. Both carbon and hydrogen are in short supply on the lunar surface. There’s islands of water (permanently shadowed craters and rilles) and carbon (meteor impacts) but finding both in the same place is probably hard (we don’t actually know).

    So we’re talking about cells that can travel very long distances while using very little carbon. I guess you could call these spores. They have lots and lots of oxygen, silicon, iron, calcium, aluminium, and magnesium to play with. Locomotion can be solved in lots of ways.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s