Harness the Grey Goo — Self-Replicating Machines

Evan Chan
Evan Chan
Nov 2 · 8 min read

Most people (especially if they clicked on this article) will know about the grey goo scenario; an apocalypse of multiplying nanobots taking over the planet. But are machines like these an existential threat… or one of the greatest tools available to mankind?

Nanobot swarms are commonly mentioned as an end-of-world scenario; but they hold great potential.

Space technology has always been one of my favourite areas to study, and I spend lots of time exploring the impact that different equipment, government, and culture would have on a space-faring civilization, and interests of a spacefaring civilization. Mix in a healthy dose of AI and you get one of my favourite concepts; that of a self-replicating machine, especially a self-replicating miner or colonizer known as a von Neumann probe. I’ll be talking about what they are, why it’s an important concept, and if they’re a good idea to create. This will be a bit more speculative than my normal articles (since this is all theoretical so far), but before we do any speculating, we need to figure something out:

Not Just Nanobots

What are these things?

A self-replicating machine is an autonomous robot which builds more of itself from the stuff around it. This means it is capable of growing its numbers as long as it is fed with resources. However, this is only a basic function; these machines can do more, and often exhibit many of the properties usually used to define life. In fact, some people call these machines a form of life, or at least synthetic life. So far, this concept is purely theoretical, although some factories have gotten pretty close to achieving it (however, it is still financially impractical on Earth to build a robot which assembles itself).

Von Neumann probes could be sent out to create tons of things from habitats to weapons platforms

A variant of the self-replicating machine is the von Neumann probe; these devices are spacecraft which are sent to build themselves, or variants of themselves, with specific properties in order to construct infrastructure for resource extraction, habitats, or anything else that requires prior preparation before humans show up (if they even do — systems created by self replicating machines could even be fully autonomous). Von Neumann probes would be able to push the boundaries of human settlement; or create their own.

I’ll be focusing mainly on von Neumann probes here, as those are the ones that I believe will have the biggest implications.

We’re not Meant for Space

Humans are not built to live in space.

They’re very fragile outside of their home environments, require lots for sustenance, can only propel themselves by pushing off of something else, and die slowly from hostile gamma radiation which damages their DNA.

A machine has none of these problems. Until an area is ready for human settlement, it will probably be extremely difficult to survive there; we can’t transform ourselves to survive better in foreign environments (at least, not yet). However, we can build machines that are specifically tooled for other environments. This means that a swarm of self-replicating machines can build not only colonies, but entire habitats in space. Given long enough, these habitats can reach the size of planets! Habitats, unlike planets, would be hollow, because it requires far less material to channel light into the habitat than to create a solid block, not to mention the difficulty of maintaining an atmosphere. By making these von Neumann probes build hollow habitats, we could use far less material than can be mined from nearby planets — the moon alone has enough resources for probes to build multiple of these habitats!

Concept art for a space habitat. These colossal structures could house humans better than Earth ever will, and can do so with von Neumann probes.

It gets better though — Fed a stream of material, these structures could maintain themselves. This is because every surface in the habitat would, in itself, be a Von Neumann probe, and could therefore replicate and adapt at will. This frees up an unimaginable amount of human labour, and would give us the opportunity to be richer, healthier, and stronger than ever before. But there are good reasons why this shouldn’t be our ultimate goal…

Rebellious Machines and Big Companies

Like many emerging technologies like genomics and AI, self-replicating machines and complete automation as a whole raise many ethical questions. What if the machines rise up? And if they don’t, what do we do with our newfound free time? Finally, how do we govern the controllers of this newfound power?

The problem with mostly uncontrolled intelligences like von Neumann probes is that they must have an extremely strong code of machine ethics — programming to stop a machine from performing unethical tasks. Should machines go about destroying the human race, sabotaging our later missions to gain more resources, and replicating quickly and aggressively? I hope they shouldn’t. Coding in machine ethics is difficult, because the machine. A self-replicating machine race rising up to destroy us isn’t really why we’re sending out those probes.

Many influential scientists and visionaries did and do believe in the danger of a misguided AI rising against humans, or outclassing us altogether.

In fact, some people think that the machine uprising question is a serious one, and these people include Elon Musk, Stephen Hawking, and more. With increasingly powerful AI beginning to make humans obsolete at many jobs, we need to prevent our tools from becoming our masters. Contrary to popular opinion, the biggest threat to humans isn’t a killer robot or a nanobot swarm — it is a group of virtual AIs which can outclass us at every task, because that makes the human race obsolete at everything. A general intelligence is much harder to create than a program which duplicates itself, and that program can modify itself to perform more tasks better than any group of people. That is terrifying. Innovation, research, art, social contact, every facet of human life will be better handled by something inhuman. If those AIs believe it is better or necessary to stop babysitting humans, our civilization would collapse. This becomes a huge issue, because if we send them all over the galaxy, they will be heading very far away from us — so far that it would take literally millions of years to send messages back and forth. In that time, they could transform from a burgeoning colony to a huge space empire each time we receive a message, maybe even accidentally engulfing our system. That’s like starting dinner and heading out for a couple of hours, coming back, and finding that your pot has converted your entire house into a production facility to manufacture said dinner!

On top of this, there is the problem of AI monopoly — if only Facebook, Amazon, Apple, Microsoft, and Google have control of the AIs, and those AIs can outclass humans, then these companies extend beyond the reach and control of government, because controlling the AI becomes more important than controlling the people behind it. If self-replicating machines were in the hands of only large companies, then they could own the galaxy (if there aren’t any big alien Googles and Facebooks to own the other half). I hope most people would agree that a profit-driven entity owning most of the galaxy isn’t exactly our view of a bright future!

Tech giants have a real chance to get out of control if self-replicating machines are developed under their banners.

Slow and Steady Wins the Race

I think we should create von Neumann probes and self-replicating machines. I believe we are far from a machine uprising and currently, the benefits far outweigh the risks if we play our cards right. As more research on cutting-edge AI and these technologies comes out, these opinions will likely change.

Some things that our modern society demands are harsh and unforgiving to produce. Minerals, industrial materials, and chemicals are all raw materials, the extraction/production of which wrecks the environment and people’s lives. These resources are thrown into the endless pit that is our consumerist economy in Western countries. This sort of cycle isn’t good for anyone, and our supply of resources, and environmental hardiness, here on Earth is bound to begin to dry up.

Mining can be very harmful to both people and the environment alike; von Neumann probes move mining off Earth, along with every other resource extraction technique.

Self-replicating machines solve every single one of those problems.

That said, these machines are a huge risk. Overall, I believe that these machines should be given not only strong codes of machine ethics, but networked, not only for communication, but to stop replication at a certain point. This helps get around the issues of constructing machine ethics that can be misinterpreted in a bad way; if we stop them from replicating, they are forced to stagnate, effectively disabling their growth and change. Of course, regulating the machines would be difficult, and even today government regulators and companies debate how they should be regulated; but with enough deliberation and careful planning, as well as only modest excursions within our solar system, we should be able to exploit thousands of times more resources than before with these technologies.

Reach for the stars, but take it slow. If we get careless with these self-replicating machines, the consequences could be immense, but if we get it right, we’re bound for greatness.

If we use self-replicating machines right, and take our time, the stars can be ours.

Takeaways/TL;DR

  • Self-replicating machines can duplicate autonomously. Von Neumann probes do it in space, either to prep for humans or make their own colony. These things don’t exist yet, but we’re close.
  • Von Neumann probes have a huge advantage over humans, because they’re less vulnerable to the hazards of space, and can be designed around them.
  • Machine ethics is a huge problem as AI grows in strength, and if self-replicating machines come to be, they could replace humans at every task. Big corporations could also abuse these devices.
  • Building these things can help us save the Earth, but take it slow and think it through. You don’t want to mess this stuff up — the grey goo is a terrifying scenario.
Evan Chan

Written by

Evan Chan

Hi, I’m Evan, a 13-year old innovator at The Knowledge Society.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade