One possible disaster facing the earth in the not too distant future is the grey goo problem, a hazard of the development of nanotechnology.

What is nanotechnology?

The normal progress of technology is towards making smaller and smaller versions of things by progressively reducing the size of the components*.

Nanotechnology reverses that approach, and instead of making things progressively smaller, starts with the most basic building blocks possible - individual atoms and molecules. Devices are then built up using the minimum number of these fundamental building blocks. What results is the smallest possible device allowed by the laws of physics - further miniaturisation is fundamentally impossible.

The foundations for nanotechnology were laid in a speech given on December 29th, 1959 to the American Physical Society by later Nobel Laureate Professor Richard P. Feynmann, entitled "There's plenty of room at the bottom". In it, he described how, merely by "writing"* using direct manipulation of atoms on the surface of a metal, it was physically possible to store the full text of every book ever written in a pamphlet you could carry in your hand. Furthermore, if one were to encode the information somehow, much more space could be saved - the full sum of recorded human knowledge could be stored in a piece of dust barely visible to the unaided eye. He offered no description of how this might be achieved, but noted that there is nothing in the laws of physics to prevent us from doing it - it's just a matter of technology.

But the possibilities extend well beyond the maximum possible density of data storage. Nanotechnology also encompasses mechanical devices and computers too small to see even with a normal microscope. The name "nanotechnology" refers to the fact that such devices would be of the order of a few nanometres across - a nanometer being one millionth of a millimetre.

Assemblers

A nanotech assembler is a device which can physically rearrange matter, atom by atom, according to some program to produce a desired result. The first faltering steps towards this aim have already been taken - a team of scientists in San Jose has managed to write the letters "IBM" (their sponsor) in individual xenon atoms on a crystal of nickel.

The aim of nanotechnology is to be able to manipulate matter atom by atom to produce whatever you want. The ultimate device - nanotech's "killer app" - would be a universal assembler. Such a device would incorporate some significant computing power, molecular manipulators, and some form of power conversion - either running on ambient heat or possible solar powered.

A simple example of the sort of thing a universal assembler could do would be to turn graphite (from the lead of a pencil) into diamond. Both are simply different arrangements of identical carbon atoms. The assembler would simply alter their arrangement, atom by atom. And since all organic matter contains lots of carbon (by definition - and that includes things like old plastic bags, used tyres, horse manure etc.) you could use that as raw material for your little diamond factory. And since you have control over the very atomic structure of the diamond, every one you turn out will be flawless and as big as you like in any shape you like. Want a greenhouse in your garden made of a single greenhouse-shaped diamond? No problem. Want a rocket engine combustion chamber lighter and stronger than anything ever built? Done.

Of course, the universal assembler won't just do carbon atoms. In principle it will be able to assemble anything, given the constituent atoms. Most of the things we throw away contain carbon, hydrogen, oxygen, nitrogen, and a couple of dozen other common substances. All this garbage could be used as raw material for universal assemblers. These devices will be able to turn toxic chemical waste into rump steaks, horse manure into life-saving drugs, and lawn cuttings into petrol - for free. They'll be able to swim the seas, gobbling up oil spills and turning them into plankton, or cruise your bloodstream in their millions converting fat deposits on your artery walls into pleasant, side-effect free euphoric drugs.

Of course, developing and building a universal assembler will be expensive. Although they will most likely be too small to see with the naked eye, they will be the most complex devices ever made. The cost will be immense - for the first one. We are familiar with the concept of "economies of scale" - building the first Ford Model T cost millions of dollars, but because millions were built, they could each be sold for a reasonable price.

But even economies of scale don't apply to molecular assemblers. The first one will be vastly expensive. The second and subsequent ones will not be cheap - they will be free. The first task for the first assembler off the production line will be to build a copy of itself, using the discarded prototypes as raw materials.

And herein lies the possibility for disaster.

The perils of geometric progression

Imagine you have a molecular assembler on a table in front of you. You can't see it - it weighs only one thousandth of a gram. You instruct it to make a copy of itself. This process takes about a second - computing speeds for the assembler are high because it is so small, and the actual building process happens as fast as it can move atoms, which is quick indeed.

So after a second, you have two assemblers on the table in front of you - the second one constructed by the first using atoms it literally picked up off the table. You still can't see them. Each one of the two then carries on and builds a copy of itself.

Hang on to your chair, because after only twenty seconds, there'll be over half a kilogram of assemblers in front of you, furiously building copies of themselves from whatever comes to hand (except each other). Probably best, actually, if you get out of your chair, because after another ten seconds the pile of assemblers will weigh over half a tonne, and the table and most of the room will be gone - used up as raw material. You still can't see the individual assemblers - just half a tonne of seething dust consuming everything it touches.

Assuming for a moment that this rate of reproduction could be sustained, the assemblers will have consumed every atom on the planet - you included - in a little over a minute and a half.

Don't Panic

This is an extreme example, and makes many invalid assumptions.

Assemblers may take minutes or hours to copy themselves. They will not necessarily have instant access to all the atoms needed to build a copy - if you only have access to carbon, for instance, you can't make anything except graphite, diamond and fullerenes. Nanotech machines will not be able to change an atom of carbon into an atom of boron - that requires an altogether different level of energy expediture. They may have difficulty cooling themselves sufficiently to operate at those kind of speeds. And responsible designers will program in safeguards to prevent runaway reproduction. The planet will not be destroyed within hours of the invention of the first universal assembler.

Panic Now

Of course, successful nanotechnology will have all sorts of possible ill effects. First of all, as soon as you can effectively feed, clothe, house and educate every single person on the planet for free, money becomes obsolete. Most of the people in power are there because they control access to resources - money, healthcare, food - and if all these things become available to everyone for nothing, those in power will suddenly find themselves superfluous. It's difficult to predict what effect this will have on global society - but it's unlikely to be pretty.

The avenues nanotech opens up for more of man's inhumanity to man are manifold and unpleasant. Tailored mechanical "viruses" to eliminate specific nations, towns, ethnic groups, or even specific people in a variety of unpleasant ways are an obvious possibility. Weaponry built of nanotech would change warfare beyond recognition - no armour can stop a molecular assembler, it would literally crumble to dust under the onslaught, or more likely be remade as something dangerous to the person it had been protecting.

But all these deliberately conceived nightmare scenarios pale next to the biggest threat posed by nanotech - the grey goo problem.

The Grey Goo Problem

Almost all organisms on earth* depend directly or indirectly on the sun. Food chains have at their base organisms which convert sunlight into energy. Plants do this using chemicals such as chlorophyll. Natural selection and fierce competition means they've grown reasonably good at it over the several billion years they've been doing it - but soon they may have competition they can't keep up with.

A nanotech assembler needs power. It could construct solar panels which absorb all the light that hits them - no wasteful reflecting the green wavelengths like plants do. It could build these things as small or as large as it needs to - and since the macroscopic design of plants IS very efficient, it's likely to look like a plant, with branches and overlapping leaves. But since its small scale design is so much more efficient than any plant - near 100% efficient use of the light hitting it - it would displace any plant from any ecological niche. Not a problem if your assemblers are confined to the lab - but an accidental release of these devices into the global ecosystem could result in a mass extinction unprecedented in its scope, devastating in its speed, and from which the earth would never recover.

Even if molecular assemblers were only 1% more efficient at turning sunlight into power than organic plants, they'd begin displacing them immediately. Insects, birds and animals wouldn't be able to eat these machines, so they'd begin to suffer. The maths of geometric progression alluded to above would mean that this displacement would occur not over thousands or millions of years, as is usual in nature, but in a matter of hours or days.

And eventually, when all plant life had been displaced, and all animal life died out, a terrible quiet would settle over the earth. The entire planet would be covered in a film of solar-powered self-replicating assemblers, all near-identical - a grey goo. And unlike every other mass extinction in this planet's history, there'd be no way back - no obscure class of organism to rise up and take over as the mammals did after the dinosaurs, because by their design the nanotech machines would be the very optimum energy users possible. Nothing could ever compete with them, so nothing could ever replace them, except better versions of themselves, built by themselves. And since they would, by design, be self-repairing, they'd never die out. No amount of climate change could affect them, until the sun exhausts its hydrogen fuel, expands into a red giant and envelops and destroys the earth.

The Blue Goo solution

One suggested solution to the problem of "grey goo" is "blue goo" - special "policeman" nanotech devices designed specifically to recognise and disassemble molecular machines which are out of control. The blue goo would be deliberately released into the world, and allowed to replicate to a pre-determined level, there to wait and monitor the activity of other nanotech and act in case of runaway self-replicators.

It's a physically possible solution to the problem - but the human race has a long history of developing technologies which destroy the environment well before they develop the technologies to control them. With nanotech, we will only get one chance - the first accidental release could be the end of all life on earth.

Conclusion

Nanotech promises a bright future for humanity - if we can control it. If we don't, or can't, it may be our last invention.

Further Reading

The definitive factual text on the subject of nanotechnology is "Engines of Creation" by K. Eric Drexler.

The definitive science fictional treatment of nanotechnology and the grey goo problem is surely "Blood Music", by Greg Bear. This exists both as a short story, which can be found in the Greg Bear anthology "Tangents", and as a full length novel.