Nanomunchies

There’s a Twilight Zone, I think (I get all my references secondhand through Futurama) in which a scientist mixes up all the “evil” genetic bits from every animal in the world, creating a superevil hybrid species which he hopes will conquer the planet. As his creation takes his first steps out from the misty incubator, the scientist draws back in horror, for he has simply re-created man.

Aside from being a pretty out-there reference, this episode (assuming I’m not making it up) expresses the best argument I can think of countering the “grey goo” paranoia sweeping the scientific pop press. The fear is that very tiny replicating devices, made using some new nanoengineering technique just over the horizon, will be able to reproduce at an exponential rate, eventually “eating” the entire planet and replacing it with a grey goo of nano-soup. The theory has two things really going for it: the mental image of death at the hands of an invisible swarm, and the idea that the creation of only a single replicator would be enough to destroy the world. Of course, no one really knows how to engineer complicated things at that scale, or how to engineer replicating things in general. Like the fear that we would ignite the atmosphere with the first atom bomb or more recently create black holes with new particle accelerators, the fact that a single mistake would literally be the end of the world justifies the concern. I haven’t seen anyone try to reason through the argument in any logical way, however, so I decided to do so myself, and hope it helps anyone not in the nano/replicating field see why grey goo really isn’t a problem. Or, if you like, it is a problem, but generally not grey or world-destroying.

[A caveat before I begin: I am not a nanotechnologist, though I have had read in the field, and study replication in general. If my arguments regarding nanoscale interactions (or anything else) are wrong, definitely post comments so I can clean them up.]

A fundamental piece of information we are all taught in high-school physics is that “everything is made of atoms.” Look around, and everything is made of the same few dozen components at the basic level. Nanoreplicators work at this scale, right? So if you were some sort of device that could “eat” atoms and reproduce, you’d have an almost infinite supply of food, nothing could stop you, and because of your exponential growth, you’d eventually take over the universe. This does actually work, in a loose sense. You could consider an atomic bomb as this type of replicator, where unstable atoms break and transmit their instability to several? others. This only works with very particular unstable atoms, however (which is why the atmosphere doesn’t ignite), and the nanoscale is actually molecular rather than atomic.

Nanoscale structures, in general, are not able to split or process atomic energy in any way. What they can do is form chemical/mechanical bonds (different perspectives, same bond) with other structures. For any type of complex interactions required for replication, they have to be able to break these bonds too (otherwise they would just clump together). The interaction bonds therefore have to be REVERSIBLE. At the macroscale, you see the same phenomenon. I could glue a hammer to my hands with epoxy if I wanted to, and get a fantastic grip, but if I finish up work I might want to grab a pint glass instead. Enzymes in my body process food energy (and everything else) in a reversible way too, otherwise we’d run out of enzyme, and the system wouldn’t be self-maintaining. There are a lot of things replicating organisms aren’t able eat because of this, for example, rocks.

There are “rocks” at the nanoscale too. Sometimes, such as the case of mineral crystals, they are actually rocks, but many chemical bonds in general are much too strong to be considered reversible by something of that size. The upshot is, if nanoreplicators could exist, they’d have the same trouble we do at the macroscale finding appropriate food molecules.

A second argument goes as follows: Assume that it is possible for a nanoscale replicator to be produced that turns everything into grey goo. Such a replicator would have MANY, MANY fewer parts than a typical cell. If such a replicator was possible, evolution has not discovered it in 4 billion years (because just discovering it once would be enough) but HAS discovered many more complicated limited replicators. In addition, the natural world is to this day continually performing a huge variety of distributed interactions on the nanoscale, while researchers in the nanoscale still struggle to create simple shapes. I’m more concerned about nature, quite frankly.

To shift tack again, let’s consider the properties required for replication. It is a general property of replicators that they must be made in some way from their food, “You are what you eat.” If a nanoreplicator is able to eat me, what would it have to be made of? Humans are mostly protein, fats, and water. Hydrocarbons and water basically. Any replicator made of hydrocarbons and water then potentially has the ability to eat me, but also, has NO potential to eat metals or rocks, or anything non-hydrocarbon based. This nanoreplicator would have to have some sort of hydrocarbon processing mechanism, let’s call it a metabolism, for lack of a better word. Also, the replicator would need some reversible interactions with these hydrocarbons, the catalysts which would need to be hydrocarbon-based themselves. There is a word for these too, enzymes. Some sort of carbon-based control structure as well, otherwise the whole system wouldn’t be repeatable. DNA. Suddenly this is looking like a rather bulky nano-scale device. A variety of enzymes, with an enzyme control structure, possibly needing to be packaged in some sort of hydrocarbon lining (cell wall), seems something more at the microscale. And voila, through (admittedly brief and dirty) logical deduction, our nanoreplicator has become some sort of a cell.

There is a way around this process, though. What if, instead of taking all that baggage with it everywhere it went, the nanoreplicator simply used our own cellular machinery to reproduce itself? That way it would be more specialized than the cell-sized replicator (and so of danger only to other living things), but able to be much, much smaller. Like a virus.

Good thing we have a way to protect us from these nanoreplicators, our immune system. Of course, there’s always the (very real) danger that a new virus or cell could be engineered that we wouldn’t be able to defend ourselves against, but that’s currently feared under the biotechnology umbrella and we needn’t worry about it twice. To put the Twilight Zone I mentioned earlier back into context, our grey goo fears after all this analysis are really just our well-known biotechnology fears.

For all of the reasons above, grey goo is not a concept anyone needs to worry about. Not because replicating technology isn’t dangerous, but because replicating technology has fundamental characteristics that make a universal replicator totally implausible. Specific replicators pose a danger to us, such as cells and virii, but evolution and our unlucky ancestors provided us with strong defenses against them. Nanotechnology may indeed provide us with new and dangerous playthings, but goo is fortunately not one of them.