Self-replicating nanobots: time for a new law of robotics?

Like I said, we are in the early stages of imagining and designing nanotech machines… there may well be bacteria sized members of the swarm, and they will be the most numerous, but there has to be largr members of the swarm to transport and replicate and control… at least in the earlier designs (perhaps later the whole system could be incorporated as nanoscale machines, but that could be far in the future.)
No, the Fraggle Rock analogy is very good- each machine would be designed to be the best size for it’s function, and the active site of construction /destruction may be occupied by bacteribots, but that is how you digest and metabolise too.


SF worldbuilding at
http://www.orionsarm.com/main.html

Smiling Bandit,

As Revtim referred to earlier, Drexler has written what is considered ‘the bible’ on the subject of nanotech. If you haven’t read his works you can get started for free here:

http://www.foresight.org/EOC/

http://www.foresight.org/UTF/Unbound_LBW/

The “gray goo problem” is addressed in the EOC chapter 11, “The Threat from the Machines” section.

Replicators are discussed extensively in other portions of the books.

It’s been several years since I read both of these books, and I’ve only casually kept up with the latest achievements. However, I do not recall* any of the frailties you have described about nanobots. Your statements contradicting Drexler’s theories seem to be founded in some recent practical limitations that I am not aware of. Could you please provide a link (or any other cite) to the latest discoveries that have formed your opinions? I’d like to get caught up on the facts.

Thanks

(* of course my memory sucks… I can’t wait for the days of the nano-Cray wet-ware embeds.)

Thanks for the link. I will read it when I have a little time.

Its a matter of physical limits of materials. there is simply not space and energy enough to permit nanobots to do much. The “Grey Goo” scenario relies on nanobots being able to produce their own power from harvested materials, being able to replicate themselves from harvested materials, and being able to chop up nearly anything.

The first problem is not insurmountable, but much, much harder than it seems. remember, no lifeform (humans aside) on the planet earth does much besides take in energy. And, these are microrobots, which can’t hold onto significant amounts of it. Bacteria hold onto only enough to divide. And colonies of bacteria never take over the palce, because they always run into the energy limit of their environment. Nanorobots would have to get all the power they need and store it, which amounts to us humans remaking the path of life. Not an easy task.

The second problem is also more difficult than it sounds. You can’t simply find the raw materials very easily. Plus, what happens when they use up one source? They have no idea where the next one is! This makes them use of lots of power wandering around tasting for the resources they need. This is where things get tight, because the swarm MUST grow or its shrinking from power loss. Of course, it could just eats own dead, but that just uses up more power. Ergo, to replicate, it will have to be in an environment where it can constantly take in new energy and still have access to raw materials.

And that assumes that a nanobot CAN replicate itself. This is a nontrivial problem. Can you build a DNA strand that can lead to the creation of a robotic lifeform? Can you write one that makes even a simply germ? What about a virus? No? There almost certainly will have to be a factory unit of some kind.

by Eburacum45

But this doesn’t help. You now need to make and organize a whole economy, much like ants. You still didn’t say how the nanobots are going to “know” what to do. You can’t ignore this problem, and you canot simply build control systems. The only real things that can work is a chemical adapter, but that’s hardly very accurate, and further increases the already difficult task of replicating the “swarm”.

Yes, in fact it conflicts with our current understanding of the physical limits of computer systems, so I’d say that we’d have to break several laws of physics first.
The last problem is the real doozy though - how are you going to make nanobots that can take apart any structure? Even if they could take apart any chemical structure, they will be expending energy if they aren’t somehow gainig it. And how many “tools” will they need?Barring some awesom scientific discovery of a Universal Solvent, you are going to have to have some damned big nanomahcines to hold the chemical swiss army knife. take the two together, and the nanomachine also MUST have a chemical Universal Constructor to all the peices together into new nanobots. If it isn’t, they whole group of nanomachines WILL die from energy exhaustion.

And then, of course, there is the question of exactly “why” it would go cutting up everything. Unless it had some kind of impulse to do so to everything it comes into contact with, it will just sit there and die.

But at the last, a strong electromagnetic signal will simply kill them all. Theya re basically a tiny bit of metal, so pumping them full of jiuce - and they REALLY cannot take much - will kill them dead.

I don’t think they will be metal, although they have metal parts and crystalline parts- whatever they require they will have. Electricity would only affect them in the same way that it would affect gritty cheese.

The control system is likely to be an artificial intelligence of some sort- as you point out, controlling a nanotech swarm
(not a nanobot swarm- that is merely a bacterial infection)…
computing nodes are envisaged by Drexler on the nanoscale, so control inside the smallest machines is not seen to be a problem;

  • communication between the smallest machines and the energy bearing machines and the matter compling machines and the transport machines is likely to be a problem, as you also point out… there have been many suggestions, optics and chemical markers and nanofilaments are possibilities- there might be analogues to slime molds and fungi as well as ant and bee swarms…
    but mostly, the solutions will be ones we have not yet arrived at yet.
    Thanks for all the arguments- they are instructive and helpful for those who are tryingto evisage these constructions.

Finally, is the grey goo scenario plausible?
I would say no, until picoscale manipulation is available, which may be never.
But if you imagine a swarm of cow-sized, rat sized, flea-sized, amoeba sized, virus sized machines, which descend on a field of crops or the blacktop roads in a town and convert them into anti-personnel mines…
you have an idea of what might be the future of warfare in a hundred years or less.


SF worldbuilding at
http://www.orionsarm.com/main.html

In addition to a law against nanobot reproduction, I propose that we pass laws against Jedi knights getting married, and I propose that we have all starship crews trained in using the deflector dish as an energy weapon against Borg cubes. I also think something should be done about the goddamned Daleks.

In case it isn’t obvious, being worried about self-replicating nanobots strikes me as being sci-fi-fueled idiocy.