8 replies to this topic

[A941]

When will we "see" the first assembler at work?

In 20, 30 years or much later?

[Mind]

You must be asking about nano-assemblers, because robots that can build other robots (macro-scaled things) already exist.

[maestro949]

There's no need to miniaturize mechanical machinery as nature has already shown us how to assemble structures at the nano level. By simply mixing the right molecules you should be able to assemble nearly anything. Whether it's biocompatable or not is another story

[Andrew Shevchuk]

There's no need to miniaturize mechanical machinery as nature has already shown us how to assemble structures at the nano level.  By simply mixing the right molecules you should be able to assemble nearly anything.  Whether it's biocompatable or not is another story

Most synthetic chemists would disagree. The big problem with the more complex structures like the ones Drexler and Merkle have modeled is that nobody knows how to make them right now. This is not to say that they won't work as intended, since the simulations suggest they will, but chemical synthesis tends to be a rather messy and inefficient process currently. Getting molecules to react in a very specific way to form things as complex as those machines either can't be done with traditional methods, or it's so arduous and inefficient that it's not worth the effort. That's why we need mechanosynthesis, to ideally have positional control over the locations of atoms one at a time, rather than trying to coax two molecules into combining in a specific way based on a lot of conditions that are difficult to control.

Whether mechanosynthesis itself is feasible is the great question. It's easy to forget that the structures nature made are not mechanosynthesized, but assembled through standard reaction processes. Therefore I am less convinced that the "existence proof" argument proves mechanosynthesis is possible. Now, of course I want it to be possible, at least as much as anyone else here. But it'll be difficult without proofs of concept to make any claims.

[maestro949]

Getting molecules to react in a very specific way to form things as complex as those machines either can't be done with traditional methods

I would ague that the concept of miniaturizing mechanical structures for the sake of assembling increasingly complex functional components seems somewhat misguided at this point. Nature has numerous examples of building a variety of machines and component parts that perform all sorts of complex functions. Pumps, electron transport mechanisms, communications mechanisms, branes and other scaffolding, protein distribution mechanisms, energy conversion and storage, locomotion mechanisms, etc. Couldn't we simply reuse many of these as building blocks for the more complex functions that we need? We are getting to a point where we can even insert the genes to express the appropriate proteins to generate all of these widgets. Surely we'd still need to figure out how to glue these together into a set of tools that provide utility but I see biomimicry being our fastest path to anything that resembles true nanotech machinery in the near-term.

[Cyberbrain]

Just thought this video might be appropriate

[Andrew Shevchuk]

Couldn't we simply reuse many of these as building blocks for the more complex functions that we need?

I think the short answer is yes, to start with. I doubt that physics requires our biology to be as complex as it is (if no simpler lifeform were capable of our level of intelligence and functionality that would be a terrible blow to AI prospects), and so most of the components could be considered suboptimal in various ways. These components may well be capable of constructing new complex systems, but certainly less efficiently than we could do with our own designs. Sooner or later we're going to be limited by the variety, usefulness, and efficiency of existing components anyway. It's probably just easier to figure out how to do it from the ground up given the eventual flexibility.

[maestro949]

It's probably just easier to figure out how to do it from the ground up given the eventual flexibility.

I agree that eventually this will be true. Using amino acids as base building blocks probably doesn't make sense long-term due to the waste of space and number of unnecessary atoms being used. In the short-term though, using biocompatible component parts would seem to make sense given our ignorance as to how to construct molecular assemblies that can withstand the constant battering and chaos that ensues within cells, the bloodstream and extracellular matrix. Assuming they're even non-toxic, wouldn't many of our own engineered creations that we put into cells be completely torn to shreds or fused into some odd protein mass of muck?

[Reno]

Just thought this video might be appropriate

That video is going to be sooo outdated in just a decade or two.

It kind of reminds me of this one.
TheInternet