All asexual species found in Nature, are by definition both molecularly self-assembling (SA) and self-replicating (SR). I agree with Chris, that the assembly of complex molecules is an example of self-assembly. Though chemists usually do it as a bulk process, while Nature generally does it molecule by molecule (using enzymes). To me most of what the biotechnology industry does is reengineer nanotechnology based SA-SR systems. Where Nature doesn't completely fullfill the full vision of molecular nanotechnology is that not all of the atoms in a bacteria or eukaryotic cell are precisely assembled (cells use solution chemistry) and the programs the cells execute are stored in ROM (e.g. DNA), rather than RAM, so their software isn't easily modified.

We are approaching the point where we will have the complete blueprints (the genomes) for hundreds of micron-sized SA-SR systems (bacteria) and a few larger organisms. See my Genome Page for a list. Being able to reprogram or replace these blueprints is one of the reasons I have argued for the past 6 years that we will eventually conquer aging.

Scientists currently believe SA-SR systems in Nature require around 350 different parts to manage these tasks. That is if they live in a relatively rich environment within cells. If they live in a resource poor environment where they have to work much harder for their resources, they may require a thousand parts or more (this is an informed guesstimate on my part). So while we do not currently understand all of the processes involved in SA & SR in Nature, we certainly will eventually. Given that 350-1000 parts is substantially less than that found in a car, it seems feasible for us to build SA-SR systems.

Moshe Sipper maintains a very good page on artificial self-replication. There is related collection of links involving the Origins of Life by Daniel Segré.

There is also the infamous NASA report, coedited by Robert A. Freitas, Jr., Advanced Automation for Space Missions, put online by the Molecular Manufacturing Shortcut Group. Another version is here. It deals with the issues involved in the design of macroscale self-replicating factories.

Robert also unearthed this absolutely wonderful example of how you can dissassemble and reassemble living cells (self-replicating systems): K. W. Jeon, et al, "Reassembly of Living Cells from Dissociated Components," Science 167(1970):1626-1627. This is Ref. 158 from Nanomedicine , where self-assembly is reviewed on pages 45, 47 and 66, and self-replication (replicators) on pages 59 & 65-67.

I think one of the primary reasons that there is little SA equipment is that there are economic justification problems. Designing something complex (e.g. a computer or a car) is difficult enough. To design them so that they self-asssemble as well, you would need to have a phenomenally huge market in order to recover the engineering costs. The only way around this is to design and build something small that can replicate itself into something macroscale. Bio-engineered seeds for example could be considered nanotechnology based self-assembling replicators.

I suspect if one stretches a little, you could provide examples of human things that do self-assemble, just not at the molecular level. File boxes that are purchased flat but magically fold into a 3-D box or jigsaw puzzles that you could make such that placed in a shallow box you could shake them into assembly. One of the problems is where do you get the energy to drive the self-assembly? In Nature or organic chemistry it comes from the heat of the environment (or gets harvested from sunlight or stolen from another organism). For macro-scale self-assembling systems you would have to supply the energy externally or build it into the entity (potentially making them dangerous). I'd be concerned if I bought a bicycle for my child if it came in a small box that said on the side "assembles itself"!

Enough academic work has been done on them to prove that, self-assembly is theoretically feasible. I think one can provide examples of self-assembling systems we are beginning to engineer using pre-existing parts from Nature. Gene-therapy vectors based on viruses that require helper cells (i.e. the virus particles self-assemble with a synthetic genome, but require virus components that are manufactured by the cells) are a very good example. Larger scale self-assembling systems, do however seem complex (and therefore costly), so one of the reasons we may not have them is the simple fact that human assemblers are cheaper and perhaps safer.

One has to keep in mind that nanotech has not yet successfully built (or synthesized, rather) an efficient assembler. In fact, overcoming this obstacle is the chief goal of the first-ever nanotech firm, Zyvex. The writer isn't necessarily ignorant, he might just be very sceptical.