Over on Metamodern, there's a discussion of whether advanced molecular manufacturing systems could be developed by a small private effort.
Drexler thinks it's not possible today, and I agree with him that it doesn't look very likely; enabling technologies are not yet at the point where a small effort can succeed. But what if special-purpose nanotechnologies become increasingly sophisticated, continually out-competing the possibilities of general-purpose manufacturing for another 20 years?
Even though the cost of developing general-purpose manufacturing (where the space of products includes the manufacturing system) will continue to drop rapidly - perhaps exponentially - there may not be enough commercial incentive to develop it in the near future. The end of this projection is that it may not be developed until it becomes really easy.
I have thought for years that, even if no one did any work toward general-purpose nanomanufacturing, around about 2030 it would be a moot point - it would be something that a university lab could cook up in a few semesters, but many of the applications would have been achieved already. So, if no one tries to develop a nanofactory, then in two decades or so we may have one - and it won't be that big a deal - at least not at first...
On the other hand, the technology of DNA shapes has advanced with incredible rapidity over the last decade. With comparable advances in nanoscale actuators, it might not be all that difficult to put together a basic machine that could build general-purpose DNA-machine manufacturing systems. They wouldn't be diamondoid nanofactories, but they might be a major enabler for further work in that direction.
The world could have had digital programmable computers before 1900, but too many people didn't really see the point. Perhaps, in the end, molecular manufacturing will be driven not by technological capabilities, not even by recognition of the possibilities, but by the appearance of a problem that only molecular manufacturing can solve.
Of course, that depends on the definition of "problem" - it doesn't seem to be a problem that tens of millions of people per year die of preventable causes - but that's another blog post.