[exratione]

I think the general consensus in the futurist community is that this vision requires molecular nanotechnology, which means first sorting out the basics of mechanosynthesis in carbon and simple nanoforges. There are groups working on the foundation technologies, but not many of them in a deliberative way with nanoforges as a destination (E.g. Zyvex).

Translating the section of the present human-based manufacturing industry needed to support a modern village into a box that can be built and maintained by a much, much smaller industry seems hard to do unless said box is something very similar to a nanoforge.

There are important differences between what you can do with a mature general purpose rapid prototyping system and a nanoforge. Foodstuffs directly from raw feedstock, for example. But this has all been debated at great length in sections of the transhumanist community for decades now. There is a lot of good writing on this topic out there to be discovered if you go looking.

[lsparrish]

I'm not sure why people seem so convinced that MNT is a hard requirement here. To be sure there are some components that require fine micrometer to nanometer level precision, but existing mechanical and/or chemical approaches do work, otherwise we wouldn't have things like computer chips

You can draw fibers, use cantilevers, exploit the wavelength properties of laser light, electromagnetically control the path of ionized materials in a vacuum, use piezoelectric actuators that convert current to angstrom level movements, and so on. Not to mention the many approaches to coating a surface with a very thin layer: vacuum deposition, spin coating, electroplating, etc.