Common misconceptions about atomically precise manufacturing

Almost everything will be buildable - often misunderstood

It is often thought that APM is supposed to be able to produce almost anything (often formulated: all allowed structures permissible by physical law) including e.g. food, wood, plastics and metal parts but this is surely not the case.

Take a look at the "mechanosynthesis" and you will find that the range of materials and strucuctures targeted lies in a very narrow range. The magic lies in the diamondoid metamaterials that emulate properties above the atomic level.

This is not to say it will be impossible for all times to assemble materials lying outside the set of now targeted materials. When the technology will have been around for quite a while [advanced extensions] may be able to do this but this is way beyond the scope of any current day APM attainment project.

No food

Advanced APM is not in any way intended to be a means for food production. Structures out of solvated weakly linked non stiff proteins and lipid layers are a good example of "anti-diamondoid" materials.

To be sensible food tissue construction via advanced mechanosynthetic means (e.g. a pie like this hoax [1]) must be quite a bit faster than biological machinery. This may be expectable but at this point the highly diverse tool-tip chemistry at cryogenic temperatures and at the threshold of stability needed poses a prohibitively high barrier. That is barely any exploratory engineering can be applied here. Further some kind of hierarchical assembly that completely replaces the natural system would be needed.

Attempting to create exact copies down to atomic resolution of an original tissue at this point seems ridiculously complex. Some kind of very advanced scan (atomically precise disassembly) of the original would be needed to be performed in advance. Trying to compress quasi-random atom configurations data hierarchically like in diamondoid APM systems would probably lead to strange unnatural compression artifacts. The need to produce everything in a frozen state (ice crystals) might be a problem but one of the most minute ones.

Attempting to create genetic twin tissue (avoiding the need for a complete scan) has the problem that information extraction from DNA to a spacial (not only typological) atom and molecule configuration is not straightforward to say the least. There's not only the forward protein folding problem but also the yet unsolved riddle how body shape at all scales is encoded.

Note: Plants are already self replicating and thus cheap. Most people just don't grow all of the plants they consume because they need space, sun, soil, and often industrial post processing. APM will bring all the other stuff to the same or lower price level per mass. Including means for easier plant breeding.

No disassembly

It is often thought that the capability of taking things apart atom by atom would be become available just when one starts to be able to put things together atom by atom. This is far from true. Taking things apart atom by atom is a much harder problem that is probably impossible in many cases. For the details visit the "atomically precise disassembly" page. Beside other factors the inability to consume just about anything harshly this limits the grey goo scenario.

APM is not nanotechnology

Nanotechnology is a huge field and a big part of research is done on interesting things that are on the verge of falling apart.