An email correspondent asked an interesting question (thanks, Wendy!) about whether we've missed the moment to regulate molecular manufacturing.
I see two main sources of regulation, plus a distant third:
1) Commercial regulation, dealing with the economic implications of a manufacturing revolution.
2) Military regulation, dealing with the weapons and geopolitical aspects.
3) Maybe criminological regulation, since new products can often be used for new forms of crime.
I think commercial development of molecular manufacturing will likely be too fragmented for effective regulation. There may be some regulations that have substantial negative impact, comparable to the creation of software patents in the mid-1980's. But I suspect most commercial regulation will be trying to solve non-problems (defend obsolete monopolies) after the fact, and won't have too much impact.
A major question on military regulation is whether molecular manufacturing will be developed mainly on the commercial side, incrementally, as computers have been, or whether it will be developed first in a "Nanhattan" project. In the latter case, it might be seen as dangerous (especially if used in war in an especially flashy and gruesome way) and susceptible to blanket regulation. That could hamper commercial use for decades. But if molecular manufacturing develops incrementally, it may be no more regulated than computers or biotech.
Regulation of crimes and other social impacts will probably be about as effective as regulating illegal drugs.