I'm reading a story, "Therapy 2000" by Keith Roberts, in an edited collection of World's Best Science Fiction 1970. The main character edits advertising material for a printed publication, placing it in its proper location on the page. This used to be called "pasteup" because it used to be done with physical scissors and glue.
The author looked thirty years into the future, and tried to imagine in what way technology would change the job. Here is his description:
In fact Travers used a Grant and Digby, a bulky combination of epidiascope and dyeline printer that enabled images to be enlarged, reduced, squeezed, expanded and jazzed up at pleasure before being fixed by the simple pressure of a button. It was a nice machine; some illusion of privacy might be gained once Travers had involved himself in the intricacies of its various folding black plastic hoods.
Of course, in just about half of those thirty years, the graphical user interface and WYSIWYG software created desktop publishing. No need for the intricacies of folding black plastic hoods anymore. That's because computers are a general-purpose data processing technology. From Heinlein's space ship pilots loading three-dimensional cams into their navigation computers, to Samuel R. Delany's colonists on Triton reading microfilm, even the brightest science fiction authors failed to understand just how useful and powerful digital computers could be.
A general-purpose, digital, nanoscale manufacturing technology will be just as flexible as computers. As I think about all the futurists who have underestimated computers, perhaps it shouldn't surprise me that molecular manufacturing is similarly underestimated.