If you have a large amount of purified protein it is trivial to measure out a certain quantity.
The ribosome is one such reactor... In fact, some recent studies suggest that features of the ribosome help to aid the folding of proteins to their native state. Although the primary sequence of proteins determines their structure and folding, in practice, it is pretty difficult to get large unfolded proteins to fold back into their native state.
It's not completely ridiculous, but I'm just pointing out some potential problems with your idea. After all, if you can think of a solution to these problems, the idea won't be so ridiculous after all. Chemical synthesis makes it much more easy to add unnatural features to the protein you're synthesizing (such as unnatural amino acids, amino acids with the wrong chriality, sites for attachment of specific post-translational modifications, etc.), so it could definitely offer some advantages over recombinant expression.