They made a really big, really stable molecule, inside an E Coli bacterium.
Synthetic machinery that we design could be made functional inside the protein complex so as not to disturb background cellular processes.
Receptors for specific cell types could be attached to the outside so that the complex is taken up and a payload is delivered.
All gene-encoded proteins are defined with atomic specificity. Well, atomic specificity as long as you ignore the existence of stable nuclear isotopes.