Engineering DS 2 -> 3 -> 3.1, 3.1A ->4, etc. is not entirely easy. Determine DNA sequence and controller changes (what or whats, where? here, there, there too, etc.). Create and insert sequence(s). Test spinneret function and output. Test loom compatibility. Test final product(s). Test genetic stability of changes over many generations and fiber specs forever.
Given combinatorics, kinetics and nature of nature, that is a complex undertaking and probably why DS 2, 3, 4 etc will be "baby" steps while/until silk worm genetics etc. work toward complete programmability (after all the programmable computer age really started in English looms). Then bigger steps can be taken at reasonable cost. Computer. Print me some new shoes, etc.
Today's state of the gene-art: insert this here, is a bit like tweaking one or a few gene(s) in a complex and incompletely understood network and creating a specific, custom product. Oh, you like a bit more or less X in the product? Without excellent understanding, the response might be: OKaaaaaaay, we will get back to you next year.
If E. coli (or some other bio-reactor) can create at reasonable cost, fibers of length that can be blended into a thread and subsequently woven into product (... and that is a big set of IFs), custom fiber engineering and blending will be cheaper, better, faster ... and unfortunately easy to steal. Years if not a decade away but coming this way.