To me, it seems the gaps are the same as software products:
Software: Business Guy w/ idea -> Developer + Servers -> Product
Hardware: Business Guy w/ idea -> MechE/EE + Manufacturers -> Product
Are you saying the MechE/EE are harder to find/coordinate and/or typical software startup guys don't have the skills to develop hardware?
I agree with you - when we can solve this gap - then we've got something cooler than any one manufacturing process.
Whether it's harder to find and coordinate ME/HW resources is hard to judge. It doesn't appear like that to me and the disciplines have been around longer so there's more maturity in the industry. Typically you won't find the MEs bickering about the latest framework or tool to use. There's more of a get-er-done attitude, especially since the best MEs tend to be more practical minded people in the first place, more likely to go out and pour some concrete on the weekend instead of spending it at their computer screen.
My team mates frequently joke that my job as FW engineer is easier and I tend to agree - only half jokingly. On the other hand, more than once we've sent out firmware fixes to the field that prevented a gear box from stripping or some part from overheating so everything has its place and value.
One gap is simply the body of knowledge necessary to spec and drive that process. The second gap is actually knowing who does what, and who has a good history of meeting their promises. Right now to bridge both gaps you need to go through intermediaries, but it's easy to immagine a world where technology solves both information problems. And to me that's a lot more exciting than 3d printing.