Code is not a bottleneck. Specs are. How the software is supposed to work, down to minuscule detail. Not code, not unit tests, not integration tests. Just plain English, diagrams, user stories.
Bottleneck is designing those specs and then iterating them with end users, listening to feedback, then going back and figuring out if spec could be improved (or even should). Implementing actual improvements isn't hard once you have specs.
If specs are really good -- then any sufficiently good LLM/agent should be able to one-shot the solution, all the unit tests, and all the integration tests. If it's too large to one-shot -- product specs should never be a single markdown file. Think of it more like a wiki -- with links and references. And all you have to do is implement it feature by feature.
So... coding. :P
Processes, tools, and diligence vigilantly seem the most apparent path. Perhaps rehash the 50 year old debate of professionalization while AI vibes coding is barking at the door, because what could possibly go wrong with even less experience doing the same thing and expecting a different result.
I don't think we'll reach this promised land™ until incentives re-align. Treating software as an assembly line was obviously The Wrong Thing judging by the results - problem is how can we ever move to a model that rewards quality perhaps similar to (book) authors and royalties?
Owner-operator SaaS is about as close as you can get but limits you to web and web-adjacent.