I'm still new in developing my data professional career, but here is my two cents.
I personally view what I do and what others do (better) as experts in all things data (or at least the pursuit of). It's a different mindset, a different set of skills and objectives that normally are not a primary concern of a developer.
Because of that, putting more weight on the developer to control the schema through code along with everything else is just minimizing the need for data professionals and consolidating those responsibilities to one position. A position that already is overworked, under resourced and possibly not even specialized in the consolidated areas.
That's a problem simply because the profession, something we are trained or training ourselves to do, is considered easy enough to just consolidate and stick all on the developer to develop, scale and maintain.
I think that's a decision that ultimately will lead to more issues in the long run, especially when the only benefit is making it easier for the developer to be more flexible with the data as opposed to letting two professionals correctly identify what is best in both areas as most of us do today.