I'm with you, it's certainly not the most elegant solution to the problem, but we're basically Doing Things That Don't Scale. (http://paulgraham.com/ds.html)
As we grow, we identify bottlenecks, and we write software to automate those processes; I imagine we're iterating towards something like what you're envisioning. (At least internally, if not externally.)
But your last couple sentences got at an important point... as much as I'd like to see this problem solved with software, there are still a lot of unavoidable, messy, people-related issues that, currently, still require manual finesse.
As we grow, we identify bottlenecks, and we write software to automate those processes; I imagine we're iterating towards something like what you're envisioning. (At least internally, if not externally.)
But your last couple sentences got at an important point... as much as I'd like to see this problem solved with software, there are still a lot of unavoidable, messy, people-related issues that, currently, still require manual finesse.