Internally it could work if the teams understand that their services are never done - they're part of a living organism, and the responsibility of a team assigned to a service is to keep it working. Shit will break constantly, but that's not a problem as long as it gets fixed. It's labor-intensive, but done right, we're talking few devs being busy maintaining a process that benefits thousands, or hundreds of thousands of their colleagues. It's what the internal development is meant to be.

At some point in our industry, "service providers" started thinking of themselves as kings, instead of what they were supposed to be - servants.