The middle man was there to make sure everyone understood each other. Now simple things like taking a screenshot requires programs to speak multiple different languages.
And that's the point. X11 provided a framework for all of this. Not perfect, but deliberately extensible. The Wayland alternatives are so much more complicated and brittle, and have to work around the limitations of Wayland. The effort expended by groups like freedesktop.org to architect and standardize new features would have been (and was) so much more productive in the X11 world.
There's also talk of D-Bus eventually being replaced with Varlink [0], showing that this whole ecosystem is still very much a moving target. What feels like endless breakage and friction right now is really just the noise of a massive architectural migration in progress. Eventually, the dust will settle, the APIs will stabilize, and the whole stack will become boring, reliable, and really good. That is how I prefer to look at it.
If you want to see things move in a specific direction or address any particular problem, the best path forward is to get involved and participate in the projects that interest you.