KDE 3 would have been a much better example. But these days are gone, too.
What about R and Jenkins? Are there paid developers working on these?
I feel confidant in saying that yeah, there probably are. They're being paid to implement features and fix bugs found by their companies, who use R and Jenkins.
How so?
Blender cannot cope with the kind of stuff that is thrown at commercial DCC tools in feature production. Parts of it don't scale, others fall short of what would be required. Many features are kind of there on paper, but are not fleshed out that well.
Blender faces a sort of a chicken and egg problem: if you don't push it, you can't find out what needs fixing. But nobody in the industry is willing to try because they know that it will break too easily.
This is sort of why the Blender Foundation creates these open source movie projects: they are there to figure out what needs fixing. A ton of new features had to be created for each new short. Last time I bothered to check, many of these features were stuff that would have been considered absolutely essential in VFX workflow at the time the Blender developers managed to implement it.
The downside to these projects is that they also lead to the addition of features like compositing and editing workflows that are out of place in a DCC tool (especially the editing part). Blender started to overextend itself and ended up with too few developers to push all of these areas equally well.
Blender is striving to meet the features and capabilities of Maya and 3ds Studio, not the other way around.
A more "business" orientated sentence:
Blender is required to to split its focus to provide a complete movie/game making package including a movie editor, shader editor (i.e. photoshop), and animation editor, an audio editor, a physics editor, a compositing editor, and a game editor; Maya and 3ds Studio do not.
Blender just can't keep up. It's a good tool, no doubt about it, but few professional users will end up using it to make money.
Over the past decade or so, Blender has been on par, and even ahead of "industry standard" tools.
It isn't required to split its focus. It is able to. That is something proprietary software from a company would fail at. It's something that free software and collaboration succeed at.
The biggest benefit is that it has all of these tools in a cohesive environment.
Instead of a photoshop/GIMP-like menu of functions, Blender is a cohesive manageable UI.
There are, in fact, a wide variety of professionals using Blender to make money. You don't hear about it, because telling you what they use is not their perogative.
What features to Maya/Max/etc. have that Blender is really behind on today? I can't think of any.
Now, there is at least Neutron as an open source compositing tool that seems to be filling that particular hole now much more competently than Blender ever could.
I do not understand how you can consider Blender's UI "manageable" in any way. There are much better ways to build UIs for enormously complex DCC tools. My go-to example for that has been Maya for a long time, but since a couple of years, Autodesk has bolted on features in a way that spectacularly breaks the internal consistency, logic and general discoverability of features in Maya. Even though, Blender's UI is still an unsorted mess compared to that.
Expanding feature sets is something commercial software fails at? The biggest reason nobody professionally uses the open-source alternatives is they're missing features.
There are plenty of examples where they have done just that. Blender is probably the most obvious.