Seriously, KDE isn't new. Desktop has really shitty defaults, which is a major problem considering how many people don't know how to change the settings. I'm talking about "not a computer person" people, here.
Are you actually disagreeing, though? We both seem to agree it needs work. And you're correct the challenge is partly an institutional one -- recruiting and retaining design talent and manpower, and fashioning new ways of working together. It's something we're aware of and working on, and if you'd like to work on it with us we'll welcome you with open arms.
The KDE community isn't unaware of its weak spots and isn't satsified with its weak spots; addressing weak spots is Real Work[tm] though, and doesn't always happen over night. We'd like to think we've made some progress (for example we've identified some things that don't work from past attempts, and that's useful institutional knowledge as well), though. And that it's very much worth doing.
tl;dr If open source is bad at product design, we'd rather not accept that as law of nature and work out how to get better. If that's a challenge you have an appetite for ...
My corporation has a strict KDE/Qt-first policy for any linux apps, due to the useability and comfort that comes from a KDE desktop.
I still keep my users at KDE 4.x for now, because 5.x is not quite ready yet. but it's very close.
so, I guess this is me actually disagreeing with your actual disagreement.
This is open source -- you get to see how the sausage is made.