We already did that. It was called Plan 9 From Bell Labs[1]. And while it gave us UTF-8, procfs, 9P, etc, it failed to become a popular OS.

[1] https://en.wikipedia.org/wiki/Plan_9_from_Bell_Labs

The important question here is, why? Was it because of those features? Or something else?

I don't know, but one major factor in OS adoption is network effects. "The rich get richer", and it's very hard to introduce a new and different system, no matter how good it is.

Unlike Unix, which was basically free software and therefore extendable and available for companies to build their own versions of, Plan 9 was proprietary from Bell Labs.

Ah, Plan 9. Vastly superior to its predecessor in every way, except the one that counts: bums on seats.

And, for a long time, price/licensing.

The thing about the crufty being on top instead of inside it is that we can replace it from time to time.

(Too bad the replacements aren't always better...)

Having cruft in your top layer is fine and healthy; a sign of a living, evolving system that is still learning how best to do its job and hasn’t prematurely ossified.

The problems occur when a still-crufty layer becomes set in stone, ensuring its base defects go on to pollute everything built on top. That may happen either because its designers lack perspective and cannot accept their precious design is flawed; or because everyone else starts building on top before it is ready for prime-time, then reject all attempts to straighten it out because that would require them to rewrite as well.