Of more recent software "clones", there is guix vs nix, mcron vs vixie-cron, shepherd vs systemd, gnutls vs openssl, all high-quality implementations of good ideas.
Sure, lsh has a way to go before it's a viable alternative to OpenSSH, but I wouldn't say it's low-quality. I don't think I've come across a GNU project which felt "half-put-together", rather the opposite.
The GNU projects' motivation is very clear-cut and not at all different from e.g. the BSD camp or suckless. How are they questionable?
Shepherd is hardly a clone of systemd. It's older and has much fewer features. They are both init systems.
The relationship of Guix and Nix is also not that of rivalry. Guix uses the Nix daemon and subscribes to the idea of functional package management, but everything else is quite different.
(I realise you put "clone" in quotes and wrote "all high-quality implementations of good ideas", but I thought I should clarify anyway.)
I'm working on GNU because I share its vision of providing more computing freedom to people, blurring the lines between admins and users as is the case with Emacs (self-documenting, trivially extensible and modifiable), Guix (trivial inspection and modification of all packages on a system, a large step towards freedom from system administrators, independent software profiles giving each user what they want), and the Hurd (even more freedom from sys admins).
I'm not even talking about the many compilers in GCC, which make it feasible to build a completely free system (such as GuixSD).
You seem to misunderstand the GNU project, and I find it unpalatable to reduce what you call "China" to a cardboard cutout for the sake of a flawed analogy.