I disagree that it is time to retire LaTeX, since it is not just about formulas.

In the humanities, linguists are very fond of it for the use of diacritics and tables etc, while the humanities overall greatly profit from bib(la)tex. Furthermore, the KOMA-Script document classes give excellent calculated full-text page layouts without having to play with geometry. Additionally, hyphenation usually just works, microtypography is an awesome add-on etc.

I have yet to see something providing results that look as professional while at the same time allowing to be used with arbitrary writing programs (text editors!), being free & free and being as well documented.

It seems to me projects like that are great for quickly writing some math formulas, do some non-formal papers for school and so on. But I can't see how replacing LaTeX is even in the intended scope of the project nor how it can come close to it.

I think it's good enough for writing a journal paper. The journals would probably need to stay with LaTeX for the powerful typesetting and layout features, but converting from this MathML-Markdown mix to LaTeX should be trivial for them.

I've not seen a non-TeX layout engine that produces remotely acceptable results for setting mathematical material, for instance for line-breaking equations. I'm sure it is possible but I don't believe it exists at the moment.

Very much this, I never quite harnessed the power of \LaTeX\ despite using specialised editors such as LyX. Although I'd recommend AsciiDoc over Markdown because it's easy to convert to (e)book, whereas publishing Markdown is done by converting it to \LaTeX\ first. (AsciiDoc merely uses it "under the bonnet" for PDF conversion.)

pandoc mybook.md -o mybook.epub

If that produces a reasonable result, I stand corrected.