For these points, I think it's crucial to distinguish "Myers-Briggs style tests" from The Myers-Briggs Type Indicator.
The first thing, MB-style tests, are amorphous and have been improved in various ways since the invention of the test.
That second thing is a formalized test with licensed practitioners. It's been used to disqualify people from jobs, and has produced a huge amount of secondary literature/analysis which requires the binary distribution to be meaningful. This formalized test has been worth a small fortune for the people who own it, and their training seminars and public literature aggressively go against both points 1 and 2.
It's fair to criticize MBTI as, say, needing to be binary. 16personalities is a secondary group which is doing a different thing than the still-popular, rigorously-bounded original.