I have no idea how this became the widely held belief about markdown. Markdown is based on Gruber's buggy Perl implementation of a rich text format that's intended to be fairly readable while editing the raw source files. Gruber had zero concerns for how other people would use it, he didn't even support standardisation efforts that tried to resolve the contradictory informal spec he had published.
To be fair, his use case was/is for writing very simple things (blurbs on his blog), that simply don't get into painful corner cases. It's a simple tool that works great - for simple things. It can fail hard on more strenuous cases. As you say, he resists attempts to resolve the ambiguities, yet, he also guards the name "Markdown" - which has all the mindshare - causing headaches for others who try to resolve those ambiguities. So, we get things like CommonMark.
