While reading the thing I could not help but feel like what I'd call "sensible markdown" is both very close and bijective to org syntax. The trouble comes when I need to define what is "sensible markdown" which is a thing I never attempted to do, but pragmatically works very well. It seems like it's CommonMark + GFM (for tables, checklists, and syntax-aware fenced code blocks mostly) + MarkdownLint[0], the latter enforcing stringent rules that makes Markdown very consistent and portable as well as prune the nonsensical stuff (like double space before newline or underline headings). It's fantastic to use with VScode[1] or vim-ale.
Some pet peeves:
I really don't ever get the argument about Markdown URLS. Being able to do both [](url) vs [][ref] is awesome (the latter I liberally use when writing long form docs). If one finds it so annoying one would make the mistake like three times and then it'd be burned into one's brain due to selective pressure. Or just use [][ref] style (which is more readable anyway since you don't have the URL right in the middle of a sentence).
Also, very personal thing but my brain seems to hate /italics/ as it branch predicts into parsing that as a regex.
What drives me nuts though is all the variants and ad hoc syntaxes like JIRA's or this very board's, which sit right in the uncanny valley of nonsense.
[0]: https://github.com/DavidAnson/markdownlint
[1]: https://github.com/DavidAnson/vscode-markdownlint