First, the author links to WHATWG HTML but labels the link "W3C". Then the article talks about "XHTML" which hasn't been a thing since 2009 or so. Finally, the author pulls HTML validation results from W3C's nu validator site via curl, but you can just run it locally as a Java app, which also might've the benefit that you're in control over the HTML version to validate against.
To the latter point, W3C's most recent recommended WHATWG HTML snapshot (with not even a newer proposed recommendation) is HTML Review Draft published January, 2020 [1]. That recommendation , being a product of W3C's and WHATWG's "memorandum of understanding" prohibiting W3C changes to WHATWG specs (or does it?), happily accepts a number of long-standing WHATWG design decisions W3C had always rejected, such as liberal HTML heading nesting levels and the so-called HTML5 outlining algorithm. While that part of the spec was finally rewritten last year by "properly" (wrt to the MoU debate) editing the upstream WHATWG spec, it didn't make it into a new W3C HTML version.
I apologize for labeling the link incorrectly. That is the standard the W3C recommends anyway: https://www.w3.org/html/
The article doesn't discuss XHTML; it simply highlights that a validated document remains reliable even after the standard has become obsolete. I suggest you reread. You don't have to run the nu validator locally; there's an API, which I used in the article.
I'm afraid the issues between W3C and WHATWG are not in scope for the article. It looks to me, you're more interested in that debate rather than the subject of the article.
