All of the claims they make apply to any specification, and yes, divergence is necessary to make progress.
A standard is a commonly agreed to specification, frequently ratified in one or another international organization (ISO, IETF, ECMA, W3C...). The main value of a standard is in ensuring interoperability where that matters more than all the other concerns raised.
Eg. we'd never have much of the internet if people didn't simply settle on the IP (v4) protocol.
Eg. AC chargers had a bunch of different, diverged "standards" for pretty much restricted use-cases (those 1.5mm x 4mm connectors and then micro- and mini-USB). Text encodings had multiple standards for encoding the same text (eg IBM, Windows code pages and ISO encodings) without unification attempts.
In both of these examples, there is one unifying standard added (USB-C and UTF-8 + Unicode) that did stop the proliferation of new standards.
But majority of things never result in one unifying standard that can do everything win: even SGML brough up in this discussion is an example. CORBA also springs to mind.