The author references a Brooks publication in his rant, but I wonder whether he has considered the lessons of Brooks' other works. One of the problems with developing lots of software that has to work together is scaling up to actually do just that, and one of the choices made to make this happen is to reduce the communications overhead even if that means people duplicating effort, making their own technology choices, and so on. If the resulting system can withstand this "anarchy" both in terms of resources and maintainability then such a tradeoff has to be considered acceptable because the alternative is that you don't deliver anything at all.
(In "The Mythical Man Month" I seem to recall Brooks going into some detail about the distribution of paper manuals to development groups on a regular basis during the development process. Although the overhead of such communications would be less in today's networked reality, it emphasises the kind of effort you need to coordinate groups within a monolithic project.)
I think the author conflates "cathedral" with "software engineering" and "bazaar" with "hobbyist practices" when in fact both styles of development are best practised with proper software engineering techniques. There are probably plenty of cathedral projects suffering from the kind of unsophisticated techniques that the author bemoans.