I find it quite amusing that the author is railing against HTML at least in part because it's practically impossible to build a new web browser at this point, a

I find it quite amusing that the author is railing against HTML at least in part because it's practically impossible to build a new web browser at this point, and then moves to PDF instead. In my time working with PDFs, I've found that generating them in ways that can be read with the most popular PDF readers is cryptic and difficult, and even parsing the ones made from the most popular creators is hard. I would definitely not pick PDF over HTML in regards to how easy it is to implement a good reader or writer. And there's plenty of authoring tools for HTML already, so the "ecosystem already exists for PDF" doesn't track either. Even the complaint about churn makes no sense to me, because there's no need to upgrade your tools constantly. If you're using something that produces good HTML today, it'll produce good HTML in a decade, too. OTOH, if you have a problem that could be automated, you're a lot more likely to be able to create that tool for HTML than PDF, and it's quite likely that someone else already has for HTML, but not PDF.