Just a caveat to that statement, you can literally do interactive and dynamic 3D graphics rendering in PDFs: https://helpx.adobe.com/acrobat/using/enable-3d-content-pdf....
You can also embed JS in PDFs: https://helpx.adobe.com/acrobat/using/applying-actions-scrip...
Even most simple interactive things can easily not work correctly even in more widely spread PDF readers.
IMHO PDF is in many ways worse then HTML, it's just that this ways are less commonly used, but if you start a PDF instead of HTML trend it's just a matter of time until this "not so compatible" aspects of PDF become widely used by some people.
This guy is arguing that removing JS is what makes the web better. Having published, static, paper-like content is the way forward.
As someone who has had to extract data from large sets of PDFs and modern web presentation formats, I'm not a fan of either, really. Even verifying that a visibly presented string exists in a PDF document programmatically can be a non-trivial task, as with a given website as well. That to me says a lot.
For what it's worth, the same objection occured to me. The use of scripting I've seen in PDFs has been use-supporting and consistent with their book-like feel.