It's not just the people with JavaScript turned off: it's people who, for example, rely on the Readability or Readable bookmarklets or Safari's Reader functionality.
In general turning documents into programs deprives users of those documents of a kind flexibility that they enjoyed when documents were just data.
And does it not bother you that web-browser development has gotten so complicated and labor-intensive that there are exactly five organizations with the resources to maintain a web browser?
What hope do operating systems with very small userbases like Plan 9 have of ever running a web browser capable of displaying correctly the majority of web site?
Browser complexity closes off certain opportunities: e.g., about 15 years ago, a blind programmer named Karl Dahlke wrote a "command-line web browser" called "edbrowse" that has a command language similar to the line editor ed. Is it OK with you that the fraction of web pages browsable with edbrowse keeps going down?
Another way that making the web a richer application-delivery platform reduces the options available to users: approximately nobody bothers to maintain a local copy of the web pages they have browsed (which would be among other things a useful insurance against web pages disappearing from the web) because it is so complicated to do.
And then there is the loss of consistency useful to readers. For example, when you click on a link, then hit the Back button, the browser used to always put you back at the same place in the web page that you were when you clicked the link. Not anymore: for example, if you click a search result on hnsearch.com, then hit Back, you are taken to the start of the web page containing the search results with the result that you have to scroll through results you've already sifted through just to get back to the state of progress you were in when you clicked the link.
A possible reply to that is that the maintainer of hnsearch.com should fix his web site. But the number and variety of "losses" or "regressions" like that one is so large -- and increasing so fast -- as to make me doubt that webmasters will ever get around to fixing most of them, particularly since webmasters on average are less technically knowledgeable than, e.g., programmers are.
Selecting an extent of text in preparation for copying it is another thing that has become less consistent and controllable over time: sometimes when I just want to select a few words, slight movement of the cursor while dragging will cause an entire adjacent column of text to be selected or de-selected according to rules that are essentially unknowable to the reader.
In the past, for about 15 years, the space key consistently meant "scroll down a screenful" (provided the page as a whole has the focus -- as opposed to, e.g., a TEXTAREA in the page). The desire to turn the web into an applications-delivery platform caused the web site to gain the discretion to map the space key to something else, which is a gain for authors of web apps, but a loss for readers who used to be able to depend on consistent behavior from the space key.
In summary, although I am happy that many thousands of applications developers are now able to make good livings without becoming a "tenant" or a "captive" of a platform owned by a single corporation, I am sad about how complicated, tedious and mystifying it has become to use the web to consume static content -- and how expensive (in programmer time and effort) it has become to put static web content to uses not foreseen and provided for by the author of the content.