WA makes it possible to replace entire browser runtimes with proprietary stuff. For example, you can compile WebKit to WA, with typography rendering and flow layout on a Web Canvas, and "Save as HTML" and other basic functionality disabled. I find it very irritating that TBL/W3C is promoting this and EME. I think W3C should be seen as what it is - a self-proclaimed standardization company acting in the interest of whoever pays the hefty membership bills struggling to keep relevant.
But still, you can inspect that WA code, and WA doesn't have access to the DOM. Inspecting the EME packet is a crime in many jurisdictions, and designed to be difficult.
WA makes sense as what it is intended to do: give a memory efficient alternative to JavaScript in circumstances that warrant it.
What are circumstances that warrant it in a web browser? The things it makes possible can be had with native apps. What it does make possible is a new model for software and content sales. Which is precisely why I find it has no place in a web spec.
I wish you were right, but the market has decided that the web is the next cross-platform framework.
If this has to happen, let's make it work in a way that doesn't let the stupidest ideas survive. Using WA still requires JS to interact with the user. Which is far better than the monstrosities that JS is being forced to perform at the moment.
Some simple, relatively sane things that WA can do, but JS is a bad choice for:
Whereas EME is more, trust this native code please, and give it all the access it asks for?