Why does it matter? Turing completeness is a low bar. Both exp(exp(N)) and log(log(N)) algorithms solve the AI problem in a finite number of steps, but one of them is a useless abstraction, while the other one really works.
I'm a little late to the party. What are those containers really?
One feature that web really needs can be called a "container" or an advanced iframe. That super-iframe would allow to intercept or change the global JS API for the underlying site and everything that app transitively loads. For example, the site wants to use getUserMedia for tracking? Fine, because the container it's in has replaced getUserMedia with a mock, in a way not detectable by the site. This would allow to see what JS API the site uses. This would effectively kill fingerprinting.
Author here. If you follow the github link, you'll see a few sample images and a live demo. It uses WebAudio to capture audio samples, runs FFT to calculate ACF and renders the result in polar coordinates with WebGL. Works on desktop (Ubuntu) Chrome at 60 fps.
Yes, you need to remove spaces from the link, because the HN's spam filter is ludicrous.