A Study on the Prevalence of WebAssembly in the Wild [pdf]

I've been less optimistic about WebAssembly than most HN commenters, but I was still surprised by this paper. WebAssembly is being used at all on less than 0.2% of sites; those uses boil down to just a couple of hundred of unique modules; and half of the use is malicious.

Why is WebAssembly such a flop? I would argue it is because it solves a problem that very few web developers face (pure computation speed) at the cost of worsening a problem that very many web developers face (complexity of the web platform).