Nice that the author is enthusiastic; but everything he praises has been around for many years or has a rather hobbyist touch (e.g. what he writes in "The Edge" section). Webassembly has undoubted potential, but in the original use case (browser applications) it is still clearly too slow and inflexible; and interestingly, the adaptation seems to take place more in areas that are well occupied by Node, CLI or the Java VM. I just recently looked again at the current state and studied current literature with the intention of writing a WASM backend for Oberon+; but the technology probably needs a few more years to be worthwhile; at the very least, a built-in GC (or at least a means to scan the stack) should be available, otherwise the implementation would become incredibly inefficient.
GC implementations over Wasm definitely have perf issues. eg. Go compiled to Wasm has these with its GC. The main way when compiling to Wasm now is to not use a GC, which is reasonable depending on the application in question. It's quite reasonable for games or graphics applications.
