Very disappointing to see that there are almost no benchmark quoted in this thread, only words like "snappy" or hypothetical questions. I have been reading articles to understand if converting some of our JS code to WebAssembly could lead to significant performance improvements and other benefits, and I have yet to come to a conclusion. I have seen dumbed-down examples that don't actually mean anything in the real projects, some anecdotal numbers, and other performance comparisons based on proprietary codebase which don't really help. I would advise that people be cautiously optimistic about WebAssembly -- the benefits may not be worth the effort.
