Like pascal_cuoq, I was bothered by the apparent lack of rigour here. The problem is a mismatch of expectations. The post is presented like a paper (references, novel algorithm, comparisons, future work), but it isn't a paper. Viewed as a blog post or work-in-progress, it's very cool. Viewed as a paper (which are available on the Web nowadays), it has some misleading statements in it. There is nothing anti-wonder or anti-knowledge about pointing that out.