The blogpost you linked to is explaining, in detail, with a worked example, that zero-knowledge proofs are possible. The only thing in it that even slightly matches the reddit guy's view is the sidenote near the beginning pointing out that just because someone claims a cryptosystem is zero-knowledge, doesn't mean they're actually right, or that the cryptosystem is secure.