This sounds like a different variant of the "rampant layering violation" argument that was used against ZFS. It is sometimes healthy to challenge old assumption

This sounds like a different variant of the "rampant layering violation" argument that was used against ZFS. It is sometimes healthy to challenge old assumptions. If QUIC turns out better than what we have in every respect that matters "for the end user", as this article is about, what would be the harm?