That's in a different place (p. 11), but the gist is still the same. Even my quote includes "The WhatsApp server has no access to the client’s private keys".

I guess it wasn't clear, but I was trying to refute the claim implied by the Twitter post (by showing that the document still claims that WhatsApp servers don't have access to the private keys).