It doesn't. The question peer review, or refereeing asks is whether an article is fit to be published, based on a quick reading of a paper, essentially, "will we embarrass ourselves if we release this?" Peer review most definitely does not prove a paper correct or beyond criticism. These idiots genuinely imagine that science progresses one paper at a time, and peer review is a means of proving those papers correct. Used in the tweet above, it is just a way of derailing discussion by introducing a red herring and demanding it be refuted.