@dang, this is a great policy and should be incorporated explicitly into the guidelines, IMO. I would also suggest applying this to scientific articles, given how fluffy university press releases can be, as well.
University press releases tend to be better, although they're clearly promotional. The trouble with most scientific papers is that they're highly specialized and, depending on the field, hard for outsiders to read. In those cases we usually recommend submitting the best third-party article and linking to the paper from the comments.
A similar category is legal rulings, which came up yesterday: https://news.ycombinator.com/item?id=24357078