> Perhaps Google is doing a decent job eliminating crap from their search results (debatable), but that’s not an outcome of a naive application of PageRank.

Is this actually the case? We don't have a good way to test it, but I believe that you wouldn't see the crappy articles I referred in the first N results even with vanilla PageRank (and for a very large N at that).

Keep in mind that my definition of fake news is pretty narrow here: it's "Pope Francis endorses Trump" rather than "Fox dismisses COVID-19 threat" or spin of that sort, which is probably what you meant by crap.

There will certainly be blind spots like the ones above, but my goal is to eliminate the most egregious miscreants and the method outlined will achieve that (well, we'll test it and see how it goes).