Take a look at arXiv.org and tell me out of all of them which papers seem to stand out.
There's too little time in a day and in a life for anyone to become independently aware of all happenings of personal interest. We rely on other to curate what we see, and it happens that many parties vie for our attention. Hence the development of propaganda and news aggregators, like the one you're on now.
I recommend reading Edward L. Bernay's work Propaganda for a more thorough understanding about the role of "PR" in society. It's a good foundation. I actually remember reading another rather good post by Paul Graham on the subject, which you might find interesting.
Sure. But if you're posting to hacker news, post the article itself. No reason to haul a writer in to muddy it up.
EDIT: And to be clear, the idea of curating arxiv is absurd. But so is the idea that Stanford would have anything interesting to say about their own research.
Even on HN, where we never tire of self-praise, I think the article is more helpful to the average reader than the paper itself. What percentage of us are in a position to critically examine the experimental methods? I don't know, but my naive assumption is that a conversational presentation of the publication and its consequences is more welcome than the publication itself. Regardless, the primary source is one hyperlink and university subscription away for all who care to read it.
Of course it's absurd, that was my point. And obviously I disagree that Stanford has nothing of value to say about its researchers' work: I think they've managed well in this case.
I agree it's unsettling. I still haven't formed a strong position on Bernays' normative assertions. If you have any thoughts on the subject, I would be interested to discuss them.
It's still weird to read about a paper via press release. Papers should stand on their own, not require PR.