I changed it from the article title to a shortened version of the URL slug (/well-try-to-help-you-follow-the-police-attacks-on-journalists-across-the-country). That's often a legit source title, along with the HTML doc title, and other places that articles tend to reveal what they're actually about (
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...).
The reason I changed it is that "U.S. police have attacked journalists more than 110 times since May 28" is more baity. First you've got the aggressive verb "attacked". And then you've got the specific facts and figures (100 times, May 28). Why do these things make a title more baity? I don't know, but they are somehow active ingredients in the psychology of titles, which is why headline writers use them. I think of them as sharp edges. If you throw a spiky thing into a crowd, it gets more attention than a beach ball—but the quality of attention is not conducive to a reflective discussion on the internet (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...). People react to the sharp bits.
It's in HN's interest to rewrite baity titles (indeed the site guidelines request it: https://news.ycombinator.com/newsguidelines.html) but this is 1000x more important on inflammatory topics of the moment, like this one. We want HN to discuss this, but we want to support the discussion to stay in the intended spirit of the site: thoughtful, respectful, curious conversation. The material is provocative enough, and a title doesn't need sharp edges to make readers here care about the topic.
To clear up a misconception that sometimes arises: I'm not saying the statement in the sharper title is false. Obviously it's legitimate to use an aggressive verb to describe an aggressive action, and if the facts and figures are true then it's obviously legitimate to make a factual statement with them. As a reader I have zero problem with that title. But the moderator perspective is different: the question for a moderator, on HN at least, is always: what is the prospective effect on forthcoming discussion? Or to put it pseudotechnically: given choice A vs. choice B, what is the diff between the probability distribution of threads that A is a prefix of, vs. threads that B is a prefix of? (This is just a metaphor, but it reflects how we think about this. See https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que... for more.) In the case of these two alternative titles, experience has shown that there's a significant difference in probable outcomes, and titles are by far the biggest influence on threads, so it's potentially a big deal.