Everybody has their pet peeve about things that are "too popular" on Hacker News, see this classic:
http://ontology2.com/essays/HackerNewsForHackers/
The trouble is people don't agree on what those are!
The next article in that series
http://ontology2.com/essays/ClassifyingHackerNewsArticles/
never went viral because it was too emotionally positive. (E.g. rather than complain about the danger of A.I. in the hands of Google/Facebook/Baidu why not apply it to your own needs?)
I get sick of the all the "I competed in Kaggle and all I got was this lousy T-Shirt" posts that you see on HN, but HN readers who look at that article superficially might not realize that I am solving my own problem, facing down the issues of applying ML to a business problem which has not been carefully curated.