You might be surprised at how much curation goes into those seemingly mechanical lists. I try to exclude the cases that aren't interesting, and unfortunately it's pretty hard to write software the distinguishes the interesting from the uninteresting. In fact, if I knew how to write such software, I could automate my job!
There's a bit of a difference there in that an automated list that's half as good as what you'd manually curate is nearly as useful as your version, especially if people know it's automated. This isn't at all the case for most other aspects of moderating that depend on human input. The previouslies don't really have to be dang-quality.
2019 https://news.ycombinator.com/item?id=19132724
2018 https://news.ycombinator.com/item?id=18230258
2016 (1 comment) https://news.ycombinator.com/item?id=12749342
2014 https://news.ycombinator.com/item?id=7726496