The paper itself [1] and a Wired article covering it [2], were received very poorly on HN when they were originally posted before. [1] https://news.ycombinator

The paper itself [1] and a Wired article covering it [2], were received very poorly on HN when they were originally posted before. [1] https://news.ycombinator.com/item?id=9399090 https://news.ycombinator.com/item?id=9399090 [2] https://news.ycombinator.com/item?id=9398399 https://news.ycombinator.com/item?id=9398399 I honestly didn't expect that. My impression was that people are generally afraid that this will be used as a way to auto-ban people just for disagreeing with popular sentiment. I think what people didn't understand was that this analysis is heavily based on community and admin actions, primarily banned users and deleted posts. While there is some analysis of the comments themselves, like readability and word similarity, this is not some SkyNet auto-ban bot. They are predicting how likely a user is to be banned, based primarily on the community's response to that user. My point is that if systems like this are used to flag or ban people for disagreeing with the group think, then the problem is with the community, not the algorithm. Edit: typos and clarity