@dansup Won't that lead to filter bubbles that amplify unwanted stuff like conspiracy theories? IMHO ML inherently leads to exactly that kind of problems.