I'm thankful for "Like" and "Share" being positive words. It would be much worse if they were "Dislike" and "Shame". The reinforcement strategy behind AI recom

I'm thankful for "Like" and "Share" being positive words. It would be much worse if they were "Dislike" and "Shame". The reinforcement strategy behind AI recommendations (on Facebook, YouTube, Spotify) is to continue showing more similar content. This leads to extremism. To encourage diversity, we the programmers need to intentionally change the algorithm to present opposing views. Mix some 80s disco in with that punk rock. Remind people that there is another perspective. We can also encourage "thank you"s - actively creating backlinks so that people know where the thought came from originally, and can learn more. Hacker News has a great system of self-moderation where downvotes are allowed but only after the user has received 1000 upvotes to establish their credibility. First contribute well, then earn the right to criticise. I hope we can find ways to translate this model into other domains, even real life.