In the last few years, I have consumed news/content mostly via Social media, and I keep noticing the redundancy and narrowed recommendations of (FB, Twitter) algorithmic timelines. Since I didn't want to use other apps to consume news, I initially tried to manually personalize my timelines. After several attempts, FB remains a black box, without easy ways to adjust/influence the "algorithmic" timeline. I started using Lists on Twitter and manually explore new content. I have tried to hack a tool to diversify the content, and have better success with reddit/twitter APIs, FB has strict API. To my surprise, I had better results using RSS feeds + script to cluster content -> slack summaries periodically, (Old) RSS feeds helped to passively monitor good content.<p>It's almost 2021, what social media platforms do you still use?. Is it just me, or do you notice such problems with algorithmic timelines.<p>As a user, I just want to have a better algorithmic timelines, with no harmful personalization that cage us in our filter bubble.<p>As a (ML) developer, I want to have a better API/system that can enable me to customize or deploy my recommendation/news trending models; why do we still need to use one "recommendation model for every user on social platform" that is unknown for end user, I believe user/community should have flexibility to deploy their own recsys models.