From Hacker News to TikTok - How Algorithms Learned to Hook Us

The podcast examines the development and consequences of social media algorithms, comparing early platforms like Reddit and Hacker News, which used community-driven, gravity-based models to balance content freshness and popularity, with modern platforms such as Instagram and TikTok, which leverage machine learning to create hyper-personalized feeds. It highlights how these algorithms can unintentionally amplify divisive or harmful content by capitalizing on psychological tendencies, such as engagement with emotionally charged or controversial material, as illustrated by Reddits "controversial" sorting mechanism and Facebooks 2017 shift toward prioritizing "meaningful social interactions," which inadvertently promoted polarizing content. The discussion also contrasts TikToks real-time, data-dense algorithm with YouTubes batch-processing approach, while addressing broader concerns about algorithmic addiction, mental health risks from content escalationsuch as exposure to extreme or harmful materialand the challenges of reconciling user engagement with ethical and psychological well-being.