# Algorithmic Curation Algorithmic curation is the automated selection and ranking of content based on predicted user interest. Rather than showing everything chronologically, platforms like Facebook, YouTube, TikTok, and Twitter use machine learning to decide what each user sees. Algorithms optimize for engagement (clicks, watch time, shares), which often means showing emotionally charged, confirming, or outrage-inducing content. This is the mechanism behind [[Filter Bubbles]]: by personalizing feeds to maximize engagement, algorithms inadvertently narrow exposure and amplify bias. [[Tristan Harris]] and the Center for Humane Technology argue these systems exploit psychological vulnerabilities. Alternatives include chronological feeds, transparency about ranking factors, and "bridging" algorithms designed to show cross-cutting perspectives rather than maximize engagement. ## How It Works | Step | Process | |------|---------| | Data collection | Track clicks, time, shares | | Prediction | Model what user will engage with | | Ranking | Sort content by predicted engagement | | Feedback loop | Engagement confirms predictions | ## Human-sourced alternatives Another way out is to replace predicted engagement with explicit human recommendations. [[HackerTube]] (2026) only shows YouTube videos that [[Hacker News]] commenters recommended, and its "For You" feed runs from state in the browser, without accounts. The pool is much smaller, but every item passed a person who chose to recommend it. [[HackerTV]] (2026) goes one step further and drops personalization altogether: the same videos run as 24/7 TV channels, and everyone sees the same program at the same time. ## References - https://en.wikipedia.org/wiki/Algorithmic_curation ## Related - [[Filter Bubbles]] - [[Persuasive Technology]] - [[Attention Economy]] - [[Tristan Harris]] - [[HackerTube]] - [[HackerTV]]