Tech & AI
How recommendation algorithms decide what billions of people see
12 min
The feeds that shape modern attention are driven by recommendation engines optimizing for engagement. Explore how these systems actually work, what they optimize for, and the consequences of handing curation to machines.
Listen on the app, request early access:
Show notes
Recommendation systems use a two-stage architecture to filter billions of items into a small ranking pool.
YouTube shifted its primary metric to expected watch time to prioritize retention over simple clicks.
Algorithms monitor micro-decisions like scrolling speed and dwell time to predict user interest.
The exploration-exploitation trade-off introduces new topics like woodworking to prevent user boredom.
Creators often homogenize their visual style using high-contrast thumbnails to satisfy algorithmic preferences.
New legal mandates in the European Union require platforms to offer non-profiled chronological feeds.
In this episode
- 1Intro1 min
- 2The Architecture of the Feed3 min
- 3The Currency of Engagement3 min
- 4The Feedback Loop and the Echo Chamber3 min
- 5The Future of Curation2 min
- 6Outro1 min
Sources
- Understanding Social Media Recommendation Algorithms
- How Algorithms Actually Decide What You See
- How do recommender systems work on digital platforms? | Brookings
- Recommenders under fire: Echo chambers, disinformation, and what really happens inside the ranking code – TechGrid Media
- From Hacker News to TikTok - H… - CoRecursive: Coding Stories - Apple Podcasts
- Why Algorithms Show Us What We Claim Not to Want | Psychology Today
- How TikTok’s algorithm is affecting you, from the content you see to what you can do to regain control | Technology | EL PAÍS English
- How Recommendation Algorithms Actually Work - YouTube
- How Platforms Decide What You See — Ranking & Recommendation Systems Explained
- How Recommendation Algorithms Work: The Science of Your Feed - InsideTheSystem
- Algorithms
- Twitter's Recommendation Algorithm
- News Feed ranking, powered by machine learning - Engineering at Meta
- Reverse Engineering the YouTube Algorithm - Gal Lahat
- Scaling the Instagram Explore recommendations system - Engineering at Meta
- Deep Neural Networks for YouTube Recommendations
- Understanding Social Media Recommendation Algorithms
- On YouTube’s recommendation system - YouTube Blog
- Recommending What Video to Watch Next: A Multitask Ranking System
- Design a Recommendation System (Netflix / YouTube / TikTok) - The HLD Handbook
- The YouTube video recommendation system
- Deep Neural Networks for YouTube Recommendations
- YouTube performance FAQ & Troubleshooting - YouTube Help
- How YouTube Finds Your Next Video in Milliseconds
- ML System Design Case Study: YouTube Video Recommendation Engine - ML Digest
- How YouTube Decides Which Video to Show 2 Billion People - The Algorithm Fully Explained - AIWala News
Fylom generates episodes like this on any topic you're curious about.
Fylom episodes are researched and written by AI. Automated checks help catch inaccuracies, but episodes aren't reviewed by a human and AI can still get things wrong. Treat them as a starting point, not a source of record — more in our accuracy disclaimer.