The Echo-Chamber of Your Social Media Feed
summary
Claim: recommendation pipelines on Google YouTube and Facebook concentrate attention and can radicalize rhetoric by surfacing what clicks not what’s comprehensive. Explains crawling, 20+ million new texts per day, indexing into barrels and a lexicon hit list, and quality vs relevance scoring with PageRank math (matrix iteration, average PR = 1, dampener d = 0.85). Shows advertiser bidding via Adwords keyword auctions and thousands of relevance signals so personalized click‑through rates amplify articles into echo chambers.
AI-written summary · about
chapters (7)
- 0:00Social media algorithms intro
- 2:50Search engine indexing
- 8:30Data to personalized results
- 18:30Rhetoric and click-through rates
- 23:46Political echo chambers
- 28:02Ethics of search engines
- 30:56Q&A: Improving algorithms
related talks
44mHacking the Algorithm: Survive the Era of Digital Cult Worship
19mPopping the Filter Bubble
23mML-Powered Search & Recommendation 101
03mServing Billions of Personalized News Feeds with AI
32mPlatform-supported Auditing Of Social Media Algorithms For Public Interest
56mWelcome to the Machine
06mHilary Mason interviewed at Web 2.0 Summit 2011
16mMaking recommendations explainable