CoRecursive: Coding Stories From Hacker News to TikTok - How Algorithms Learned to Hook Us
36 snips
Mar 2, 2026 They trace the evolution of ranking systems from simple upvotes to modern personalized feeds. They explain how metrics like 'meaningful' interactions and watch time warped platforms toward divisive, attention-grabbing content. They unpack TikTok’s rapid, dense signal learning and how short-form feeds narrow interests. They end by exploring practical ways to resist or reset recommendation engines.
AI Snips
Chapters
Transcript
Episode notes
News Feed Exploits Social Comparison
- Facebook's News Feed tapped innate social comparison—gossip and social standing—not complex code, to keep people engaged.
- EdgeRank simply ranked posts by closeness, interaction frequency, type, and recency to surface friends' updates.
Metric Shift Turned Facebook Toward Rage
- Facebook optimized for 'meaningful social interactions' (comments, shares, reactions) in 2017, which raised engagement but amplified anger.
- Optimizing that metric rewarded divisive posts because anger drove comments and shares.
Growth Incentives Locked In Harmful Choices
- Facebook couldn't easily roll back the engagement optimization because reducing divisive content risked slowing growth and hurting revenue and stock.
- The tradeoff created a company-level incentive to tolerate harmful side effects.