Musk Open-Sources X Algorithm: Grok Summarizes Interaction Weights
elonmusk · x · 2026-08-14
Elon Musk stated that open-sourcing the 𝕏 algorithm aims to actively seek criticism to improve the fairness and quality of recommendations.
According to a summary generated by Grok after reading all 370,523 lines of code, the algorithm does not assign a 'quality' score to posts. Instead, it guesses what actions a reader might take and multiplies each by a specific weight. The total sum determines the post's value, with the highest scores ranking at the top of the For You feed.
Production weights include: copy link/share (+20), reply on an original post by mutuals (+20), quote (+5), any reply (+5), share via DM (+5), and follow from post (+4).
More from Apps
- Cornell's DebunkBot AI Convinces 25% of Conspiracy Theorists to Change Minds — stevenstrogatz · 2026-08-14
- Customer Support's Future: Multi-Tier AI Agents Escalating to Solve Issues — round · 2026-08-14
- DebunkBot: AI Chatbot Convinces 25% of Users to Abandon Conspiracy Theories — stevenstrogatz · 2026-08-14
- Enterprise AI Adoption Curve: Six Stages of Brand Asset Automation — briannekimmel · 2026-08-14
- Gemini 3.7 Flash Demo: Turning Research Papers into Interactive Simulations — melvinjohnsonp · 2026-08-14
- Open-Source MCP Server Optifeed Radar Checks If AI Engines Recommend Your Products — erdemgezer · 2026-08-14