X Open-Sourced Their Algorithm—So I Analyzed It
https://github.com/hal-shin/x-algorit...
X recently open-sourced its recommendation engine—now powered by Grok, its in-house AI model. Instead of guessing how the algorithm works, I analyzed the actual code to understand how posts are selected, filtered, scored, and ranked in the feed.
In this video, I break down:
• How X chooses in-network vs out-of-network posts
• How Grok scores posts using multiple engagement signals
• Why diversity and posting cadence matter more than volume
• The four concrete actions you can take to improve reach and engagement
After applying these insights myself, I saw noticeable increases in impressions and profile visits within weeks.
If you’re trying to grow on X and want to understand how the platform actually decides what gets shown, this video will give you a clear, code-level mental model—without the hype.
0:00 - Intro
0:51 - Algorithm Explained
2:49 - First Tip
3:46 - Second Tip
4:17 - Third Tip
5:08 - Fourth Tip
6:12 - Outro
Company: https://overseed.ai/
X/Twitter: https://x.com/_halshin
LinkedIn: / halshin
Twitch: / halshin