Twitter changed its algorithm last week. Most of what you read about it was half right. We read the open-source code line by line and compared 21 October with 29 October. Three things changed: 1. Equal trimming When your feed has more posts than the 2,800 the ranking model can score, every source is now cut by the same share. Before, the last source in line lost the most, and that's where posts under 2 hours old sat. 2. A warning label Posts labelled low quality or misinformation now carry a notice. Everyone sees it, followers included. On low-quality posts it blocks likes, replies, reposts, bookmarks and shares. 3. The AI slop label Accounts flagged as "LLM slop" get a spam label that hides them from people who don't follow them. It now lasts 7 days instead of 30. What didn't change: not one ranking weight. The last changes were on 13 and 14 October. What to do this week: → Post originals when you can be around for the first 2 hours → Create posts people read to the end → If you use AI, edit until the post reads like you The third change is the one nobody covered. It's now in the code: generic AI content costs you reach. Full breakdown, with links to every file in the code, in the first comment.