X Releases Algorithm Source Code, Reveals Post Ranking Weights

Negative interactions limit a post's visibility sharply. A predicted report carries a weight of 468 in the negative direction – 468 times the positive weight of a like. A predicted mute carries a negative weight about 118 times a like. A "not interested" interaction is about 86 times as large in the negative direction, and a block is about 62 times as large.

The system's weighting structure suggests that inflammatory content designed to provoke angry replies – often called "ragebaiting" – may not be a reliable path to growth. The company did not comment on whether that outcome was intentional.

Also released on Thursday was a pilot feature called "Under the Hood." The feature lets participating users see aggregate information about labels applied to their account and posts that can affect visibility. The source code shows that X's systems classify content and accounts across categories including spam and adult material. Those labels feed into a separate system that determines whether a post is shown to users.

X said the feature is active only for a "randomized test group of eligible accounts" and that a broader rollout will depend on the feedback the company receives.

Thursday's release is the latest in a series of open-source moves by X. In January, Musk called the algorithm "dumb" and pledged to make the system more transparent. "At least you can see us struggle to make it better in real-time and with transparency," Musk wrote on X. "No other social media companies do this."