/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

X makes the source code for its For You timeline available on GitHub and adds tools to let users see if X's ranking systems “shadowbanned” them

X is significantly expanding its open-source codebase, which includes the app's “For You” algorithm and its core ranking engine …

TechCrunch Sarah Perez

Context & Ripple Effects

X has been progressively exposing how its recommendations work: Twitter released only part of its For You code in 2023, while X published a core recommendation system built around a Grok-based transformer in January. The new release extends that transparency effort from the core engine to the timeline code and adds a user-facing accountability layer.

The disclosure sits alongside a narrower boundary on data use: X previously barred third parties from using its API or content to train foundation models. Code can therefore become more inspectable without opening the underlying platform data for outside model development.

First-order effects

  • X users gain a tool to check whether the platform's ranking systems limited their distribution, turning a previously opaque complaint into a product-level status check.
  • Researchers and developers can inspect the For You timeline implementation on GitHub, expanding on X's earlier release of its core recommendation system.

Second-order effects

  • X will face more specific scrutiny over whether the behavior of its published ranking code matches users' observed reach and the results returned by its new checking tools.
  • Developers can study X's recommendation architecture, but X's API restriction keeps them from using X content to train competing foundation or frontier models.

Third-order effects

  • If X continues pairing code releases with user diagnostics, recommendation transparency shifts from periodic source-code disclosures toward ongoing, user-verifiable platform accountability.
  • The combination of open code and restricted data access points to a split model of AI transparency: systems may be inspectable while the data needed to reproduce or train them remains controlled by the platform.

The trend: Platforms are separating algorithmic transparency from data and model-access rights, opening more implementation detail while retaining control over the inputs that confer competitive leverage.

Discussion

  • Moushmi Banerjee Moushmi Banerjee on linkedin
    X open-sourced its feed algorithm  —  The xai-org/x-algorithm repo shows how the “For You” feed is built end-to-end:  —  • Candidate generation (in-network + discovery) …
  • Ritvik Jhawar Ritvik Jhawar on linkedin
    X just pushed its biggest open-source update yet to the algorithm running the “For You” feed.  —  The buildup made it interesting. …
  • @aendra.com Ændra on bluesky
    They published the algorithm about two months ago it seems:  —  github.com/xai-org/x-al...  How useful this is for interrogating bias without firehose access to the platform is another question.  Whether it allows Musk to manually promote posts is another entirely.
  • @xbusiness @xbusiness on x
    Today X is taking a major step towards unprecedented transparency. By expanding the open-source For You ranking code and launching a new tool that clearly shows visibility-limiting labels, X is giving people a clear view into how posts are ranked and filtered. This is important
  • @xopensource @xopensource on x
    https://x.com/...
  • @gergelyorosz Gergely Orosz on x
    This part in the middle pretty telling: “This is important for advertisers and brands” Sounds like X is struggling to get advertisers at the rate they want to be a money maker. I dont see how this would help them tho. Advertisers want ROI on ad spend comparable to Meta/Google
  • @iterintellectus Vittorio on x
    they actually posted the weights of the algorithm [image]