Facebook is adding a feature that lets people directly tell its recommendation system which topics they want to see more or less often. The Verge reported on October 13 that the tool, called Your Algorithm, will allow users to add and manage interests affecting both the main Facebook feed and Facebook Reels. Meta says recommendations will respond immediately when a person marks a topic as interesting or uninteresting.

The controls will appear in several places. Users can open the three-dot menu on an individual post to reach Your Algorithm or select an interested or not interested option. Meta may also place those choices beneath posts and ask follow-up questions to understand what a person liked or disliked. The company says people can describe their preferences in their own words, with recommendations updating in real time.

A glowing speech bubble directs preferred and unwanted topic symbols onto separate paths.
Users will be able to describe what they want to see more or less of in their own words.

Meta is presenting the feature as part of a broader set of AI-powered personalization tools. According to The Verge, the company is using large language models more extensively inside its ranking systems because they can interpret the interests, context and intent expressed through natural language. That creates a different kind of input from a like, share or pause: instead of requiring the system to infer a preference from behavior, the user can state one explicitly.

Facebook is not the first Meta service to receive this approach. Your Algorithm launched on Instagram earlier in 2026, while Threads introduced a related feature called Dear Algo. Bringing the concept to Facebook extends conversational feed control across Meta’s largest social platforms and turns natural-language instructions into a common interface for shaping recommendations.

A bright user-control lever sits among many hidden inputs inside a recommendation machine.
Meta has not detailed how much weight direct preferences will carry against other ranking signals.

The announcement does not explain how strongly an explicit request will outweigh the many other signals Facebook uses to rank content, how long a preference will remain active or whether the controls affect areas beyond feed and Reels recommendations. It also does not say whether users will be able to inspect why particular topics were inferred before they changed them. The promise of an immediate response is therefore clear, while the deeper mechanics remain opaque.

Even with those unanswered questions, the feature changes the relationship between a user and the feed. Recommendation systems have traditionally learned from accumulated behavior and offered limited correction through controls attached to individual posts. A conversational layer could make those corrections broader and faster, especially when a person’s interests change temporarily. The test will be whether the system treats direct instructions as durable authority or merely one more signal among many competing for control of the screen.