Meta is extending its effort to make social-media recommendations feel less like a black box and more like a negotiable setting.

Earlier this year, the company introduced “Your Algorithm” on Instagram, then followed with the comparable “Dear Algo” tool on Threads. Facebook is now next in line: its version of Your Algorithm will allow people to add and manage subjects they want to see across the main Feed and Facebook Reels.¹

The rollout gives users several routes to reshape what appears on screen. From a post’s three-dot menu, they can select “interested” or “not interested”; those options may also appear directly beneath posts. Meta says the system can ask follow-up questions, enabling people to explain in their own words what they want more—or less—of.

The company’s central claim is speed. “You can easily tell us, in your own words, what you want to see more or less of, and your recommendations will update in real time,” Meta says.¹ That is a notable shift from the familiar model of quietly training the feed through likes, follows and watch time.

Behind the new controls is Meta’s broader push to use large language models in ranking. The company argues that LLMs can account for “interests, context, and intent behind natural language,” making the algorithm something users can effectively talk to—and one that adapts within the same session.¹

For users, the promise is more agency over an often opaque feed. For Meta, it is also a new way to refine personalization: not by abandoning algorithmic ranking, but by inviting people to help direct it in increasingly explicit terms.