# Facebook adds an AI-tunable "Your Algorithm" feed control Meta is extending Instagram's Your Algorithm topic controls to Facebook and says large language models now play a larger role in ranking what users see, with changes it claims take effect within the same session. Canonical URL: https://freelancenews.online/news/facebook-adds-an-ai-tunable-your-algorithm-feed-control-a579d945 Published: 2026-10-13T19:17:15.210Z Updated: 2026-10-13T19:17:15.210Z Source published: 2026-10-13T17:01:31.000Z Event date: Not established Review status: source-reviewed Review method: Automated comparison against retrieved source text; not independent fact-checking. ## Report Meta is bringing the "Your Algorithm" feature it previously introduced on Instagram to Facebook, giving users a direct way to add and manage the topics that appear in their feed and in Facebook Reels. The company announced the change as part of a broader set of what it describes as AI-powered tools for customizing what people see on the platform. The mechanism is topic-level rather than post-level. According to the source, users can add and manage topics they want to see, and can mark individual topics as "interested" or "not interested." Those same two options are also reachable from the three-dots menu on individual posts, and Meta says the buttons may additionally appear underneath posts in the feed. Meta's stated behavior after an update is immediate: the company says recommendations "respond right away" once a user makes changes through the feature. It also says it may ask follow-up questions to get more detail about what a user liked or disliked, and that users can describe in their own words what they want to see more or less of, with recommendations updating in real time. The feature is not new to Meta's product line. Your Algorithm launched on Instagram earlier this year, and the company also introduced a comparable "Dear Algo" feature on Threads. The Facebook rollout therefore extends an existing control pattern across a third surface rather than introducing a novel concept. The more consequential technical claim is about ranking. Meta says it is using large language models more heavily in its ranking systems, arguing that LLMs can incorporate the interests, context and intent behind natural language, which the company frames as "an algorithm you can actually talk to." In Meta's description, telling the system what you want more or less of produces recommendations that adapt within the same session to reflect current preferences. That framing matters for how the control is likely to behave in practice. A topic list is a persistent preference signal, while the natural-language and in-session adaptation described by Meta is closer to a conversational adjustment layer. The source does not explain how the two interact, how long an in-session adaptation persists, or whether stated preferences override engagement signals the ranking system already collects. For freelancers, designers and developers who rely on Facebook for client discovery, community groups or distribution of their own work, the practical value depends on whether topic controls actually shift what surfaces. A designer who wants more of a specific craft or tooling topic can now name it directly instead of relying on inferred interest, but the source provides no measurement of how much the feed changes as a result. There is also a maintenance cost implied by the design. Topic lists and interested/not-interested signals are explicit inputs, which means they can drift out of date as a person's work changes. The follow-up questions Meta says it may ask suggest the company expects users to refine preferences over time rather than set them once. The announcement sits inside a wider pattern of platform-level personalization controls, but the evidence here is limited to Meta's own description of the feature. No independent testing, third-party measurement or user data is supplied, and the source does not include rollout dates, regional availability, or whether the feature reaches all Facebook users at once. One detail worth separating from the headline: the source material also contains unrelated items about Apple's "Welcome Home" event, including a smart display, a new Apple TV 4K with built-in microphones for Siri AI, and a home collection with a video doorbell and smart lock. Those items are not part of the Facebook algorithm change and should not be read as connected to it. The honest limitation is that almost everything here is a vendor claim. Meta says recommendations respond right away and adapt within the same session, but the supplied evidence contains no method, sample or external verification for those statements. Treat the real-time behavior as described by Meta, not as an established result. For this audience, the reasonable interpretation is that Facebook now offers a more explicit lever over feed composition than passive engagement alone, and that lever is worth testing against actual outcomes such as reach on posts, group visibility and discovery of relevant work. Whether it delivers a meaningful change in what freelancers and developers see remains unverified until users compare their own feeds before and after adjusting topics. ## Key points - Meta is extending the Your Algorithm topic-control feature from Instagram to Facebook, covering the main feed and Facebook Reels. - Users can mark topics as interested or not interested, including from the three-dots menu on individual posts, and Meta says it may ask follow-up questions. - Meta says it is using large language models more in ranking and claims recommendations adapt in real time within the same session. - Your Algorithm previously launched on Instagram earlier this year, and Meta also shipped a similar Dear Algo feature on Threads. - No rollout dates, regional availability or independent verification of the real-time behavior are provided in the source. ## Practical implications — editorial interpretation Editorial interpretation: freelancers, designers and developers who depend on Facebook for discovery or client reach can now state topic preferences explicitly instead of relying only on inferred interest, which is worth testing against real outcomes such as post reach and group visibility. The source does not establish how much the feed actually changes, so treat the control as an experiment to measure rather than a guaranteed distribution fix. ## Limitations and unknowns All behavioral claims come from Meta's own description; the evidence includes no independent testing, sample, method or user data. Rollout dates, regional availability and whether the feature reaches all Facebook users are not stated. The source does not explain how topic lists interact with the natural-language, in-session adaptation, or how long those adjustments persist. Unrelated Apple product items appear in the same source and are not connected to this change. ## Sources - [1] theverge.com: Facebook will let you customize your feed algorithm using AI https://www.theverge.com/tech/1010344/meta-facebook-your-algorithm-ai Retrieved: 2026-10-13T19:17:01.898Z ## Claim references - Meta is bringing the Your Algorithm feature to Facebook, letting users add and manage topics in their feed and Facebook Reels. [source 1] - Meta says recommendations respond right away after users update their preferences. [source 1] - Users can reach the interested and not interested options from the three-dots menu on individual posts, and Meta says it may ask follow-up questions. [source 1] - Meta says it is using large language models more in its ranking systems and that recommendations adapt within the same session. [source 1] - Your Algorithm previously launched on Instagram earlier this year, and Meta also launched a similar Dear Algo feature on Threads. [source 1]