Your Algorithm Facebook is live. On October 13, 2026, Meta announced a new AI-powered feature that lets Facebook users describe — in their own words — what they want to see more or less of in Feed and Reels. Type “show me more vegetarian recipes but fewer dessert videos,” and Meta says your recommendations update in real time, within the same browsing session.
This is more than a settings page. Meta is putting large language models directly inside Facebook’s ranking systems, so written intent becomes a ranking signal alongside likes, comments, and watch time. If you make money from content — or plan to — the rules of organic reach just changed.
What “Your Algorithm” actually does
The feature is a dedicated space where users add and manage the topics driving their recommendations. The fastest way in is the three-dot menu on any post: it now offers a “Your Algorithm” option plus “Interested” and “Not interested” buttons. After you tap one, Facebook may ask follow-up questions — in plain language — to learn exactly what you liked or disliked.
The technical shift is the real story. Until now, Facebook’s recommendations were built almost entirely on inferred behavior: clicks, watch time, lingering. Now natural-language statements feed the same ranking stack. In Meta’s official Newsroom announcement, the company puts it this way: “LLMs can incorporate the interests, context, and intent behind natural language, unlocking an algorithm you can actually talk to.”
Meta’s example shows the tool summarizing a user’s tastes back to them, recommending topics like music festivals, fitness, and videography — while honoring explicit bans such as “hiking” and “horror movies.” A suggestion box, bolted onto the most powerful recommendation engine in social media.
A rollout two years in the making
Facebook is not the test lab — it is the final stop:
- December 10, 2025: “Your Algorithm” debuts on Instagram Reels in the US.
- February 11, 2026: Threads launches “Dear Algo” (US, New Zealand, Australia, UK).
- April 15, 2026: Instagram extends Your Algorithm to Explore.
- June 10, 2026: Instagram unifies it across Feed, Reels, and Explore in English worldwide.
- October 13, 2026: Facebook gets Your Algorithm plus Interested / Not interested buttons.
Instagram head Adam Mosseri framed the philosophy in June: LLMs can make opaque ranking systems understandable by showing why content appears and letting users state preferences explicitly. Two years of Instagram and Threads testing means the Facebook version arrives battle-hardened.
What Meta hasn’t said
Honest reporting means flagging the gaps. Meta has not published a market-by-market rollout timetable. It has not disclosed how much weight typed requests carry versus viewing habits — whether a stated preference beats a thousand accidental clicks is unknown. It has not said how long a preference sticks, or whether preferences affect ads. Treat this as a stronger input into the recommendation system, not ownership of it. The algorithm still ranks. You just got a steering wheel.
The money angle: reach is becoming steerable
For creators and side hustlers, organic reach is revenue. Here is how to play the new game.
1. Write for stated intent, not just engagement bait
Ranking systems that parse semantic meaning reward content that is easy to classify. A caption that clearly says what it is — “beginner sourdough tutorial: 3 mistakes that kill your starter” — matches a user’s stated intent far better than a vague hook. Audit your last ten posts: could an LLM describe each in one clear topic sentence? If not, your positioning is leaking reach.
2. Prompt skills are now a distribution skill
The same skill you use to get better answers from ChatGPT — articulating exactly what you want — now shapes what billions see. Creators who understand how LLMs parse intent can reverse-engineer discoverability: name your niche the way your audience would type it, and the machine does the matchmaking. Prompt engineering just became a marketing skill.
3. Niche down — sharp positioning wins
When users can say “more woodworking, less politics,” sharply defined creators win the match and generalists blur out. Pick the lane, state it in your bio in plain words, and repeat the topic vocabulary in every post. You are no longer optimizing for the algorithm’s guesses — you are optimizing for your audience’s sentences.
Do this before the weekend
- Test it as a user. Open the three-dot menu on a post, tap “Your Algorithm,” and read what Facebook thinks you are into. Free market research.
- Rewrite your bio and pinned posts with the topic keywords your ideal follower would type.
- Retitle your next five posts so each names its topic in the first line. Measure reach for two weeks against your baseline.
Frequently asked questions
What is Meta’s “Your Algorithm”?
An AI-powered feature, announced October 13, 2026, that lets Facebook users add and manage the topics driving Feed and Reels recommendations — by describing preferences in their own words instead of relying only on likes and watch time.
How do I use Your Algorithm on Facebook?
Tap the three-dot menu on any post and select “Your Algorithm,” or tap “Interested” / “Not interested” on posts. Facebook may ask follow-up questions so you can explain in plain language. Recommendations update in real time.
Does Your Algorithm replace the Facebook algorithm?
No. Meta describes it as an additional signal for its existing ranking systems. The algorithm still ranks and recommends; users now have a direct way to steer it.
Is Your Algorithm on Instagram and Threads too?
Yes. Instagram launched it for Reels in December 2025 and unified it across Feed, Reels, and Explore in June 2026. Threads offers “Dear Algo” since February 2026.
Will my preferences affect the ads I see?
Meta has not said. The announcement covers Feed and Reels content recommendations only.
Keep Reading
Sources: Meta Newsroom (announcement: “Your Algorithm, Your Words: New Ways to Shape What You See on Facebook,” October 13, 2026); TechCrunch’s report on the launch.