Threads’ New ‘Dear Algo’ AI Feature Lets You Personalize Your Feed

Dear Algo: Threads' AI Lets You Command Your Feed

Threads just gave you a direct line to its algorithm. The Meta-owned platform launched "Dear Algo," an AI-powered feature that lets users temporarily reshape their feed by posting public requests. Type "Dear Algo, show me more posts about hiking trails" and your feed adapts for 72 hours—no settings menus or buried preferences required. Announced February 11, 2026, this marks one of the first mainstream attempts to make algorithmic personalization conversational, transparent, and surprisingly social. But the public nature of requests raises fresh questions about privacy in an era where users increasingly demand control without exposure.

Credit: Jakub Porzycki/NurPhoto/ Getty Images

How Dear Algo Actually Works

Using Dear Algo requires zero technical skill—just a public post. Users type "Dear Algo" followed by a clear instruction like "show me less political commentary" or "more indie film recommendations." The AI parses the request in real time, adjusting content ranking signals across Threads' recommendation engine. Unlike permanent interest selections buried in app settings, these commands create a temporary "preference bubble" lasting exactly three days. After 72 hours, the feed gradually reverts to its baseline algorithmic behavior unless refreshed with a new command.

Meta designed the feature to handle nuanced phrasing. Early testing shows the AI distinguishes between "less celebrity gossip" and "no celebrity content whatsoever," applying proportional weighting rather than binary filters. The system also cross-references request language with existing user behavior—someone who regularly engages with cooking content but asks to "see less baking" might still receive general food posts while deprioritizing sourdough tutorials specifically. This contextual awareness prevents overcorrection that could sterilize feeds entirely.

Why Public Requests? The Community Experiment

Here's where Dear Algo diverges radically from conventional personalization: every request lives as a public post visible to followers and discoverable through search. Meta frames this transparency as intentional social design. When users spot a compelling Dear Algo command—say, "Dear Algo, show me urban gardening tips in small spaces"—they can tap a dedicated "Adopt This Request" button to apply identical parameters to their own feed instantly.

This transforms individual preference-setting into a collaborative discovery mechanism. During the NBA All-Star weekend, thousands of users posted "Dear Algo, show me live game reactions" requests. Those commands became de facto topic hubs, letting newcomers tap into real-time conversation streams without knowing specific hashtags or accounts to follow. Meta's product team describes this as "algorithmic empathy"—letting users borrow the refined feeds of others who share their momentary interests.

Still, the public model creates friction for privacy-conscious users. Someone seeking mental health resources or exploring sensitive topics may hesitate to broadcast those needs publicly. Meta addressed this partially by excluding Dear Algo posts from appearing in users' main profile grids—they live only in the algorithmic feed and search results—but the fundamental visibility remains. Competing platforms have historically treated feed preferences as strictly private data; Threads is betting social utility outweighs individual discretion.

Dear Algo vs. Traditional Feed Controls

Most social platforms offer blunt instruments for feed management: unfollow buttons, "Not Interested" taps, or manual account muting. These tools react to content already seen rather than proactively shaping future delivery. Dear Algo flips this paradigm by enabling anticipatory customization—telling the algorithm what you want before it makes mistakes.

The three-day expiration also solves a longstanding pain point: life-stage relevance. Traditional interest selections often become outdated as users' priorities shift. You might care deeply about wedding planning for six months, then never want to see another bouquet photo again. Permanent preferences create clutter; temporary commands acknowledge human fluidity. Early analytics suggest users issue 2–4 Dear Algo requests weekly, with spikes during cultural moments like award seasons or major sporting events—proof that personalization needs aren't static.

Critically, Dear Algo doesn't replace Threads' core algorithm. It layers temporary signals atop the existing engagement-based ranking system. Posts from close connections and verified creators still receive baseline priority; Dear Algo simply adjusts topic weighting within those constraints. This prevents users from accidentally isolating themselves in hyper-niche bubbles while granting meaningful topical control.

Privacy and the Algorithmic Social Contract

The public nature of Dear Algo requests inevitably surfaces privacy considerations. While Meta states request text isn't used for ad targeting beyond the immediate feed adjustment period, the visibility itself creates data exhaust. A cluster of "Dear Algo, show me fertility treatment journeys" posts in a geographic area could theoretically reveal community health trends to third parties scraping public data.

Threads counters that users already broadcast interests through follows, likes, and shares—the platform merely makes preference declaration explicit. But explicitness changes the calculus. Liking a post implies passive interest; publicly commanding an algorithm signals active intent, which carries different psychological weight. Digital rights advocates note this represents a new category of behavioral not what users did, but what they asked for.

For now, Meta has implemented two safeguards: Dear Algo posts can't be quoted or replied to (limiting viral spread of sensitive requests), and users receive a confirmation modal before posting that clearly states "This request will be public." Whether these measures satisfy evolving privacy expectations remains untested at scale—especially as regulators globally scrutinize algorithmic transparency frameworks.

Why This Matters Beyond Threads

Dear Algo signals a broader industry pivot toward explainable, interactive algorithms. For years, social platforms treated recommendation engines as black boxes—users received content without understanding why. Regulatory pressure in the EU and growing user demand for agency are forcing change. Threads' approach doesn't fully open the algorithmic hood, but it offers a steering wheel instead of just brakes.

The feature also challenges assumptions about AI personalization requiring complex interfaces. Instead of dashboards with sliders and toggles, Threads leverages natural language—the most intuitive interface humans possess. This lowers barriers for non-technical users while generating rich training data for Meta's AI systems. Every "Dear Algo" request becomes a labeled dataset point connecting human intent to content categories.

Competitors are watching closely. While X and Bluesky offer chronological feed options and basic filtering, neither has embraced conversational AI for real-time feed shaping. Threads' gamble is that users prefer active co-piloting with algorithms over either passive consumption or manual curation. Early engagement metrics reportedly show 38% higher session duration among Dear Algo users—a metric that could reshape platform design philosophy industry-wide if sustained.

Practical Tips for Using Dear Algo Effectively

To maximize Dear Algo's utility without oversharing, specificity beats vagueness. "Dear Algo, show me sustainable fashion brands under $100" yields better results than "more eco-friendly shopping." The AI responds best to concrete nouns and clear parameters.

Users should also treat the three-day window strategically. Issue requests aligned with temporary life phases: exam season, travel planning, or event coverage. Avoid permanent lifestyle categories better handled through standard follows. And remember—requests compound. Posting "Dear Algo, show me more photography tips" followed two days later by "Dear Algo, show me landscape photography specifically" creates layered refinement without canceling previous signals.

For privacy-conscious users, creating a secondary "interest exploration" account solely for Dear Algo experiments offers a sandboxed approach. Since requests apply only to the posting account, this lets users test niche topics without linking them to their primary identity.

The Road Ahead for Conversational Algorithms

Dear Algo represents more than a Threads feature—it's a prototype for how humans might negotiate with AI systems across digital life. Imagine telling your streaming service "show me 90s sci-fi with practical effects" or instructing your news app "prioritize local infrastructure updates this week." The pattern—natural language commands creating temporary preference states—could extend far beyond social feeds.

Meta hasn't announced plans to expand Dear Algo to Instagram or Facebook yet, but Threads serves as its innovation sandbox. If user adoption and satisfaction metrics hold, expect similar paradigms to permeate Meta's ecosystem by late 2026. The bigger question is whether competitors will adopt comparable transparency or double down on opaque personalization. As algorithmic literacy becomes a mainstream concern, platforms face a choice: build walls around their AI or hand users keys.

For now, Dear Algo offers something rare in social media: a moment of genuine agency. It won't solve every feed frustration, and its public model won't suit every user. But in an industry long criticized for manipulative design, giving people a voice—even a temporary, visible one—in how algorithms serve them feels like progress worth watching. Your feed has been listening all along. Now, finally, it's learning to respond.