For the past year, Threads users have had one major complaint: “I want control over what I see.”
Unlike the reverse-chronological bliss of the early Twitter days, Threads has always relied on algorithms to serve you “recommended” posts from people you don’t follow. It’s been a source of friction for users who feel at the mercy of the machine.
Now, Meta is finally handing over the keys.

This week, Threads began rolling out a new AI-driven feed customization feature that allows users to actively tweak the algorithm powering their “For You” page. But this isn’t just a simple “Show more/Less” button. It represents a fundamental shift in how social platforms are approaching personalization.
Here is what the update actually does, why it matters, and what it signals for the future of social media algorithms.
- What the New Feature Actually Does
Previously, training your Threads algorithm was a passive activity. You liked a post, you followed a user, or you silently scrolled past. The AI watched, but it never explicitly asked what you wanted.
The new update introduces on-device AI controls that let you adjust the weight of the algorithm in real time.
Key Capabilities:
- Adjusting “Interest” Sliders: Users can now tell Threads to show more or less content related to specific topics (e.g., “More Photography,” “Less Politics”).
- Preferred Account Boosting: You can designate specific accounts whose content you want prioritized, even if they aren’t viral.
- Real-Time Feedback Toggles: If a recommended post misses the mark, you can flag it with context (e.g., “Not Interested in This Topic” vs. “Not Interested in This Creator”).

Unlike the old system where feedback was a silent whisper, this new feature is a loud, direct command to the algorithm.
- Why Meta Is Doing This Now
This move is not purely altruistic; it is a direct response to user retention data.
The “Empty Room” Problem: Threads grew rapidly at launch, but keeping users inside the “For You” feed has been a challenge. When users feel they have no control, they don’t fight the algorithm—they leave the app.
The TikTok Effect: TikTok’s algorithm is famously opaque, yet addictive. However, users are growing fatigued by the lack of transparency. Threads is betting that Explainable AI will become a competitive advantage.
The Decentralization Vibe: As platforms like Bluesky gain traction by offering algorithmic choice, Meta is preemptively neutralizing the argument that it is a “black box” dictatorship.

- The AI Mechanics Under the Hood
While users see simple toggles and sliders, the backend is surprisingly complex.
This feature relies on lightweight on-device machine learning models. Instead of sending every “Show less” click back to Meta’s massive data centers to retrain the main model, the app adjusts your feed ranking locally on your phone.
Why this matters for privacy:
- Immediate feedback doesn’t require your data to be uploaded instantly.
- It reduces the “creepiness” factor where you mention something out loud and then see an ad for it.

It’s a hybrid approach: The global model predicts what most people want, but the on-device AI overwrites it based on your specific adjustments.
- What This Means for Marketers and Creators
If you are a brand or creator on Threads, this update fundamentally changes the rules of engagement.
- The “Viral or Bust” Model Softens:
Previously, the algorithm was a black box. You either caught lightning in a bottle, or you were invisible. Now, users can manually boost creators they love. Loyalty now trumps virality. Building a small, dedicated audience that toggles your “Preferred Account” switch is now more valuable than chasing random likes. - Topic Authority Becomes an Asset:
If users tell the algorithm they want “More SEO tips” or “More vegan recipes,” the platform needs to serve that. Creators who are clearly associated with a specific niche will be surfaced more aggressively by the AI when users tweak those sliders. - Community Management > Broadcasting:
The “Show less from this account” button has always existed. Now, the “Show more” button is equally powerful. Marketers need to shift from campaignsto continuous valueto earn that affirmative user action.

- The Bigger Picture: The Rise of the “Tunable” Feed
This Threads update is part of a larger industry pivot away from the “Omniscient Algorithm.”
For the last decade, tech companies operated under the assumption that their AI knew better than the user. The attitude was: “We have the data, trust us.”

That era is ending.
Users are demanding agency. They don’t just want to be fed; they want to steer. By introducing tunable AI feeds, Threads is acknowledging that the perfect algorithm doesn’t exist—but a transparent one does.
The Bottom Line
Threads is betting that the future of social media isn’t a smarter algorithm, but a quieter one.
One that listens to your explicit commands rather than just guessing based on your lingering gaze.

For users, this means less noise. For marketers, it means the death of spray-and-pray content. If your audience has the power to manually turn down the volume on your industry, you better make sure they want to keep listening.
Have you tried the new Threads AI controls yet? Are they making your feed better or just more boring? Let me know in the comments.
