The social feed is starting to behave less like a black box and more like a conversation.
For years, platforms gave users a few blunt tools: follow this account, like this post, hide that video, tap “Not Interested” when something felt wrong. But the real decision-making still sat with recommendation systems that interpreted those signals at scale. A new TechCrunch report on user-controlled algorithms points to a sharper shift: Threads, Instagram, TikTok, and Bluesky are all moving toward feeds that users can more directly shape.
That does not mean users are taking full control of the algorithm. But it does mean the old bargain is changing. Instead of asking people to passively react to what the system serves, platforms are beginning to ask what they actually want to see.
Threads makes the algorithm request private
The clearest example is Threads. On June 16, 2026, the platform launched “Your Algo,” a feature that builds on its earlier “Dear Algo” experiment from February. The first version asked users to publish a public post addressed to the algorithm, such as “Dear Algo, show me more posts about podcasts,” to influence what appeared in their feed.
“Your Algo” removes the performance from that process. Instead of posting a public request, users can privately tell Threads they want to see more or less of certain topics. They can also choose how long the instruction should last: one day, three days, or seven days. TechCrunch gives the example of someone asking for more baseball content and less stressful news.
That time limit matters. It treats interest as something temporary, not permanent. A user may want more sports during the playoffs, less politics during a stressful week, or more parenting content for a few days without turning their entire feed into a permanent identity file.
Instagram and TikTok expose the dials
Instagram is moving in the same direction, but through visibility. In early June, it launched “Your Algorithm,” a tool that lets users see and control the topics shaping recommendations across their feed. That is a notable change for a platform whose discovery engine has often felt powerful but difficult to interrogate.
TikTok is also pushing more direct input through its “Manage Topics” controls, which let people adjust how much they want to see from different content categories in the For You feed. The important shift is not that platforms are suddenly giving away the keys. It is that they are admitting users may need more than a dislike button to correct the system.
This is a practical response to a very real product problem. Recommendation feeds work best when they feel uncannily relevant. They also break quickly when the system overreads a short-term behavior. Watch one renovation video, linger on one celebrity drama, or doomscroll one news cycle, and the feed can decide you have become someone else.
Bluesky shows the harder version
Bluesky has approached the problem from a different angle. Its custom feeds let users choose and pin alternative feeds built around different ranking systems, communities, or interests. That is not the same as telling one central algorithm what to do. It is closer to choosing which algorithmic lens you want to look through.
The distinction is useful. Threads, Instagram, and TikTok are giving users more input inside a platform-controlled recommendation system. Bluesky is making feed choice part of the product itself. Both approaches point in the same direction: the feed is no longer just a stream of content. It is becoming a user-adjustable interface.
The prompt is spreading
The unexpected part is that this shift is not limited to social feeds. It rhymes with what is happening elsewhere in consumer tech. Pinterest is turning shopping into something users can ask for. Google is trying to make the smart home respond to natural language. Threads is letting people talk to the algorithm directly.
The common thread is not simply personalization. It is negotiation. Users are being invited to state intent instead of only producing behavioral signals in the background. That changes the interface. It also changes the expectation: if people can tell a system what they want, they will be less patient when the system keeps guessing wrong.
The feed becomes a negotiation
The tension is that most users do not want to manage another dashboard. They want the feed to get better without turning social media into admin. That is why these controls are still relatively simple: topic prompts, sliders, custom feeds, short durations. The platforms are trying to make control feel lightweight enough that people will actually use it.
But even lightweight control changes expectations. Once users can tell a feed what they want more or less of, the algorithm becomes something they can challenge. And once that happens, relevance is no longer only something platforms infer in the background. It becomes something audiences can actively request.
For platforms, that is a way to keep people engaged. For users, it is a small reclaiming of agency. For brands and creators, the consequence is sharper: the next advantage may belong to accounts people actively ask the algorithm to bring closer.