The social media platform X is planning a significant shift in how users experience its content feed, with a goal to transition to a fully AI-powered algorithm by November. Company owner Elon Musk announced the change, stating that users will soon be able to directly instruct the platform's AI, Grok, to customize their timelines.
Key Takeaways
- X aims to implement a purely AI-driven algorithm for its main feed by November 2025.
- Users will be able to provide direct text commands to the Grok AI to adjust the content they see.
- The initiative is designed to help users discover niche communities and interests beyond mainstream topics.
- The move faces the challenge that a large majority of X users are passive and may not actively customize their feeds.
A New Approach to Content Curation
X is moving away from its current content ranking system toward a model driven entirely by artificial intelligence. The objective is to create a more personalized and engaging user experience by better understanding and catering to individual interests.
Nikita Bier, the head of product at X, outlined the vision for this new system. The goal is to help users break free from what he described as the "mainstream algo and the political crusades" and instead connect with communities centered on their specific hobbies and interests.
"The goal for your X timeline is to get out of the mainstream algo and the political crusades and find your niche. You should be able to post about your interests and have friendly, relevant people chime in," Bier explained in a recent post.
This strategy aims to improve user retention by making the platform a more relevant and welcoming space for diverse interest groups. The company acknowledges that it needs to refine its systems to achieve this, especially for new or less active accounts that may still see generic or sensational content.
Grok-Powered Customization
The core of the new system will be Grok, X's proprietary AI model. According to Elon Musk, the platform is on a fast track to integrate this technology fully into the user experience.
"The algorithm will be purely AI by November, with significant progress along the way," Musk stated. He added, "By November or certainly December, you will be able to adjust your feed dynamically just by asking Grok."
This feature would allow a user to type a command, such as "Show me more posts about vintage synthesizers" or "Show me fewer posts about politics," to directly influence the algorithm. This method of explicit feedback is more direct than traditional engagement signals like likes or shares, which can be ambiguous.
A Growing Trend in Social Media
The concept of user-directed algorithms is not unique to X. Meta's Threads platform has reportedly been testing a similar feature. Users have spotted an experimental function that would allow them to tag a profile named "@threads.algo" to provide feedback on their feed, suggesting an industry-wide exploration of more direct user control over content curation.
Musk also committed to transparency, promising to "open source the algorithm every two weeks or so." This would allow researchers and the public to examine the code that determines what millions of users see on the platform, a move intended to build trust and accountability.
Challenges of User Adoption and Behavior
While the idea of a fully customizable feed is appealing, its success depends on whether users will actively participate. A significant hurdle for X is the passive nature of its user base. Industry data has consistently shown that a large portion of social media users consume content without actively engaging.
The Engagement Gap
According to platform analytics and market research, approximately 80% of X users are considered "lurkers," meaning they read posts but rarely or never post, comment, or like content themselves. This behavior pattern presents a major challenge for a system that relies on active user input for customization.
Many modern platforms, most notably TikTok, have built highly effective recommendation engines that require no direct input. TikTok's algorithm learns a user's preferences simply by measuring how long they watch a video. This passive learning model has proven to be extremely successful and has conditioned users to expect personalization without effort.
The question for X is whether users, accustomed to this convenience, will be willing to take the extra step of typing commands to refine their feeds. History suggests that convenience often wins out over control.
Historical Precedent for User Inaction
Past events in the social media landscape provide insight into user behavior. In 2018, following the Cambridge Analytica data scandal, there was widespread public outrage over how Facebook handled user data. Calls to delete accounts and demand more privacy controls dominated headlines for months.
However, the long-term impact on user behavior was minimal. In a congressional hearing, Meta's CEO Mark Zuckerberg confirmed the company had seen "no meaningful impact" on platform activity. Users expressed a desire for more control, but in practice, few changed their habits or took the time to adjust complex privacy settings.
This precedent suggests that even with powerful tools at their disposal, the majority of users may default to the path of least resistance, letting the algorithm make decisions for them rather than actively managing it.
The Future of Engagement on X
The success of X's AI-driven timeline will ultimately depend on balancing powerful customization tools with the reality of user behavior. If the process of instructing Grok is seamless and intuitive, it may see some adoption. However, it is unlikely to become the primary way most people interact with the platform.
The initiative represents a fundamental choice for the platform's direction: will it build a system that caters to a highly engaged minority who want granular control, or will it focus on passive personalization for the silent majority? The outcome of this transition, set to be completed by the end of the year, will be a critical test for X's long-term strategy to retain and grow its user base.
Another potential, though controversial, path platforms might explore is the use of AI-driven accounts to simulate engagement. These AI personas could interact with users on their niche topics, creating the feeling of an active community. While this would be an artificial solution, it could provide the positive feedback loop that keeps users posting, even if the engagement isn't entirely human.