Neng Nana
In a strategic pivot aimed at reshaping the user experience within its notoriously volatile reply sections, X (formerly Twitter) has announced a significant tweak to its recommendation algorithm. The platform is now prioritizing content from “mutuals”—accounts that users follow and that follow them back—in an effort to foster more meaningful engagement and reduce the “battleground” atmosphere that has characterized the platform’s public conversations for years.
The update, announced by X’s Head of Product, Nikita Bier, marks a deliberate departure from the algorithm’s previous state, which apparently failed to adequately weigh the social connections that define a user’s inner circle. As the platform continues to grapple with shifting user sentiment and external criticism regarding its safety standards, this change represents a tactical attempt to rebuild the sense of community that once defined the service.
The Technical Gap: Why Mutuals Were Being Buried
The decision to elevate mutuals was prompted by an internal realization that a fundamental piece of social architecture was missing from X’s recommendation engine. According to Bier, the company identified a specific deficiency in how the algorithm evaluated relationships between users.
For some time, power users and casual observers alike had noted that the reply sections on X often felt dominated by strangers, viral provocateurs, or accounts with no discernable connection to the original poster. By failing to account for the “mutual” status—a hallmark of genuine social networking—the algorithm inadvertently diluted the voices of people the user actually knew or respected.
“We’re rolling out a small tweak to boost visibility of your posts to your mutuals,” Bier stated in a post regarding the update. He candidly acknowledged that the previous algorithmic gap had caused friends to “appear less” in reply threads. The result, he noted, was a digital environment where the comment sections felt more like a “battleground with people you don’t recognize.” By re-inserting this social weight, X hopes to encourage the formation of interest-based clusters, allowing for more cohesive, topical discussions rather than the fragmented, high-conflict environment that has become the platform’s default setting.
A Chronology of Algorithmic Evolution at X
To understand the weight of this change, it is necessary to view it within the broader timeline of X’s product evolution since the platform’s acquisition by Elon Musk in late 2022.
2023: The Shift Toward Paid Visibility
Following the transition to private ownership, the platform underwent a radical shift in its monetization and visibility models. The introduction of X Premium (formerly Twitter Blue) fundamentally changed how the algorithm surfaced content. By prioritizing replies from verified subscribers, the platform inadvertently incentivized "engagement bait"—replies designed specifically to provoke or capture attention rather than contribute value.
Early 2024: The Pivot to AI-Powered Discovery
Recognizing the limitations of the engagement-driven model, X began experimenting with more sophisticated discovery tools. In April, the company introduced “Grok-powered custom timelines.” This feature allowed Premium subscribers to generate and pin feeds based on specific niche topics. Nikita Bier, who has become the public face of many of these product updates, positioned this as a way for users to “dive deep into your favorite niche.”
Late 2024: The Return to Human Connection
The current "mutuals" update serves as a corrective measure for the problems created by the 2023 shift. If the goal of the 2023 updates was to maximize reach and monetization, the goal of the current update appears to be the preservation of the user base through the restoration of community.
Supporting Data: The Challenge of Platform Safety
While X frames this algorithmic tweak as a product improvement, it arrives against a backdrop of severe criticism regarding the platform’s safety and health. The "battleground" atmosphere that Bier refers to is not merely a product design failure; it is a central concern for regulators, advertisers, and advocacy groups.
In May, the Gay & Lesbian Alliance Against Defamation (GLAAD) released its annual Social Media Safety Index. The findings were stark. X scored a dismal 29 out of 100 points, ranking it as the lowest-performing major social media platform for LGBTQ+ safety. The report cited a widespread failure to enforce moderation policies, leading to an environment that the authors described as “rife with anti-LGBTQ hate, harassment, and disinformation.”
When analyzed in this context, the “mutuals” update takes on a different dimension. By creating a “bubble” of people a user already follows, X may be attempting to create smaller, safer pockets of conversation. However, critics argue that this does little to solve the systemic issues of hate speech that persist on the broader platform. If a user’s mutuals are not the source of the toxicity, they might experience a temporary reprieve, but the platform’s overall architecture remains unchanged.
Official Responses and Internal Logic
The internal logic driving this update is rooted in the belief that social media should act as a bridge between known entities rather than a megaphone for anonymous conflict. Nikita Bier has been instrumental in articulating this vision. His communication style represents a shift from the more volatile, top-down announcements of the past, opting instead for a more transparent, “product-manager-to-user” dialogue.
Bier’s assertion that the algorithm previously ignored mutual relationships suggests that the platform’s technical debt is significant. When a platform as large as X realizes that its foundational sorting mechanism is missing a core social signal, it highlights the complexity of the machine learning models at play.
Internal company sources suggest that this is only the beginning of a larger project to “de-anonymize” the reply experience. The company’s focus on “interest clusters” implies that they are looking to replicate the success of smaller, community-focused platforms like Discord or private group chats, which have gained popularity as public town squares have become more hostile.
Implications: Can an Algorithm Fix Culture?
The fundamental question remaining is whether a technical adjustment to visibility settings can fundamentally alter the cultural trajectory of X.
The Fragmentation Risk
By prioritizing mutuals, X runs the risk of creating echo chambers. If the algorithm successfully hides voices that a user does not follow, it may effectively silence dissenting opinions or broader societal perspectives. While this may make the platform feel more “friendly” and “safe” for the individual user, it could exacerbate the polarization of the platform as a whole, as users become increasingly insulated from content outside their own circles.
The Impact on Creator Economy
For content creators and journalists who rely on X for reach, this change is a double-edged sword. On one hand, it guarantees that their content will be seen by their most loyal followers—those they have built reciprocal relationships with. On the other hand, it may make “breaking out” or going viral to a new audience significantly harder, as the algorithm will now prioritize existing network density over raw engagement velocity.
The Advertising Paradox
Advertisers have long expressed discomfort with the “battleground” nature of X’s reply sections. If the algorithm can successfully clean up the reply threads, it could potentially win back some of the brand-safety-conscious advertisers that fled the platform. However, if the result is simply a more fragmented user base, the long-term sustainability of the platform’s ad-revenue model remains in question.
Conclusion: A Delicate Balancing Act
X’s move to boost the visibility of mutuals is a pragmatic acknowledgment that the platform’s current algorithmic trajectory was alienating the very users it needed to retain. By pivoting back toward social graph-based recommendations rather than pure engagement-bait, the platform is attempting to perform a delicate balancing act.
The company is trying to reclaim the intimacy of a social network while maintaining the reach of a public square. Whether this change is enough to move the needle on platform safety—or whether it simply hides the cracks in a deteriorating foundation—remains to be seen. What is clear is that the battle for the soul of X is shifting away from the high-level policy changes of the past and toward the granular, invisible mechanics of the algorithm.
As users navigate these changes in the coming weeks, the success of the update will be measured not by the number of replies generated, but by the quality of the interactions within them. If the “battleground” begins to look more like a community, X may find that it has finally found a way to leverage its remaining social capital. If, however, the toxicity persists despite these changes, it will only reinforce the argument that the issues at X are structural, cultural, and perhaps beyond the reach of any single algorithm.
Ammar Sabilarrohman
Iffa Jayyana
