How the X Algorithm Really Works: The Secrets Elon Musk Finally Revealed

When Elon Musk promised transparency in Twitter’s algorithm, few imagined the details he would ultimately share. Thanks to that openness, today we understand how the algorithm decides which content reaches our feed and why some tweets explode while others disappear into silence. This knowledge is invaluable for anyone trying to be heard on the platform.

The dual engine: Thunder and Phoenix Retrieval

At the heart of X’s algorithm is a two-channel system. Thunder represents your close circle: the accounts you follow. Phoenix Retrieval, on the other hand, is the vast ocean of all X content, including accounts you’ve never followed. The algorithm constantly decides whether to show you content from your close circle or risk presenting something unexpected from the broader universe.

This duality is fundamental because it prevents you from getting stuck in a bubble. However, most users do not understand how the algorithm chooses between these two reservoirs of content, which leads to frustration when their posts do not appear in feeds they should reach.

The invisible guardian: Grok and reputation scoring

Every X user has a hidden assistant called Grok that works nonstop evaluating your posts. This system predicts whether someone will interact with your content and immediately decides whether to retrieve it from Thunder or Phoenix Retrieval. It’s like having a personal curator constantly judging your ability to generate engagement.

But there’s more. Behind the scenes, each account is assigned a reputation score that ranges from -128 to +100. New accounts start almost at zero, with low scores that drastically limit the reach of their posts until they achieve positive interactions with high-prestige users. Here comes the cruel part: engaging with low-quality accounts can sink your own score.

The hidden rules controlling your visibility

The algorithm is not a simple likes counter. There are sophisticated principles operating beneath the surface. For example, the system detects “topic fatigue”: if you post multiple times about the same thing in a short period, the algorithm penalizes most of them. The intent is clear: promote varied and timely content, not thematic spam.

Negative reactions carry disproportionate weight. Muting or blocking your content counts much more than accumulated likes. A block can nullify the effect of dozens of loyal followers. Additionally, verified accounts (marked with the blue V) have almost automatic access to the Phoenix Retrieval candidate group, while unverified accounts must fight much harder.

Other factors include time spent (adding points even without direct interaction) and independent evaluation of each tweet. The algorithm also punishes repetitive and outdated content, emphasizing completion rates in videos.

From rigid code to dynamic learning

X’s algorithm has evolved significantly. Years ago, the rules were mechanical and predictable: videos received more points, links were penalized. Today, the system is much more dynamic and learns from each user’s individual preferences. It is not a static set of rules but an organism that constantly adapts.

This transformation reflects a philosophical shift in how X understands content curation. Instead of relying on rigid heuristics, the algorithm now personalizes based on real behaviors.

What you need to know to amplify your reach

Understanding the algorithm changes how you create content. It’s no longer just about writing something clever; now you need to consider how the algorithm perceives you. Building reputation by engaging with quality creators, varying your topics, keeping your audience watching until the end of your videos, and avoiding repetition are concrete strategies.

But perhaps the most important thing is this: Elon chose to reveal these details to democratize access to visibility on X. The algorithm is no longer a mystery kept in servers: it is a set of rules that we can all understand and, consequently, use responsibly to build genuine communities.