Elon Musk Changes X Algorithm to Prioritize Posts People Actually Read

Elon Musk has announced changes to the recommendation system powering X, the social media platform formerly known as Twitter. The update introduces new artificial intelligence models designed to understand which posts capture users’ attention and which ones people read all the way through.

The changes could affect the content appearing in users’ For You feeds, giving greater importance to meaningful attention rather than relying exclusively on familiar engagement signals such as likes, reposts and replies.

For everyday users, creators, publishers and businesses, this development highlights how artificial intelligence is increasingly shaping the way social media content is discovered and distributed.

What Has Elon Musk Changed on X?

According to Musk’s announcement, X is introducing two important improvements to its recommendation system. One model is designed to predict how much time a user is likely to spend viewing a post, while another helps identify content that people are likely to read from beginning to end.

These models perform different jobs within the recommendation process. The retrieval system identifies potential posts that may interest a user, while the ranking system determines which posts should appear higher in the feed.

Together, the changes aim to improve the platform’s ability to recommend relevant content based on user interests and viewing behaviour.

Although Musk has outlined the purpose of the new models, the company has not publicly explained every detail of their implementation or precisely how different engagement signals will be weighted.

AI Could Change What Appears in the For You Feed

X’s For You feed recommends posts based on a range of signals intended to predict what users will find interesting. The latest changes suggest that the platform is putting greater emphasis on whether people genuinely spend time with content.

For example, a detailed technology explanation that users read completely could become more competitive in recommendations than a post that attracts quick interactions but receives little sustained attention.

However, this does not necessarily mean that likes, replies and reposts will stop mattering. The available announcement does not establish that these signals have been removed or made irrelevant.

Instead, the update points towards a recommendation system that attempts to understand user interest more accurately.

For users, this could mean a more personalised feed. For publishers and creators, it may increase the importance of producing material that delivers useful information and maintains readers’ attention.

Why Reading Time Matters for Social Media

Social media platforms have traditionally relied on different engagement signals to help determine which posts deserve greater visibility. Likes, comments, shares and reposts can indicate interest, but they do not always reveal whether someone found a post genuinely useful.

Reading time provides another possible indication of interest.

A user might like a post without reading its full content. In another situation, someone may spend several minutes reading an informative thread without interacting with it publicly.

An AI system designed to predict reading time could help identify these differences. By combining multiple signals, a recommendation system may become better at matching posts with people who are likely to appreciate them.

Nevertheless, longer viewing time does not automatically mean that content is accurate, useful or trustworthy. The quality of recommendations will also depend on how the system interprets user behaviour and balances different objectives.

What the Algorithm Update Means for Content Creators

The changes could be particularly important for creators, bloggers, news publishers and businesses that depend on organic reach from X.

If reading behaviour becomes a more significant ranking factor, creators may need to focus on making their content interesting enough for people to keep reading.

Several practices may help creators adapt:

Write compelling opening lines: Introduce the main topic clearly and explain why it matters to the audience.

Deliver useful information: Posts should answer questions, explain developments or offer meaningful insights rather than relying entirely on sensational headlines.

Structure longer posts carefully: Break complex information into readable paragraphs, use clear explanations and organise threads logically.

Understand the target audience: Technology enthusiasts, gamers, business owners and general news readers may respond to different types of content.

Prioritise originality: Share original reporting, analysis, practical advice or perspectives that provide value beyond simply repeating existing posts.

Monitor performance: Compare reach, engagement and other available analytics to understand which topics resonate with followers.

These are sensible content strategies, not guaranteed ways to rank higher. X has not published a complete formula explaining how creators can optimise their posts for the new models.

Will Likes and Reposts Become Less Important?

The announcement does not confirm that X is abandoning likes, reposts, replies or other traditional engagement signals.

Instead, the new models appear intended to improve the platform’s understanding of how users interact with recommended content. Reading behaviour and predicted time spent may become additional considerations within the overall system.

A post that receives many likes may still attract significant attention, while a detailed article or thread could demonstrate value through sustained reading.

The practical impact will depend on how X combines these signals, how the models perform and how the platform continues adjusting its recommendations.

Users should therefore avoid assuming that a single metric, including reading time, will determine whether a post goes viral.

What the Changes Mean for Ordinary X Users

For everyday users, the update could influence the variety and order of posts appearing in the For You feed.

People who regularly read technology news, sports discussions, educational threads or entertainment updates may receive recommendations that more closely reflect their demonstrated interests.

However, the outcome will depend on the effectiveness of the recommendation models and the signals used to personalise individual feeds.

Users who prefer a feed based on accounts they deliberately follow can also explore X’s Following tab, which offers an alternative to the personalised For You feed.

The new recommendation approach may improve content discovery, but it also raises questions about transparency. Users and publishers may want to understand why certain posts receive greater visibility and whether the system consistently delivers relevant information.

X Continues Expanding Its Use of Artificial Intelligence

The recommendation update forms part of X’s broader push towards AI-driven services and personalisation. Under Musk, the platform has increasingly incorporated AI tools into its ecosystem, including its integration with Grok.

Recommendation systems are an important application of artificial intelligence because they must process large volumes of content and estimate which items may interest individual users.

Improving retrieval and ranking can help a platform organise this information more effectively. At the same time, the success of such systems depends on balancing relevance, diversity, content quality and user preferences.

The latest announcement signals continued investment in making X’s recommendations more responsive to how people actually consume content.

Final Thoughts

Elon Musk’s latest X algorithm update puts greater emphasis on AI models that predict viewing time and identify posts users are likely to read completely.

The changes could influence how content is ranked in the For You feed and encourage creators to publish more engaging, useful and well-structured material.

However, X has not disclosed the full ranking formula or confirmed that traditional engagement signals are being discarded. The actual effect on individual accounts will become clearer as the updated system develops.

For creators and publishers, the most practical approach is to focus on original content, clear writing and genuine audience value rather than trying to manipulate a single engagement metric.

As AI continues to shape social media, understanding how recommendation systems work will remain important for anyone seeking to reach an audience on X.

Source context: Reporting on Elon Musk’s announcement of new AI ranking and retrieval models for X’s recommendation system.