According to X's product lead Nikita Bier, starting next week, Grok's capabilities will be fully integrated into the recommendation system, and he directly described this update as the most important change in X's history so far. Before: The algorithm mainly relied on manually written rules, sorting content based on simple metrics like likes, shares, and watch time. After the update: Grok AI will analyze every post and video on the platform in real time, using AI to understand the meaning, quality, and relevance of the content. It will score each piece of content and then decide who it’s best suited to recommend it to. For example, crypto enthusiasts might see more crypto-related content, even if it’s posted by smaller accounts. As long as the content meets relevant needs, it will be recommended. Overall, the timeline will become more user-friendly. Previously, X was more about distributing content based on behavioral data, similar to Xiaohongshu’s model. After the update, it will be more like letting AI directly participate in content understanding and traffic allocation. The specific changes are likely to be: 1. Recommendations will feel more tailored to what you want to see, rather than just pushing the most popular content. 2. Low-quality content, engagement-bait posts, and pure emotional noise will theoretically find it harder to gain traffic. 3. Smaller accounts will gain more weight—if their content has enough value, the chances of being seen might actually increase. 4. Timelines will likely become increasingly customizable, like only seeing crypto-related content or only political content. So if this update rolls out, the biggest impact will probably be on the platform’s traffic distribution model. Good content, as long as there’s demand, will theoretically get better exposure (in theory ).