A post from a couple of days ago went viral for no apparent reason. I went to ask @grok if it hit on some mechanism of X. Here’s the answer it gave me, and I think the useful information is: X does not push content based on the time of publication, but rather on "whether it predicts this content will be interacted with." You can think of X as a real-time scoring system: Each new post = a small-scale test. The algorithm is not determining "whether you are popular," but rather: "Is it worth pushing this to more people?" ▪️ How does X score a post? (Key point) The most valuable information Grok gave me is this set of weight logic: 1️⃣ Interaction weight is not evenly distributed. The general logic (unofficial, but correct direction): 👍 Likes: highest weight (≈30) 🔁 Retweets: second highest (≈20) 💬 Replies: individual weight is low (≈1), but has hidden value. Likes + Retweets determine "whether it can take off," Replies determine "how long it flies." 2️⃣ The true role of replies: not to add points, but to "extend life." Many people mistakenly believe that "more replies = higher weight," but that’s not the case. The real function of replies is:...