Let's start with a counterintuitive judgment — the default thinking is: if the functionality is insufficient, just add more features. My daily use of Planet leads to the exact opposite conclusion: what you need to do is not addition, but subtraction — first, delete all those settings that oppose "following" one by one. This is my third article teaching OKX product development. In the first two, I talked about the PC homepage and team collaboration. In this one, I discuss a more critical issue — how exactly Planet should increase its DAU. Outline of this article: 🎯 Want to increase DAU? Don’t rush to add features; what you lack is a "core task." 🚧 Your most conflicted area right now: constantly opposing "following." 🔵 First cut in the feed: replace that dull gray effect expert badge with a blue badge. 🔁 Second cut in the feed: learn from Weibo and Twitter to promote KOL recommendations. 📺 The most counterintuitive move in live streaming: while others desperately help fans follow, you are restricting it. ❤️ Bonds are the engine of DAU; tipping is just the matchstick. Enough talk, let's start the main text 🙏 🧵🧵🧵 Outline of this article Planet wants to increase DAU, but the direction has been wrong from the start. 1. Want to increase DAU? Don’t rush to add features; what you lack is a "core task." 🎯 Whether making games, products, or setting strategies, the first question I always ask is: what is the core task? Every action of a team must focus on this one task. Anything that contradicts or hinders this core task must be cut off. That’s it.

Let me start with something that might offend a lot of people — it's not just the AI era that's wrong. Even in the internet era, it was wrong; it's just that back then, people had no choice. This is my second article about teaching OKX product development. In the last one, I criticized their PC homepage. But in this one, I won't talk about how to change a specific feature; I'll talk about something bigger — what the entire team collaboration method should really look like. This approach isn't something I came up with on a whim; I experimented with it bit by bit within my own company. Here, I won't discuss the specifics of how developers write code; the core question is: how should product and development teams cooperate? Outline of this article: 🧩 First, correct a concept that has been domesticated for twenty years: division of labor itself is wrong. 🔩 The cost of over-division: everyone becomes a screw who doesn't understand the whole picture. 🎯 Product managers in the AI era: those who can't produce anything are unqualified. 🛠 Developers in the AI era: the wait-and-see, hands-off managers are no longer useful. 📡 Information gets distorted after passing through three hands; the middle layer is the biggest pitfall. 🤝 My answer: establish a tripartite AI collaboration group of product, development, and operations. 🏃 Learn by doing, not by studying before doing; start working first to push yourself. No more nonsense, let's begin the main text 🙏 🧵🧵🧵 Article outline Frontend does frontend, backend does backend, product does product, testing does testing — this division of labor you've been doing for over a decade is itself wrong. 1. First, correct a concept that has been domesticated for twenty years: division of labor itself is wrong 🧩 China Internet