I’ve been watching this weird tension for a while. Musk keeps talking about turning X into an “everything app,” basically cloning WeChat. And every time he says it, the tech press nods along like it’s some visionary masterstroke. Meanwhile, Kimi—yeah, that Chinese AI assistant you might’ve heard of—has been quietly shipping something that looks a lot more like the future.
Not the future of “super apps.” The future of agent-native interactions.
Let me rewind a bit. We all know WeChat’s magic: a single app that swallows messaging, payments, shopping, ride-hailing, mini-games, official accounts, you name it. It’s a walled garden where every service is a plugin inside one ecosystem. Musk wants that, but for X. He’s even hired a team to figure out how to graft payments and creator tools onto the bird platform.
But here’s the thing nobody’s saying out loud: that model is from 2015. It’s a monolithic approach, and it only works if you own the entire stack—identity, payment, distribution, user relationships. For Musk, it means forcing a social media app to become a payment hub, then a commerce hub, then a news hub. It’s possible, sure. But it’s slow, and it’s brittle.
Then I saw Kimi’s latest release. And I’m not talking about another chatbot wrapper. I’m talking about an agent that lives inside a canvas, with access to memory, tools, and a continuous workflow. You tell it to do something, it doesn’t just spit out text—it clicks buttons, reads web pages, fills forms, and returns results. It’s like having a personal assistant that actually does things, not just talk about doing them.
This is where the WeChat analogy breaks down. WeChat’s ecosystem is app-to-app: you install a mini-program, it’s a self-contained service. But an agent-based ecosystem is task-to-action: you describe what you want, and the agent orchestrates multiple tools—search, calculation, data extraction, payment, scheduling—without you ever leaving the chat. It’s not a marketplace of apps. It’s a marketplace of capabilities.
Kimi’s approach feels more organic. They didn’t try to build a “super app” by cramming features. They built an agent that learns to use tools as it goes. And because the agent is already connected to the web, to APIs, to memory, it can chain actions in ways that feel… natural. You ask it to find a flight, compare hotels, check your calendar, and book the best one. It doesn’t open a separate app. It just does it.
Now contrast that with what Musk is attempting. X’s current strategy is “let’s add payments, then add creator tools, then add whatever.” It’s feature-driven. It’s rebuilding a 2024 app around a 2009 social network. Meanwhile, Kimi is building a 2025 paradigm: the agent is the container. Everything else is a capability the agent can invoke.
Does that mean Kimi is about to become the next WeChat? Probably not. They’re not building a social network or a payment network. But they’re stumbling onto something more interesting: the agent as the universal interface. If this catches on, the “everything app” won’t be a single monolith. It’ll be an intelligent layer that sits on top of everything else. You just tell it what you want, and it handles the rest.
Musk might be chasing the right goal—a seamless, all-in-one digital life. But he’s using the wrong architecture. Kimi, by accident or design, might have already built the prototype for where we’re actually heading.
Now the question is: who else is paying attention?