A logic circulates widely: whoever builds the better product wins. That logic has a big problem. Let me put it plainly: Tencent never won because its products were good. It won because it copied well — always the latest to build, the fastest to copy.
Not convinced? Read the feature list. Yuanbao bolted a pile of features onto DeepSeek: voice, image recognition, even phone calls. DeepSeek itself: text chat, images barely legible. And Yuanbao still couldn’t beat DeepSeek — because DeepSeek was already inside it. Users recognize a name, not a feature table.
Then look at names. Almost no company dares put an avatar forward as its AI’s face. Only Doubao did — and Doubao is the most successful. Is that scientific? No. But users aren’t running benchmarks. They’re recognizing a person.
Someone says Tencent eventually meets things it can’t copy. Are you kidding? When Tencent starts copying, it just crushes you with money.
And when it can’t crush? Then everyone competes into the ground. ByteDance launches a cloud service, Tencent clones one. No sense of boundaries, vicious involution — in the end everyone has one, which equals nobody having one, because everything is the same.
So when you judge who wins, stop counting features. Watch the follow-up speed: the one who builds latest and copies fastest is the real winner.
