They “stole” Claude from Anthropic?

This is wild if you don't know the tea. The tea, basically, Anthropic is upset at the Chinese models for basically training on their data. The irony in this. Dude, you trained on the entire internet's data. Every blog, every book, every like, image just uploaded for free. What Anthropic is saying is that the Chinese labs created tens of thousands of fake accounts with like, Claude accounts, Claude subscriptions. And they're just making chat requests, right? Saving the transcripts from the chats and using this is basically like perfect training data, right? Which totally could have happened, but I mean, is one worse? If you're willing to put on your tin foil hat here for just a moment, some people are reasoning that these Chinese labs, the models that they're producing, are becoming very, very close in quality to Anthropic and OpenAI. Which they are. GLM 5.2 is one of these open Chinese models that just released, and it's neck and neck in terms of quality, and it's also about four times cheaper. If Anthropic and OpenAI genuinely can't get their offerings four times cheaper in like the next year, they're kinda screwed. Or if the Chinese models are banned in the U.S. Then maybe they're fine.