Technically the Chinese open source models already have a bunch of the data that Anthropic and OpenAI stole from the world through their distillations.