**[Is Meta Trying to End Chinese Open-Source Models? Muse Core Member Criticizes Kimi, Faces Backlash in Comments Section]** According to monitoring by Dongcha Beating, Zengyi Qin, a member of Meta's Super Intelligent Lab and a core contributor to Muse Spark, publicly criticized Chinese open-source models. He claimed that Meta has an order of magnitude more computing power and better data, and that Muse Spark will eventually surpass Chinese models like Kimi. He also extended his argument to the business realm: major U.S. clients like JPMorgan would switch to U.S.-based open models for compliance reasons, causing Chinese labs to lose this portion of inference revenue. Meta, he noted, has Facebook and Instagram to generate income, while Chinese model companies rely more heavily on model-related revenue. The comments section quickly fired back: if Meta hasn’t been lacking computing power or data over the past two years, why hasn’t it outperformed Chinese models yet? Others questioned how much revenue JPMorgan has actually contributed to Kimi. Some even mocked Zengyi Qin, suggesting that if this is the level of reasoning from a "Muse Spark core member," they’re now more concerned about Muse’s model performance. Muse Spark 1.2 is set to release its weights soon, and Meta will indeed add a heavyweight U.S. competitor to models like Kimi, DeepSeek, and Qwen. However, jumping from "a new strong competitor" to "Chinese models will be crushed by computing power, and U.S. revenue will be lost" seems like quite a leap. [Original Article Link]