Recent developments highlight ethical concerns over American coding tools potentially relying on Chinese foundational models. Cognition AI’s new SWE-1.5 model, noted for its “near state-of-the-art” coding abilities, was built on an unspecified open-source framework, leading to speculation about its ties to Zhipu AI’s GLM-4.6 model from Beijing. Cognition AI, valued at $10.2 billion, faces scrutiny over commercializing technologies without proper credit. Meanwhile, Cursor, valued at $9.9 billion, introduced Composer, which shows similarities to SWE-1.5 but generates Chinese reasoning outputs, raising further questions about its origins. The controversy is intensified by the permissive open-source licenses of Chinese models, allowing commercialization without attribution. Zhipu AI’s GLM-4.6, launched under an MIT license, saw a significant rise in international users, emphasizing the positive role of open-source in AI development. These developments underline the critical intersection of ethics and technology in the rapidly evolving AI landscape.
Source link
