Whether to distill or not depends on the definition of distillation. If it refers to attempting to restore the design concept of OpenAI and replicate a similar architecture from its response, I think it is highly likely. This issue can be viewed from two perspectives. Firstly, the legal provisions of OpenAI prohibit the use of its data for the development of competitive business models - which is a bit lacking in common sense. Moreover, the training process of OpenAI uses a large amount of data from commercial websites, so there are also provisions that cannot be used for competitive business models. Of course, this is very detrimental to innovators, as human progress relies on such massive and unrewarding investments. 2、 China is already a country of imitations, and experts are skilled in inventing, copying, imitating, and optimizing. Huaqiangbei is a great example. Of course, Huaqiangbei cannot cover the United States, while Deepseek, a free platform, can easily cover the whole world. It is difficult to consider plagiarism based on general knowledge for a black box research imitation. Just like when Huawei faced Cisco and developed a complete system, including commands, they didn't have Cisco's source code. The answer to whether this behavior is beautiful is certainly no, but it is also difficult to say that it violates human moral consensus.