Mark Zuckerberg, CEO of the US “Yuan” company, posted an article on the company's official website on the 10th stressing the importance of open source artificial intelligence and calling on the US government to “re-examine policies relating to model distillation and use of training data” to protect the principles that allow models to learn from other models during the development process. In an article entitled “The Future Belongs to Everyone,” Zuckerberg said that models can learn from other models is an important principle in the operation of an open source ecosystem. All models of artificial intelligence derive from human knowledge. Some have tried to describe distillation as harmful behavior, but he believes the principle that “you can learn from anything you can observe” should be protected. Model distillation is a basic technology in the field of artificial intelligence model development. It usually refers to the efficient transfer of knowledge contained in larger artificial intelligence models to lightweight models. As competition for generative artificial intelligence research and development intensifies, model distillation has also become a topic of concern and controversy in the industry. Zuckerberg said that US laboratories are more limited in terms of training data. The US should reduce these additional barriers if it wants its open source model to stay ahead in the long term. Restricting foreign open source models is not an effective solution. He believes that restricting individuals and businesses from using leading open source models, no matter where these models come from, will reduce the quality of artificial intelligence that users can obtain and further centralize artificial intelligence rather than give people more rights. Zuckerberg also said that “Yuan” will continue to firmly support open source, including open source artificial intelligence models. The company will resume publishing open source models as the company's super-smart lab is put into operation.

Mark Zuckerberg, CEO of the US “Yuan” company, posted an article on the company's official website on the 10th stressing the importance of open source artificial intelligence and calling on the US government to “re-examine policies relating to model distillation and use of training data” to protect the principles that allow models to learn from other models during the development process. In an article entitled “The Future Belongs to Everyone,” Zuckerberg said that models can learn from other models is an important principle in the operation of an open source ecosystem. All models of artificial intelligence derive from human knowledge. Some have tried to describe distillation as harmful behavior, but he believes the principle that “you can learn from anything you can observe” should be protected. Model distillation is a basic technology in the field of artificial intelligence model development. It usually refers to the efficient transfer of knowledge contained in larger artificial intelligence models to lightweight models. As competition for generative artificial intelligence research and development intensifies, model distillation has also become a topic of concern and controversy in the industry. Zuckerberg said that US laboratories are more limited in terms of training data. The US should reduce these additional barriers if it wants its open source model to stay ahead in the long term. Restricting foreign open source models is not an effective solution. He believes that restricting individuals and businesses from using leading open source models, no matter where these models come from, will reduce the quality of artificial intelligence that users can obtain and further centralize artificial intelligence rather than give people more rights. Zuckerberg also said that “Yuan” will continue to firmly support open source, including open source artificial intelligence models. The company will resume publishing open source models as the company's super-smart lab is put into operation.