According to CoinMeta, Musk proposed at the all-staff summit in Los Angeles that companies such as OpenAI, Google, Meta, Anthropic, and other AI companies should allow competitors to test their models before release. He believes that this would help to better identify potential dangerous capabilities or weaknesses. Although such a system cannot eliminate all risks, it can significantly increase the chances of detecting issues. This proposal was made against the backdrop of growing warnings within the AI industry, and Anthropic CEO Dario Amodei also calls for caution regarding increasingly powerful cutting-edge models. Nevertheless, there are still disagreements within the industry about whether to slow down development. Musk's proposal further drives discussions about the regulation of advanced AI, indicating his preference for industry-led mechanisms rather than government restrictions to strengthen the scrutiny of powerful AI models.