In the latest episode of the All In Podcast, we talked about Anthropic - what a few big shots said that I think is worth sharing Before the opening on Monday, I saw many people on Twitter worrying about whether the Anthropic incident would affect market sentiment. I have organized the core content of All In this issue for everyone to judge for themselves. First, let's talk about what has already happened Anthropic has issued an official announcement: The US government has issued an export control directive citing national security reasons, requiring the suspension of access to Fable 5 and Mythos 5 for all foreign nationals, whether they are within or outside the United States, including foreign employees within Anthropic's own company. In order to ensure compliance, Anthropic was forced to immediately shut down access to these two models for all customers. Other Claude models are not affected. Anthropic stated in the announcement that they believe this is a misunderstanding and are working to restore access as soon as possible. The scale and suddenness of this matter exceeded many people's expectations. Besides, the background of this conversation about All In Last Tuesday, Anthropic just released its latest model Fable 5, which has performance close to breaking all industry benchmarks, but the token cost is twice that of the previous generation. Then the developer community exploded for two main reasons. Firstly, Anthropic mandates the retention of all user prompt words and contextual data for 30 days, even for enterprise customers who have signed a zero data retention agreement. Secondly, and also the most infuriating aspect for developers: if the backend detects that you are using Fable 5 for cutting-edge AI research, chip design, or machine learning development, it will quietly downgrade to a lower order model without notifying you, and even rewrite your prompt words, but the fee will still be charged according to Fable 5. In the program, Jason tested it on site and asked some common questions that reporters would ask in investigative reporting. As a result, the backend directly downgraded him back to the old model. What do the big shots think about this matter David Sacks was very direct: Anthropic is using security fears for regulatory arbitrage. Dario recently called for the establishment of a government agency similar to the FDA to approve all AI models. Sacks believes that the real purpose of this is to use government power to establish a technological monopoly and strangle open-source AI. Friedberg said a more realistic thing: his own agricultural gene company, which makes completely harmless plant gene improvement predictions, has been frequently downgraded and restricted by Anthropic due to sensitive terms involving biology and genes. He said this is forcing a large number of American startups to turn to open-source models. Currently, many of the strongest open-source models in the world come from China, and the United States' own policies are pushing its technological advantages outward. Will this matter affect the Monday market? I think the short-term impact is limited, but it should be taken seriously. Anthropic has not yet gone public, so its direct impact on the public market is limited. But the things exposed behind this incident can be paid attention to. The government has started to directly intervene in the access permissions of top AI models under the pretext of national security, which is not a small signal. If this logic extends, all top AI models may face similar regulatory risks. What impact will this have on the revenue logic of the entire AI commercialization? It needs to be continuously tracked. The main factors that really affect Monday's trend are still the sentiment before the Bank of Japan meeting and FOMC. But this incident has raised the uncertainty of the entire AI regulation to a new level and needs to be put on the watchlist. The core logic of AI fundamentals has not been shaken.