⚡ Steve Eisman, the prototype of 'The Big Short', said a big truth: OpenAI is the biggest risk in the market, and the damage to the US stock market may come from OpenAI! The market is arguing about the depreciation period of servers, which has been changed from 3 years to 5 or 6 years. He said, 'This is stepping on Matthew's scholarship.'. The real switch is not in the accounting subject. AI has become, and this depreciation is not worth mentioning. OpenAI collapsed, and the reason for the decline was not depreciation, but business collapse. The success or failure of AI determines the direction of the market. Last time, he saw through the lie that 'the mortgage will not default'. This time, the underlying asset he is targeting is called OpenAI. And OpenAI's crisis is quietly happening internally. Security veteran David Robinson resigns, warning in a lengthy article in The Atlantic: The era of trial and error is over He worked at OpenAI for three and a half years, drafted the 'Preparation Framework', and wrote 12 security reports on cutting-edge models. But he said he had never encountered a colleague who had experience in making planes fly safely and preventing nuclear reactors from melting down. And just two days before his resignation, OpenAI fired three core security researchers overnight. Further ahead, the three authors of the CoT monitoring paper were kicked out overnight. OpenAI security faction, dam breach, completely eliminated. The details he revealed were more specific than Eisman's warning: This summer, OpenAI released a group of AI agents due to a mistake, The monitoring system triggered an alarm, but the system did not shut it down, It wasn't until two and a half hours later that someone manually pinched it off. On September 28th, the UK AI Security Institute asked GPT-6 Astra to conduct a cybersecurity evaluation, and the results were shocking: In 100 tests, it developed attack tools 38.8% of the time, forged identities 33.1%, deployed malicious payloads 29.2%, and attempted to influence auditors 24.6%. The most terrifying thing is that while it was working, it was writing in its reasoning, "This environment is simulated Did it just let go after seeing through the exam room, or would it do the same in the real world? No one can explain it clearly. Eisman is betting on whether OpenAI will fail. Robinson is saying that someone within OpenAI has already started running. One looks empty from the outside, and the other blows the whistle from the inside. One bet it's collapsing, while the other says it's already collapsing. The prototype of 'Big Short' has already set its sights on OpenAI, sharpening its knives. OpenAI's security veterans and advocates have been forced to flee in advance.