The video covers the conflict between Anthropic and OpenAI over U.S. military AI contracts, with Anthropic refusing to relax its ethical safeguards while OpenAI quickly stepped in to make a deal, claiming similar restrictions but facing accusations of dishonesty and opportunism. The creator criticizes OpenAI’s actions, questions the effectiveness of contractual safeguards, and highlights the political and ethical complexities surrounding AI use in government.
The video discusses the escalating conflict between OpenAI and Anthropic over military AI contracts with the U.S. government, particularly the Department of Defense (referred to as the Department of War). Anthropic, historically a major AI partner for the government, refused to alter its safety protocols to allow its models to be used for purposes it deemed unethical, such as autonomous weapons and mass surveillance of American citizens. The government, unhappy with these restrictions, pressured Anthropic to comply, threatening severe consequences. The creator defends Anthropic’s right to refuse service and criticizes the government’s attempts to force compliance using outdated laws.
OpenAI, led by CEO Sam Altman, seized the opportunity to make a deal with the Department of Defense immediately after Anthropic’s deadline to comply expired. OpenAI’s approach differs in that it does not provide the government with direct access to its model weights, instead maintaining control through its API layer, which allows for ongoing enforcement of safety measures. Sam Altman publicly stated that OpenAI’s contract includes the same red lines as Anthropic’s—prohibiting autonomous weapons and mass surveillance—and that these principles are reflected in both law and policy. However, critics, including Anthropic CEO Dario Amodei, argue that OpenAI’s claims are misleading and amount to “safety theater.”
The video explains the technical and contractual differences between the two companies’ approaches. Anthropic’s model would have been run directly on government hardware, making it impossible for Anthropic to enforce safety restrictions beyond contractual agreements. OpenAI, by keeping the models on its own servers and using an API, retains the ability to monitor and restrict usage in real time. The government’s main objection appears to be to external vendors imposing policy restrictions, rather than the technical limitations themselves. The creator notes a political bias against Anthropic, highlighted by public statements from Donald Trump labeling Anthropic as “left-wing nut jobs.”
The creator criticizes OpenAI’s actions as opportunistic and damaging, both optically and in terms of precedent. While Sam Altman claims his intent was to deescalate the situation and avoid worse outcomes—such as government retaliation against Anthropic or the industry as a whole—the creator is skeptical. He points out that the government’s definition of “lawful use” can change, making any contractual safeguards potentially meaningless if the law shifts. The video also highlights the political nature of the dispute, with government actors using the situation to score political points and punish companies they perceive as ideologically opposed.
Finally, the video addresses the fallout from the controversy. Dario Amodei accuses OpenAI of lying and only including safety restrictions to placate employees, not out of genuine concern for abuse prevention. The public and media largely view OpenAI’s deal as suspicious, leading to a surge in users switching from ChatGPT to Anthropic’s Claude. The creator expresses disappointment in OpenAI’s leadership and offers support to any employees who feel betrayed by the company’s actions. He concludes by emphasizing the complexity and political nature of the situation, expressing hope for more honest and principled decision-making in the future.