Trump Charms by Claude Mythos: Anthropic Could Make a Comeback

An Unexpected Turn in the Anthropic-Trump Conflict

The dispute between Anthropic, an artificial intelligence startup, and the Trump administration appears to be on the verge of resolution. The American president recently mentioned the possibility of a new agreement between the two parties, marking a notable shift in this tense saga. This turnaround raises questions about the motivations behind this sudden openness and the future of relations between the government and the startup.

Last year, Anthropic signed a substantial $200 million contract with the Pentagon. However, negotiations quickly fell apart when the Department of Defense demanded unrestricted access to Anthropic's AI models for all legal purposes. The startup firmly refused to allow its technologies to be used for the development of autonomous weapons or for mass surveillance. As a result, the government blacklisted Anthropic, prohibiting military contractors from using its artificial intelligences in their projects. Despite this ban, the military continued to use the Claude model in Iran, prompting Anthropic to take the matter to court, resulting in a temporary suspension of the ban.

Promising Discussions at the White House

Tensions now seem to be easing. On April 17, Dario Amodei, CEO of Anthropic, met with senior officials at the White House, including Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent. This meeting was described as a success.

In an interview with CNBC on April 21, Donald Trump expressed optimism about a future collaboration with Anthropic. "We had excellent discussions with them, and I think they are coming into their own. They are very smart, and I think they can be of great help to us," he stated. A White House spokesperson also described the meeting as "productive and constructive."

Claude Mythos at the Heart of Negotiations

Recent discussions have primarily focused on Claude Mythos, Anthropic's latest model, which is touted as the most powerful developed by the company to date. This model, launched to a select group of partners, stands out for its advanced cybersecurity capabilities, making it a major asset for rekindling dialogue with the government.

While an agreement seems to be on the horizon, details remain unclear. Anthropic is unlikely to compromise on its core principles. The real negotiation begins now: how far is the Pentagon willing to go to accept restrictions on the use of a technology it desperately desires?