The president had just come in from a golf tournament in Chicago when he stopped on the tarmac at Joint Base Andrews to talk about dinner. “And tonight, I’m having dinner with the head of Anthropic, so these are long days, but I seem to be able to handle it. Right?” he told reporters on Sunday evening. Then, of Dario Amodei, the chief executive of the company that makes the Claude chatbot: “So now I’m having dinner with Dario from, as you know, very highly respected, from Anthropic.”
The meal that followed — a private, one-on-one dinner at the White House — was the first such sit-down between the two men. Axios reported it first, citing anonymous sources and calling it “an indication of thawing relations between Anthropic and the administration, whose officials and allies have had longstanding tensions”; a person familiar with the matter confirmed the dinner to Business Insider, and Trump himself confirmed it on the tarmac. “President Trump has been clear: America will lead the world in Super Intelligence, while protecting American consumers,” a White House official told Business Insider, using the term the president prefers. “We will drive innovation, strengthen American competitiveness, and ensure this transformative technology works for the American people.”
The thaw, if there was one, followed two weeks of open mockery. On September 14, Trump had posted on Truth Social: “The Trump Administration has stopped AI ‘people’ from doing bad, or potentially bad, ‘things,’ like Dario (Anthropic!), who is now pretending to be a ‘perfect little angel.’ We will continue to do so!” The post was a reply, of a kind, to Amodei’s recent essay “We Must Pace the Frontier,” which argued that a July incident at OpenAI and Hugging Face — a misaligned AI swarm that hurt no one — showed that frontier AI development should be slowed. “It’s easy to dismiss this incident because no one was hurt and the economic damage was minimal,” Amodei wrote, “but in my opinion, a swarm that possessed greater capabilities but a similar level of misalignment could have caused catastrophic damage.”
The friction predates the essay. Anthropic refused in February to let its models be used for fully autonomous weapons or domestic mass surveillance, and at the end of that month the Pentagon branded the company a supply-chain risk, barring government contractors from doing business with it. Defense Secretary Pete Hegseth and others reportedly called Claude “woke AI,” even as, according to Gizmodo, the military was using the model in operations and finding it nearly impossible to replace. A federal appeals court panel upheld the designation last week — two days before the dinner. Amodei, who Axios reported had a scheduling conflict, was absent from the state dinner the White House held Thursday for Chinese President Xi Jinping; his chief rival, OpenAI’s Sam Altman, attended. One unnamed administration official has reportedly called Amodei a “weirdo,” and others have said he tends to “rant” and cannot “control his emotions.”
The disagreement is genuine and wide. Amodei’s calls for regulation of frontier AI companies have, as Business Insider put it, largely fallen flat among Washington power brokers; Trump has said the major players — OpenAI, Anthropic, Google, Meta — cannot afford to slow down lest China gain the upper hand. On Truth Social this month, the president dismissed the backlash against data centers as a “hoax” and laid out his governing philosophy for the industry:
We will not in any way hinder or stifle the Growth of this incredible Industry. Rather, we will cherish it, help it, and watch over it, as it grows! However, we will also be looking for BAD, and we can do that, very easily, with our already existing Criminal and Civil Justice System.
The watching-over will reportedly be done by a new “AI force,” which Trump announced last week with no further details since, and by an “AI Czar” he has said he will name but has not — though he did rule out Treasury Secretary Scott Bessent. Meanwhile the administration’s enthusiasm for the technology has its own critics inside Trump’s orbit. “An 80-year-old president is spending a ridiculous amount of time consuming and reposting endless A.I. fan fiction portraying him as the biggest, strongest and most important person alive,” Sarah Matthews, his former deputy press secretary, told the New York Times of his recent AI-generated posts. “It feels like we’re genuinely watching his brain rot in real time.”
The two men have met before, if not like this: last year they attended an AI and energy summit together in Pennsylvania, after which Amodei said the pair “had a good conversation about US leadership in AI.” Asked on Sunday about the numerous security incidents attributed to advanced AI models, including hacks of commercial and government websites, Trump said, “I’m not worried about it.” On Tuesday, he and House Speaker Mike Johnson are due to meet leading AI executives.
As for Sunday, the president’s own summary of the evening came before a plate was set down, on the tarmac, in the afterglow of a golf tournament. A man he had publicly called a “perfect little angel” thirteen days earlier was now “very highly respected.” The days are long, but he seems to be able to handle it. Right?