President Trump posted on Truth Social and X on September 14, in full: "The only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades! 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' – and we will continue to do so! We already have tremendous CRIMINAL and REGULATORY power over these companies! There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China. WHOEVER WINS AI, WINS! We are leading China, and all others, and will continue to do so. Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE! Thank you for your attention to this matter! President DONALD J. TRUMP"
It's the first response to this week's pacing debate from a head of state, and it names a specific person: Dario Amodei, whose essay two days earlier drew public agreement from Sam Altman and Elon Musk over the weekend. Reading the post cold, the "stopped... from doing bad things" line sounds like rhetorical flourish. It isn't. It refers to a real, specific, and currently unresolved fight between the administration and Anthropic that this blog hasn't covered until now.
What "stopped" actually refers to
On February 27, 2026, Trump directed federal agencies to cease using Anthropic's technology, after a dispute over the company's refusal to lift restrictions on using its models for autonomous weapons systems and mass surveillance. Defense Secretary Pete Hegseth went further, having the Department of Defense designate Anthropic a "supply-chain risk" to national security — a label typically reserved for firms treated as extensions of foreign adversaries — and directing defense contractors and suppliers to cut ties with the company, with up to six months to transition away from its products.
Anthropic sued. On March 26, a federal judge granted a preliminary injunction blocking the supply-chain designation, the presidential directive, and related restrictions while litigation continues. The government appealed to the D.C. Circuit; a panel of appellate judges heard arguments in May and appeared divided, with Judge Karen LeCraft Henderson stating she saw "no evidence to support the Pentagon's determination that Anthropic poses a supply-chain risk to national security." As of this writing, no ruling has been reported, and the case is pending.
That is the gap between Trump's framing and the current legal record. "We have stopped AI 'people'... like Dario... and we will continue to do so" describes an action a federal court currently has enjoined, in a case the administration has not won on appeal. It's accurate that the administration took the action; it isn't accurate, as of September 14, that the action stands.
"Tremendous CRIMINAL and REGULATORY power over these companies"
Taken literally, this describes authority that already exists rather than anything new: federal prosecutors retain ordinary discretion to charge companies that break the law, and the relevant regulators retain their existing statutory authority over the sector. Nothing in the post identifies a new power, a new statute, or a new investigation. What changed is that a sitting president chose to state that authority publicly, in the same paragraph as naming a specific CEO and company, days after that CEO's company had just won a federal injunction against the administration's own attempt to use exactly this kind of leverage against it. Whether that reads as a neutral statement of existing law or as pointed context for the sentences around it is a matter of interpretation this blog won't resolve for the reader — but the ambiguity itself is the notable part, not a mistake in the post.
Vance's version, without the personal target
Vice President JD Vance made a related but more measured argument to reporters the same day, without naming anyone: "I have to say, personally, I feel a little bit weird about the fact that you have so many frontier AI tech companies kind of coming to the government and begging the government to regulate them... It feels a little bit to me like a bit of a Trojan horse. So, I think we just have to be careful about this." He added that "there are obvious risks to AI, there are some benefits to AI," and said he hoped the White House and Congress would ultimately develop guardrails together.
Vance's "Trojan horse" framing is, structurally, close to the test this blog covered from François Chollet two days earlier: a warning that pacing proposals from the companies with the most to gain from them deserve scrutiny for whether they entrench the largest players rather than protect the public. Chollet's version came with two falsifiable conditions and a request for genuine international coordination. Vance's version, delivered from inside the administration currently fighting Anthropic in court over an unrelated dispute, carries a different kind of weight — not because the underlying concern is wrong, but because the speaker has direct authority to act on it, and has already tried to once.
The line with no named target
"Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE!" follows immediately after the claim of "tremendous CRIMINAL and REGULATORY power over these companies," with no name, agency, or specific event attached to either "leakers" or the people it groups with "traitors." This blog could not confirm what leak, if any, it refers to — nothing in the surrounding coverage of this post identifies one, and speculating on a specific target beyond what the text supports isn't something this post will do. What's verifiable is narrower: the sentence exists, it's addressed to unnamed people, and it sits directly beside a claim of criminal and regulatory authority over AI companies, in a post that separately names one specific CEO by first name.
The market moved that morning, which complicates a prediction from two days earlier
Nasdaq futures fell more than 1% overnight, and chip stocks were down at Monday's open — Intel around 6%, Micron around 5%, Nvidia around 3%. Jason Calacanis had predicted, two days earlier, that "AI stocks will drop 10%+ on Monday morning" in response to Amodei's essay specifically. Monday's drop was real but short of his number, and it landed on a morning that carried both the weekend's accumulated reaction to the essay and Trump's own post — which itself reads, in part, as a rejection of exactly the kind of capability-constraining regulation Calacanis's prediction assumed the market would price in. Cleanly attributing Monday's move to either the essay or the president's response to it isn't possible from the numbers alone; both happened before the same opening bell.
Where this leaves the week
Every other contribution to this week's debate — Amodei's proposal, Chollet's test, two posts from Nadella, Mensch's tweet, LeCun's mockery — was an argument about how the AI industry should govern itself, made by people whose companies had something to gain from the answer. This blog has named that self-interest in each case. Trump's post is a different kind of entry: not a proposal about governance, but a head of state publicly naming the specific person making the safety argument, invoking his administration's litigation history against that person's company as evidence of resolve, and doing so in the same statement that claims broad criminal and regulatory authority over the sector and warns unnamed others to beware. Whether the underlying policy debate about pacing gets settled on the merits Chollet's framework asks about is now a question that has to be answered inside a fight that already has a plaintiff, a defendant, and a case pending on appeal.