The facts are as follows: the actor formerly known as the Department of Defense (now called the Department of War, renamed by the great non-interventionist president of our times, because it’s more macho) had negotiated a contract with Anthropic. Under this contract, the DOW had to follow Anthropic’s fair usage policy. But then, the DOW tried to renegotiate, so that Anthropic’s AI could be used for all lawful purposes. This would allow their tech to be used for mass surveillance and autonomous weapons.
Anthropic turned them down, because they didn’t want their tech to be used for mass surveillance or autonomous weapons. The DOW found this unacceptable, because it interfered with their plans to use Anthropic’s tech for autonomous killer drones and mass surveillance.
The DOW then took an unprecedented step: they designated Anthropic a supply chain risk, meaning that doing business with Anthropic is considered a threat to American national security. This makes it illegal for companies that do business with the military to use Anthropic. Then, the DOW signed a contract with OpenAI, who after appearing to display a brief glimmer of courage, regressed back into their traditional state: ▶that of a great supine protoplasmic invertebrate jelly.
This is a clear abuse of power. The DOW is asserting its right to force private companies to sign any contract with the DOW under whichever conditions the DOW likes. If you refuse, your company will be destroyed. As Scott notes, “The “supply chain risk” designation has previously only been used for foreign companies like Huawei that we think are using their connections to spy on or implant malware in American infrastructure. Using it as a bargaining chip to threaten a domestic company in contract negotiations is unprecedented.”
It would be perfectly respectable for the DOW not to sign a contract with Anthropic. What is not responsible is to abuse the supply chain risk designation to strong-arm private companies. By the same logic, the DOW could force a private company to let them use their services for free or be designated a supply chain risk. It’s a totally crazy and un-american way of doing things. , who formerly worked with the White House, 𝕏put it very well:
Think about the power Hegseth is asserting here. He is claiming that the DoD can force all contractors to stop doing business of any kind with arbitrary other companies.
In other words, every operating system vendor, every manufacturer of hardware, every hyperscaler, every type of firm the DoD contracts with—all their services and products can be denied to any economic actor at will by the Secretary of War.
This is obviously a psychotic power grab. It is almost surely illegal, but the message it sends is that the United States Government is a completely unreliable partner for any kind of business. The damage done to our business environment is profound. No amount of deregulatory vibes sent by this administration matters compared to this arson.
Even a cursory review of the facts reveals that the “supply chain risk” allegation is totally farcical.
If Anthropic is really a supply chain risk, so that doing business with them threatens the national security of the United States, why was the DOW trying to do business with them until five minutes ago? If they’re a supply chain risk, why do DOW contractors have six months to phase out their use of Anthropic products? Surely an urgent threat to national security should be phased out sooner?
If they are a supply chain risk, then why aren’t Chinese AI companies like Deepseek considered supply chain risks? Surely Anthropic is not more of a supply chain risk than companies operating under the thumb of the CCP? And why did the government only notice that Anthropic is a threat to national security after Anthropic refused to sell their product?
Anthropic is an unusually morally upstanding company. They are unusually safety conscious. Their approach to aligning Claude is careful and decent. The people who work for them are genuinely sensible. When you read their constitution, you sense there are adults in the room. They embody exactly the values you’d want in a leading AI company.
Crushing Anthropic, one of the most innovative American companies, because they wouldn’t sign a DOW contract is idiotic, and will have a corrosive effect on the business environment. Punitively eviscerating private companies is bad both for the individual companies and for the broader business environment. It is also very likely against the law, as the claimed security risk is totally bogus.
This kind of crass bullying and abuse of power is much more typical of the CCP than America. Interestingly, this fact is admitted by certain DOW defenders:
It is true, as Mr. Carter writes, that the CCP would strong arm private companies into doing business with them, and jail those that don’t. Normally, however, we do not hold up the CCP as the ideal standard of conduct, particularly when the conduct in question is jailing dissidents. We live in America, where we have a thing called freedom!
One other very interesting approach that DOW defenders have taken has been doing a little thing called lying:
Dario isn’t in a polycule, obviously. As for the claim that he sided with Putin and Xi, if one so much as reads the second paragraph in his Wikipedia page, they will see that he is “a proponent of an "entente" strategy in which a coalition of democratic nations use advanced AI systems in military applications to achieve a decisive advantage over adversaries, while sharing the benefits with cooperating nations.” He has been a consistent proponent of stronger export controls on China. Drew is either lying or was simply unwilling to look into the facts before opining on a subject he doesn’t know the first thing about.
Anthropic didn’t cave to the CCP. It refused to cave to efforts to make America into the CCP.
Claude’s constitution, which anyone who has so much as glanced at the document would know, describes what values Claude should have. OpenAI has a constitution too (they call it a model spec, but it’s the same idea)! It makes sense to give Claude values because lots of people ask Claude for advice on moral conundrums. In addition, as Claude gets more autonomous, taking into account moral considerations will grow increasingly important. The constitution says stuff like:
Claude should try to identify the response that correctly weighs and addresses the needs of those it is helping. When given a specific task or instructions, some things Claude needs to pay attention to in order to be helpful include the principals:
[Goes on to list a bunch of sensible things]
Does this sound like a “plan to impose on Americans their corporate laws in a way that undermines the constitution?” Either Emil Michael is illiterate or he is pretending to be in order to slime Anthropic. I would like for our public officials not to openly lie to us in ways so flagrant that a minute of googling reveals that what they’re saying is false.
In addition, if Anthropic is a malignant authoritarian enterprise trying to usurp the constitution and impose their values on Americans, why was the DOW trying to do business with them? If, as Secretary of War Hegseth 𝕏suggests, Anthropic is engaged in “defective altruism,” (get it) and “seiz[ing] veto power over the operational decisions of the United States military,” then why was the DOW trying to do business with them? Our public leaders aren’t even attempting to tell a coherent story of their actions. This is the kind of petty authoritarianism combined with minimal interest in justifying their actions to the public that one sees in authoritarian countries but was, until fairly recently, not seen in America. Je suis Anthropic. Viva Anthropic.
