The Situation: Dominance Play

The Situation on Tuesday remembered the four enduring truths of the Mueller Report. 

Yesterday evening, U.S. District Judge Rita Lin of the Northern District of California issued a preliminary injunction stopping Secretary of Defense Pete Hegseth’s vendetta against Anthropic. 

The 43-page opinion will not come as any surprise to folks who have been reading or listening to or watching Lawfare over the past few weeks. Judge Lin finds—as was patently obvious—that the Pentagon’s, the president’s, and Hegseth’s actions against the AI company were retaliatory for First Amendment protected expression, unauthorized by any statute, and wrongful in treating an American company whose product has never been found to pose security problems like a foreign sabotage threat.

The case is far from over. Its sister litigation remains pending in the D.C. Circuit Court of Appeals. A quick appeal of Judge Lin’s ruling to the Ninth Circuit Court of Appeals seems all but certain; indeed, Judge Lin stayed her own ruling for seven days by way of facilitating that. And some kind of emergency docket trip to the Supreme Court seems likely in the immediate term. 

That said, I’m having a little trouble imagining the courts coming to a different conclusion from Judge Lin’s, at least based on the current record. As with the law firm cases the administration insists on appealing, you can dress this case up with a lot of fancy appellate lawyering, but it doesn’t change what is ultimately a simple fact pattern.

A company makes a product. The government wants to use that product. The government and the company cannot come to terms over the circumstances in which the government can and cannot use the product—with the company, rightly or wrongly, attaching conditions the government finds unacceptable. The company speaks publicly about the dispute. The government responds not merely by ceasing use of the product, which it has every right to do, but by labeling the product a “supply chain risk” and forbidding entities that have business ties to the company from doing business with the government. The government takes this step with specific and repeated reference to the company’s criticisms of the government and the positions it took in contract negotiations. The government has to twist the relevant statute into a pretzel to make it even remotely authorize its behavior—and even then has to acknowledge that some of the steps it took are beyond any statutory authorization. 

It’s a simple fact pattern that cannot be reconciled with a non-corrupt, market economy functioning in a civilized democratic society in which one is allowed to criticize one’s government. 

And it’s a fact pattern on which the government will not prevail—not even before a conservative Supreme Court. 

All of which raises the question: What is the government trying to accomplish here? What is the point of creating needless friction with a major up-and-coming tech giant, during a war no less, in order to set up a humiliating defeat in court at the hands of that tech giant? 

Here are two things I don’t believe this is really about: lethal autonomous weapons and mass surveillance of Americans. Don’t get me wrong. These two broad categories are what the dispute is about for Anthropic; indeed, they have become the focal point of the dispute because they happen to be Anthropic’s red lines. But I actually don’t believe that the government picked this fight because it has some top-secret drone it wants to fly that it needs Claude to pilot. And I don’t believe either that there are programs of mass surveillance targeted at Americans—at least not as the government defines “mass surveillance”—that the government regards as lawful and needs Claude to help it manage. We may be evolving towards such things, but I doubt that there is an actual surveillance program, much less an actual weapons program, that the government genuinely wants to operate but is currently stymied from doing so because of Anthropic’s terms of service.

Here’s a third thing I don’t believe this is really about: the principle that Anthropic cannot interpose itself in military operational decisions. If it had been about that, a quiet disengagement between the military and Anthropic would have been all that was necessary. Nobody—not the judge and not Anthropic either—doubts the military’s authority to cease doing business with a company that won’t provide its services on terms acceptable to the military. (That, by the way, is true of every customer.) If Hegseth’s true bottom line here was that he needs frontier AI models that will be available for all lawful uses and can’t have pesky woke tech guys imposing their values on military decisionmaking, all he had to do was take his business elsewhere. He didn’t have to set up a loser of a case in court that’s going to make the woke tech dudes into civil liberties martyrs. 

The more I think about it, the more I think this is nothing more than a dominance play. 

Harvard didn’t submit. It had to be crushed. 

Perkins Coie didn’t submit. It had to be crushed. 

James Comey and Letitia James didn’t submit. They had to be crushed. 

Six Democratic members of Congress didn’t submit. They had to be crushed. 

A guy threw a sandwich, rather than submitting. He had to be crushed. 

Kilmar Abrego Garcia didn’t submit. He had to be crushed.

Jerome Powell didn’t submit. He had to be crushed.

Anthropic didn’t submit. It has to be crushed. 

The case isn’t really about anything more than that. The underlying issue is simply the lack of submission. And the point is not to win. The point is to emphasize to the next recalcitrant entity, person, company, or institution that it will have to defend itself if it asserts its rights—because the government will relentlessly come at those who don’t submit. It will come at the little guy. It will come just as relentlessly at the multibillion dollar corporation or the university with the endowment the size of the GDP of a small country. If you’re the little guy, they will keep coming until some court lets them send you to Liberia. And if you’re a frontier AI company with the temerity not to submit, they will use the opportunity of destroying you to create opportunities for businesses associated with friendlier billionaires who know to stay on side. 

The key is to keep pushing—to raise the cost to everyone of defiance. 

That is one way of reducing defiance and enduring that The Situation continues tomorrow.