Simple rational argument:

SCOTUS says POTUS is above the law, so POTUS has collected $4B in bribe / protection money since taking office 13 months ago. Anthropic has lots of money at the moment. Why should they be allow to keep it?

Since they didn't pay off the president (enough?), his goons are going to screw with their revenue and run a PR smear campaign.

Once you realize it only has to do with Trump's personal finances, and nothing to do with national security or the rule of law, then all the administration's actions make perfect rational sense.

Open question: How much should a congress-critter charge Trump for a favorable vote? (The check should come with a presidential pardon in the envelope, of course...)

[flagged]

Except that it’s sovereign.

Because Palentir is using Anthropic.

I'm def adding "slippery semantic slopes" to my vocab.

[flagged]

We've moved beyond telling people not to forget and have entered "expect nothing less" territory

Aren't export taxes against the US constitution?

The part of this you're missing is that China doesn't want it [1].

Why? Because China will make their own. This has been obvious to me for at least 1-2 years. The US doesn't allow EUV lithography machines from ASML to be exported to China either. I believe the previous export ban on the most advanced chip was a strategic error because it created a captive market of Chinese customers for Chinese chips.

China will replicate EUV far quicker than Western governments expect. All it takes is to throw money at a few key ASML engineers and researchers and the commitment of the state to follow through with this project, which they will.

I'm absolutely reminded of the atomic bomb. This created quite the debate in military and foreign policy circles about what to do. The prevailing presumption was that the USSR would take 20 years to develop their own bomb if it ever happened.

It took 4 years.

And then in 1952 the US detonated the first thermonuclear bomb. The USSR followed suit in 1953.

[1]: https://www.tomshardware.com/tech-industry/artificial-intell...

The Trump administration tends to use this playbook.

Putting aside my take, I’m trying to objectively make sure I’m grounded on what is likely to happen next, without confusing “what is” with “what is ok”.

I think this is maybe the most revealing thing about this saga, that seemingly the U.S. government has not been training their own frontier models.

> Obviously the DoD would not want limited use.

I agree in this sense: Hegseth's Dept. of War doesn't want any restrictions. I'll try to make the case this is self-defeating, assuming one has genuine, long-term national interests at the front of mind (which I think is lacking or at least confused in Hegseth).

Historically, other (wiser) SecDefs would decide more carefully. They are aware when their actions would position DoD outside of reasonable ethical norms, as defined both by their key personnel as well as broader culture. I think they would recognize Hegseth's course of action as having two broadly negative effects:

1. Technology, Employees, Contractors. Jeopardizes DoD's access to the best technology. Undermines efforts in hiring the best people. Demotivates existing employees and contractors. Bullying leads to fearful contractors who perform worse. Fewer good contractors show up. Trumpist corruption further degrades an already lagging, sluggish, inefficient system.*

2. Goodwill & Effectiveness. Damages international goodwill that takes a long time to restore. Goodwill is a good investment; it pays dividends for U.S. military strength. The fallout will distract Hegseth from legitimately important duties and further undermine his credibility. Leading probably to a political mess for Hegseth, undermining his political capital.

* Improving DoD procurement is already hard given existing constraints. Adding Trumpist-level corruption makes it unnecessarily worse. There is already an unsavory, poorly tracked, bloated gravy train around the military industrial complex.**

** BUT... Despite all this, the system has more or less worked reasonably well for more than what, 80 years! It has enjoyed bipartisan continuity, kept scientists and mathematicians well funded, and spurred lots of useful industries. It is, in a weird gnarly way, a sort of flux capacitor for U.S. technical dominance.

There has been a terrifying decline in quality and an increase in corruption in Trump’s second administration.

Re: the hilarity part, I’m conflicted: in general, a good sense of humor is useful, but in present circumstances a stoic seriousness seems warranted.

Any documentation regarding the claim about breaking their contract?

Haven't heard that. Regardless, as someone who works with these models daily (as well as company leadership that loves AI more than they understand it) - Anthropic is absolutely right to say that the military shouldn't be allowed to use it for lethal, autonomous force.

The United States has freedom of speech. The Supreme Court has ruled that money is speech. A company can always direct their money, speech, however they like with regards to the government. Can you be sued for breach of contract? Sure. Is it a supply chain risk absolutely not.

> They are a "supply chain risk" if they can willy-nilly break their contract with US govt and enforce arbitrary rules to service.

It is the US govt that seeks to break their contract with Anthropic.

The contract they signed had the safeguards, so they were mutually agreed upon. These safeguards against fully autonomous killbots and AI spying of US citizens was known before signing.

This conflict now is because the US govt regrets what they agreed to in the contract.

> completely understandable decision from a neutral third party PoV.

Except it's not, really. If Anthropic/Claude doesn't mean the DoD's need, they can and should just put out an RFP for other LLM providers. I'm sure there's plenty of others that'd happily forgo their morals for that sweet government contract money.

No US company has to provide services to the DoD or any other branch of government. It's not "veto power" it's being selective of who you do business with, which is 100% legal.

Then you go to another supplier. But any company with proper counsel will tell them the same thing: don't break the law, which is exactly what they're trying to coerce Anthropic into doing. DoD requests do not supersede the law.

Not unless they're the sole supplier of the technology. They're saying, if you want to do this kind of thing - not with our product, but you can get it elsewhere.

No, you are the one lying trying to get political gotchas here. There is no "trying to exert veto power" absolutely anywhere, Anthropic's terms were laid out in the contract the Pentagon signed, which they want to forcibly amend. If they didn't like the terms, they didn't need to sign the contract.

Or worse: train the AI to make decisions that align with the view of Anthropic management and not the elected government. Workout telling anyone.

I’d agree it is a serious risk.