I would assume the original terms the DoW is now railing against were in those original contracts that they signed. In that case it looks like the DoW is acting in bad faith here, they signed the original contact and agreed to those terms, then they went back and said no, you need to remove those safeguards to which Anthropic is (rightly so) saying no.
Am I missing something here?
EDIT: Re-reading Dario's post[1] from this morning I'm not missing anything. Those use cases were never part of the original contacts:
> Two such use cases have never been included in our contracts with the Department of War
So yeah this seems pretty cut and dry. Dow signed a contract with Anthropic and agreed to those terms. Then they decided to go back and renege on those original terms to which Anthropic said no. Then they promptly threw a temper tantrum on social media and designated them as a supply chain risk as retaliation.
My final opinion on this is Dario and Anthropic is in the right and the DoW is acting in bad faith by trying to alter the terms of their original contracts. And this doesn't even take into consideration the moral and ethical implications.
[1]: https://www.anthropic.com/news/statement-department-of-war
The basic problem in our polity is that we've collectively transferred the guilty pleasure of aligning a charismatic villain in fiction to doing the same in real life. The top echelons of our government are occupied by celebrities and influencers whose expertise is in performance rather than policy. For years now they've leaned into the aesthetics of being bad guys, performative cruelty, committing fictional atrocities, and so forth. Some MAGA influencers have even adopted the Imperial iconography from Star Wars as a means of differentiating themselves from liberal/democratic adoption of the 'rebel' iconography. So you have have influencers like conservative entrepreneur Alex Muse who styles his online presence as an Imperial stormtrooper. As Poe's law observes, at some point the ironic/sarcastic frame becomes obsolete and you get political proxies and members of the administration arguing for actual infringements of civil liberties, war crimes, violations of the Constitution and so on.
> *Isn’t it unreasonable for Anthropic to suddenly set terms in their contract?* The terms were in the original contract, which the Pentagon agreed to. It’s the Pentagon who’s trying to break the original contract and unilaterally change the terms, not Anthropic.
> *Doesn’t the Pentagon have a right to sign or not sign any contract they choose?* Yes. Anthropic is the one saying that the Pentagon shouldn’t work with them if it doesn’t want to. The Pentagon is the one trying to force Anthropic to sign the new contract.
[1]: https://www.astralcodexten.com/p/the-pentagon-threatens-anth...
> [1] This story requires some reading between the lines - the exact text of the contract isn’t available - but something like it is suggested by the way both sides have been presenting the negotiations.
I deal with far too many people who won't believe me without 10 bullet-proof sources but get very angry with me if I won't take their word without a source :(
> "Two such use cases have never been included in our contracts with the Department of War..."
While I agree with Anthropic's position on this regardless, the original contract wording does matter in terms of making either the government look even more unreasonable or Anthropic look a little less reasonable.
The issue is a subtle ambiguity in Dario's statement: "...have never been included in our contracts" because it leaves two possibilities: 1. those two conditions were explicitly mentioned and disallowed in the contract, or 2. they weren't in the contract itself - and are disallowed by Anthropic's Terms of Service and complying with the ToS is a condition in the contract (which would be typical).
If that's the case, then it matters if the ToS disallowed those two uses at the time the original contract was signed, or if the ToS was revised since signing. Anthropic is still 100% in the right if the ToS disallowed these uses at the time of signing and the ToS was an explicit condition of the contract, since contracts often loop in the ToS as a condition while not precluding the ToS being updated.
However, if the ToS was updated after contract signing and Anthropic added or expanded the wording of those two provisions, then the DoD, IMHO, has a tiny shred of justification to complain and stop using Anthropic. Of course, going much further and banning the entire US government (and contractors) from using Anthropic for any use, including all the ones where these two provisions don't matter - is egregiously punitive and shitty.
While the contract wording itself may be subject to NDA, it would be helpful if Anthropic's statements could be a bit more precise. For example, if Dario had said "have always been disallowed in our contracts" this ambiguity wouldn't exist.
You're trying desperately to find a way that things can be at least a little normal, and I really do get it. It would be great if such a way existed. But it doesn't. I recommend you take a social media break like I'm about to, take the time you need to mourn the era of normal politics, and come back with a full understanding that the US government is not pursuing normal policy objectives with bad decisions. They hate you and they hate me for not being on their side, and their primary goal is to ensure that we're as miserable as they can make us.
Disregarding who is right or wrong for a moment, if the DoW are right (which I'm not personally inclined to believe, but we're ignoring that for the moment) -- how else can they avoid secondhand Claude poisoning?
Supposing they really want to use their software for things disallowed by Claude's (now or future) ToS, it seems like designating it a supply chain risk is the only way they can ensure that their contractors don't include Claude (either indirectly as a wrapper or tertially through use of generated code etc)
I agree that if the DoW claim is correct (and I doubt it is), then, sure, the DoW dropping Anthropic and precluding the DoW's suppliers from using Anthropic for any DoW work would be expected. However, the "supply chain risk" designation they are deploying goes far beyond that to block Anthropic use by any supplier to any part of the entire U.S. government for anything.
For example, no one at Crayola can use Anthropic for anything because Crayola sells crayons to the Education Dept. The DoW already has much less draconian ways to restrict what their direct suppliers use to build things for military applications. But instead of addressing the actual risk in a normal measured way, they are choosing to use a nuke against a grenade-sized problem. This "supply chain risk" designation is rarely used and has never been used against a U.S. company. It's used against Chinese or Russian companies when in cases where there's credible risk of sabotage or espionage. That's why that particular designation always blocks all products from an entire company for any application by any part of the U.S. Government, contractors and suppliers (which is why it's never been used against a U.S. company).
My hope is that this gets us some real concern for things that have been defended with de facto arguments (i.e. privacy) going forward.
edit: Anthropic argues that your Crayola analogy is fundamentally incorrect.
> Legally, a supply chain risk designation under 10 USC 3252 can only extend to the use of Claude as part of Department of War contracts—it cannot affect how contractors use Claude to serve other customers.
https://www.anthropic.com/news/statement-comments-secretary-...
> The Leftwing nut jobs at Anthropic have made a DISASTROUS MISTAKE trying to STRONG-ARM the Department of War, and force them to obey their Terms of Service instead of our Constitution.
(Or maybe it’s catchier to respond glibly with “never trust a child rapist and convicted felon.”)
The Pentagon feels it isn't Anthropic to set boundaries as to how their tech is used (for defense) since it can't force its will, then it bans doing business with them.
Once the hive decides you're being serious without checking, they turn the down vote button into an I disagree with you button.
This is actually one of the reasons I left Reddit. I hate to see it here.
I may not agree with what people say and it seems like he may have just been kidding or was being sarcastic, but he should be allowed to say it without being bullied and abused by downvotes.
I hope everyone will reconsider their ways.
So it's six months to phase out use of Anthropic at the DoD, but the people hosting the models have to stop "immediately".
Which miiight impact the amount of inference the DoD would be able to get done in those six months.
> Which miiight impact the amount of inference the DoD would be able to get done in those six months.
Which might not be by accident looking at the Truth Social posts which state "Anthropic better get their act together, and be helpful during this phase out period, or I will use the Full Power of the Presidency to make them comply, with major civil and criminal consequences to follow."
I would not be surprised to see this being used as an excuse to nationalize Anthropic.
The US isn't Iran, North Korea, or even China, as much as some people, including the US president, seem want to emulate those models.
According to whom?
No, not this room. The one with Hegseth in it.
Look at his other comments. He's not wrong.
Imagine a _leaded_ pipe supplier not being allowed to tell the department of war they shouldn't use leaded pipes for drinking water! It's the job of the vendor to tell the customer appropriate usage.
Go look at the package on a kitchen knife and it says not to be used as a weapon
I'm not sure this is still a useful analogy, though...
The knife manufacturer isn't obligated to sell to you in either case, I'd expect them not to cut ties with you in the self defence scenario. But it is their choice.
1. Found out you used their knives to go murdering
2. Sells knives in a fashion where it's possible for them to prevent you from buying their knives (i.e. direct to consumer sales)
Would almost certainly not "be more than happy to continue to sell to you". Even if we ignore the fact that most people are simply against assisting in murders (which by itself is a sufficient justification in most companies), the bad PR (see the "found out" and "direct to consumer" part) would make you a hugely unprofitable customer.
> the bad PR (see the "found out" and "direct to consumer" part) would make you a hugely unprofitable customer.
That... Doesn't happen.
Boycotts by people who weren't going to buy your product anyway are immaterial to business. The inevitable lawsuits are costly, but are generally thought of as good publicity, because they keep the business name in the news.
People who buy (and make) firearms are... pretty close to the exact opposite.
Also, in the cases you can't, it's generally the government stopping you, not the gun companies.
Edit: hell I get downvoted and look where the knife analogy got us. A load of weird replies miles away from anything related to AI or DoD.
Claude Opus is just remarkably good at analysis IMO, much better than any competitor I’ve tried. It was remarkably good and complete at helping me with some health issues I’ve had in the past few months. If you were to turn that kind of analytical power in a way to observe the behaviour of American citizens and to change it perhaps, to make them vote a certain way. Or something like - finding terrorists, finding patterns that help you identify undocumented people.
What we're seeing here is that when a vendor declines to change the terms of its contractual agreement for ethical reasons, the government publicly attacks it.
With the other stated reason being legal. "To the extent that such surveillance is currently legal, this is only because the law has not yet caught up with the rapidly growing capabilities of AI."
I don't think we should lessen Anthrophic's stance from technical/legal to ethical. Just as we shouldn't describe what the department of war is doing as "not renewing a contract".
Congress could come up with an act it it's for national interest.
The military isn't the typical End User.
Utter nonsense. When the US built the Blackbird, it could only use titanium because of the heat involved in traveling at that speed. But they didn't have enough titanium in the US. So the the US created front companies to purchase titanium from the Soviet Union.
Do you think the US should have informed the Soviet Union what it wanted to do with the metal?
Your comparison seems backwards
If this were a news outline writing "Department of War" I would be concerned. But in the case of the Anthropic CEO's blog post, I can understand why they are picking their fights.
It's a testament to the broken information ecosystem we're in that many people genuinely don't know this. Most will correct themselves when told. I agree with you that those who don't are not worth engaging.
Sad.
They want the department to fight wars. At least they’re being honest.
I can't see anyway this ends well for the US. I say this as both an American and a military veteran.
Frederick VII of Denmark, an absolute monarch, introduced parliamentarism without any violence or even broad public pressure.
And thats just what I can remember without digging.
[1] "only an act of Congress can formally change the name of a federal department." https://en.wikipedia.org/wiki/Executive_Order_14347
(edited to add the url I omitted)
"You can just do things" (evil edition).
So one thing to call out here is that the assumption that DoW is working on specifically these use cases is not bullet proof. They simply may not want to share with anthropic exactly what they are working on for natsec issues. /we can't tell you/ could violate the terms.
It is also dumb that DoW accepted these terms in the first place.
They will just have to recompete!
Trump has historically stiffed his contractors. Why do you think his administration would be any different with adhering to a contract?
A fact being technically available and that fact commanding widespread public attention are very different things. Anthropic's communications team understands this distinction even if you don't find it interesting. The blog post wasn't written for people who already track federal AI contracts, it was written for the much larger audience encountering this story for the first time and forming opinions about it in real time.
If the point you're making is just "I already knew this," that's fine, but it doesn't address anything about the incentive structure behind the public response.
You're ignoring the sequence of events on the ground.
If there hadn't been any been any internal pushback from Anthropic, would the directive have ever been made public?
Internal pushback and public damage control aren't mutually exclusive. A company can genuinely disagree with a client's demands behind closed doors and simultaneously craft a public narrative designed to make itself look as good as possible once those disagreements surface. In fact, that's exactly what competent communications teams do, they plan for the scenario where private disputes become public, and they have messaging ready.
The real question isn't who went public first or why. It's whether Anthropic's stated position, "we support these military use cases but not those ones", reflects a durable ethical framework or a line drawn precisely where it needed to be to keep both the contracts and the brand intact. Nothing in the sequencing you've described answers that question. It just tells us Anthropic saw this coming, which, if anything, means the messaging was more carefully engineered, not less.
[Edit: update]
I use assistive voice transcription because I'm unable to type well with a keyboard. But I'd point out that "you must be an AI" has become the new way to dismiss an argument without engaging with it. It's the modern equivalent of "you're just copy-pasting talking points", it lets you discard everything someone said without addressing a single word of it.
The fact that my sentences "thread together" is not evidence of anything other than coherent thinking. And speed of response says more about the tools someone uses than whether a human is behind them. Plenty of people use dictation, accessibility tools, or just happen to type fast.
^^^ This took me 30 seconds to speak aloud.
My speculation is the "business records" domestic surveillance loophole Bush expanded (and that Palantir is build to service). That's usually how the government double-speaks its very real domestic surveillance programs. "It's technically not the government spying on you, it's private companies!" It's also why Hegseth can claim Anthropic is lying. It's not about direct government contracts. It's about contractors and the business records funnel.
Of course they can just say - we aren’t, Palantir is.
[1] https://www.anthropic.com/news/statement-department-of-war
What will the new American reich accomplish?
Likely the same thing as all the proceeding empires - carnage, destruction, and the laughter of blood thirsty gods.
It's insane how numb I am becoming to these blurry thin lines
People like married women who changed their name, or foreign sounding people, they will be prevented to vote in 2026. ICE will guard polls to physically make people unable to reach the ballots
Specifically, he would need the US Congress to draft and pass legislation moving the date of the election. I don't know how eager they are, though, to create an unnecessary constitutional crisis.
This whole thing seems like people talking past each other, and that there’s something being left unsaid.
Anthropic doesn’t make a product that would assist with kill drones, and they don’t have the right to deny subpoenas.
I take that to mean they don't want the military using Claude to decide who to kill. As a hyperbolic yet frankly realistic example, they don't want Claude to make a mistake and direct the military to kill innocent children accidentally identified as narco-terrorists.
At least, that's the most charitable interpretation of everything going on. I suspect they are also worried that the sitting administration wants to use AI to help them execute a full autocratic takeover of the United States, so they're attempting to kill one of the world's most innovative companies to set an example and pressure other AI labs into letting their technology be used for such purposes.
Obviously the military wants to use it for that purpose since they couldn't accept Anthropic's extremely limited terms.
One can easily and immediately infer the answers to both your questions are yes.
Anthropic’s safeguards already prevent what you are describing, again the thing thar DoW has said they don’t want.
Giving precedence to words over actions is how you get taken advantage, abused, deceived, etc.
But Dario is showing weakness here by talking around it. Whatever they were asked to do, they should just be upfront about.
Anthropic is not being asked to do anything, except renegotiate the contracts. The DoW Claude models run on government AWS. Anthropic has minimal access to these systems and does not see the classified data that is being ingested as prompts. It is very unlikely that Dario actually knows what the DoW wants to do with these models. But even if he did, it would be classified information that he is not at liberty to disclose.
However the product they provide likely has safety filters that cause some prompts to not be processed if it is violates the two contractual conditions. That is what the DoW wants removed.
Did the DoW ask them to make kill drones? Because if so THAT IS A REALLY BIG DEAL.
The vagueness is irritating. He’s saying they won’t do something, the DoW is saying they don’t even want them to do that, which should resolve the issue, but hasn’t. There is obviously something else at play here.
Here's the Chief Pentagon Spokesman pointing to the same verbiage and reiterating they they won't agree to those terms of use.
> The Department of War has no interest in using AI to conduct mass surveillance of Americans (which is illegal) nor do we want to use AI to develop autonomous weapons that operate without human involvement.
Tomorrow he could change his mind to "we want to use AI to develop autonomous weapons that operate without human involvement." the issue is that he wants Anthropic to change the use terms because "We will not let ANY company dictate the terms regarding how we make operational decisions."
>>no he didn’t he actually said the opposite of that and the link you just posted says the opposite of what you are claiming
>but he might change his mind!
Okay?
>Did the DoW ask for these things?
>Did the DoW ask for that?
I showed you where the spokeperson asked for the terms to change so they could make autonomous weapons. now, you're shifting the goal posts.
Is a pundit/politician lying to you a new experience?
If the DoD did not want those things, it would not be forcing a contract renegotiation to include them, at great cost to the government.
That’s the point I’m trying to make here: Anthropic should just say the unsaid thing here.
DoW asked for the following thing: $foo. We won’t give that to them.
> DoW asked for the following thing: $foo. We won’t give that to them.
Anthropic has explicitly said that multiple times, including in the letter we are presently discussing.
$foo is the ability to use Claude for domestic mass surveillance and analysis, and/or fully-autonomous killbots.
Where is all the weird misinformation in these comments coming from?
OpenAI has already said that they’ll give up whatever info the government wants if they’re issued a subpoena; they don’t have a choice.
If anthro is saying they won’t, that’s good!
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...)
Further, what law lets the government dictate what contracts a company signs? Anthropic refused to work with them. We had a whole Supreme Court case about refusing working with customers.
But that's not the case, is it? The government can say that it's legally required to give Donald Trump a gold bar every Sunday. That wouldn't even be too far off from the outlandish claims we've seen over the past year. The Trump administration is, as Chapelle would put it, a habitual line stepper.
(What is obvious is the kind of response I will get, which is why I will ignore it and not comment further.)
I got the meaning right away, but I can appreciate if others didn't. I didn't read it as intentionally enigmatic, fwiw. Sometimes short punchy comments really land, sometimes not -- it is a risk. As you can probably tell, I err in the other direction. (:
[...]
He has endeavoured to prevent the population of these States; for that purpose obstructing the Laws for Naturalization of Foreigners; refusing to pass others to encourage their migrations hither, and raising the conditions of new Appropriations of Lands.
[...]
He has obstructed the Administration of Justice, by refusing his Assent to Laws for establishing Judiciary powers.
He has made Judges dependent on his Will alone, for the tenure of their offices, and the amount and payment of their salaries.
He has erected a multitude of New Offices, and sent hither swarms of Officers to harrass our people, and eat out their substance.
He has kept among us, in times of peace, Standing Armies without the Consent of our legislatures.
He has affected to render the Military independent of and superior to the Civil power.
[...]
For Quartering large bodies of armed troops among us:
For protecting them, by a mock Trial, from punishment for any Murders which they should commit on the Inhabitants of these States:
For cutting off our Trade with all parts of the world:
For imposing Taxes on us without our Consent:
For depriving us in many cases, of the benefits of Trial by Jury:
For transporting us beyond Seas to be tried for pretended offences:
— The Declaration of Independence https://www.archives.gov/founding-docs/declaration-transcrip...
People threw tea in Boston Harbor over less than the tariffs.
[1] https://www.nytimes.com/2026/02/27/technology/defense-depart...
Obviously Palentir and others need time to migrate off Anthropic’s products. The way i read it is that Anthropic made a serious miscalculation by joining the DoD contracts last year, you can’t have these kind of moral standards and at the same time have Palentir as a customer. The lack of foresight is interesting.
1 https://www.axios.com/2026/02/15/claude-pentagon-anthropic-c...
Congress is negligent in not reigning this kind of thing in. We’re rapidly falling down so many slippery semantic slopes.
For this administration the law isn't something that binds them, but something they can use against others.
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...
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”.
Obviously the DoD would not want limited use. Strange they don't make their own given their specific needs.
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.
That does seem to be what Hegseth is arguing, yes; and that is presumably his justification for doing something drastic here. Although I assume he is lying or wrong.
And as a cynic, let me just add that the image of someone going to the political overseers of the US military with arguments about being "effective" or "altruistic" is just hilarious given their history over the last ~40 years.
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.
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.
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.
source?
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.
> has never been invoked against an American company.
There's always a first. I am assuming it is not illegal to do that. It's a completely reasonable business decision to ensure your supply chain does not depend on things that may change against your goals. For example, you don't want to build or depend on an open source platform that you know is gonna rug pull, if you count on it remaining open source, do you? American or otherwise.
As to this being 100% legal, I'm not so sure (not a lawyer). It might not be a criminal offese, but there's a whole category of abuse of power that this may fall under if Anthropic is put under a certain status without real justification. Many powers given to the executive branch are not absolute and can't be applied arbitrarily, but require justification. Anthropic might be able to sue the government for declaring them a "supply-chain risk" without sufficient justification. E.g. they could claim that not being sufficiently patriotic in the eyes of the administration does not constitute a risk, and that since their not the sole supplier of the tech, they were not trying to strong arm the government to do anything.
> does not "influence their supply chain"
I would be wary of making this conclusion. Obviously it could conceivably influence the supply chain when you build on top of their model. If you look at the type of risks enumerated in DoD guidelines, it is not just "oh this software has vulnerability" which is what started the discussion in this subthread in the first place. There are many kinds of risks DoD needs to address, none are particularly new; including Sustainment Risk. The closest thing I remember to this case was Sun Java "no use in nuclear facility" EULA term, which LLM suggests was ignored by DoE/D because that was interpreted as a "limitation on warranty" not a "restriction of use."
I understand 'goals' and 'means to an end', but this concept of "law" evades me.
> Anthropic's terms were laid out in the contract the Pentagon signed, which they want to forcibly amend.
It's called negotiation in business. I am sure both sides are clear-eyed on what the consequences were and Anthropic made a calculated bet (probably correctly) that some segment of their employee/customer base would get wet by hearing this news and it more than offsets the lots business, thus is worth it.
Companies have gone out of business due to a big customer pulling the contract. Imagination Technologies comes to mind. This is not a rare thing in business.
> full power of the US government
Haha, I can assure you that is not even close to the full power of US government. Ask the crypto people during Biden admin for just a little more power (still not even close to "full.")
For a company of Anthropic's size, this may very well be a death sentence, even if their work has nothing to do with the military supply chain. They could have just canceled the contract, but they wanted to go full Darth Vader on them to prove a point in case anyone else thought about "negotiating" "voluntarily" with the federal government.
People have noticed.
> It's called negotiation in business.
The bad faith in this statement alone is almost equal to the sum of it in the rest of your comments.
The contract, including Anthropic's redlines, was signed more than a year ago and has been humming along with no objections from anybody. Hegseth abruptly got a bug up his ass about it last week, and demanded Anthropic sign a revised version under threat of punishment. Anthropic is simply saying "no, we will not be forced into signing a new version, you can either keep going with the original terms we all agreed to, or stop using us". The Pentagon can simply stop using Anthropic if they don't like the terms anymore (which, again, are the terms Pentagon agreed to in the first place). But what the DoW wants is to strong-arm Anthropic, using the DPA, into new terms because they abruptly changed their mind. That's not "negotiation" in any sense, that's Mafia behavior.
The actual terms of the contract aren't even relevant, this is purely a matter of tort law and whether you can bully someone into a new contact because you woke up one day and decided you didn't like the one you agreed to.
> Anthropic's terms were laid out in the contract the Pentagon signed, which they want to forcibly amend.
They want to "forcibly amend" is either within their rights per original contract, or not. One is fair game, the other is not.
Isn't agreeing to amend a contract always within their rights?
That said, many government contracts include some variant of "we can cancel at any time for any reason".
I’d agree it is a serious risk.
The current government is deeply unpopular, it's only going to get worse for them.
Whether they are good or evil depends on the hands that hold it.
In good hands, weapons provide defense, deterrence, and protection.
In bad hands, weapons hurt the innocent, instill fear, and oppress.
The hands that wield them make all the difference.
> Costco is a really popular subject for business-success case studies but I feel like business guys kinda lose interest when the upshot of the study is like "just operate with scrupulous integrity in all facets and levels of your business for four decades" and not some easy-to-fix gimmick
https://bsky.app/profile/mtsw.bsky.social/post/3lnbrfrvmss26
At peak times they run out of carts and tell the customers to go hunting in the lot for them, door greeters shouting at members across the floor, checkout queues stretch the length of the warehouse, they start half blocking the gas station entrance 30mins before close so trucks can't get in, so maybe they're turning those profit screws.
It's not their job to entertain you.
No, OP is right, their PR department is doing a great job.
Note that they are still releasing interesting research
That kind of good.
Anthropic is holding firm on incredibly weak red lines. No mass surveillance for Americans, ok for everyone else, and ok to automatic war machines, just not fully unmanned until they can guarantee a certain quality.
This should be a laughably spineless position. But under this administration it is taken as an affront to the president and results in the government lashing out.
It's probably in Anthropic's interest to throw grok to these clowns and watch them fail to build anything with it for 3 years.
Also, every other company has bent the knee and kissed the ring. And the trump admin will absolutely do everything they can to not appear weak and harm Anthropic. If it was so easy to act principled, don't you think other companies would've refused too? Eg Apple
And there is real harm here. You're reading about it - they get labeled a supply chain risk. This is negative and very tangible
Blood is on their hands already
What was Anthropic’s role in the Maduro operation? (Or we can call it state-sponsored kidnapping.) Who knew what and when? Did A\ find itself in a position where it contradicted its core principles?
More broadly, how does moral culpability work in complex situations like this?
How much moral culpability gets attributed to a helicopter manufacturer used in the Maduro operation? (Assuming one was; you can see my meaning I hope.)
P.S. Traditional programming is easy in comparison to morality.
In fact, as a patriotic American veteran, I'd be ok with Anthropic moving to Europe. It might be better for Claude and AGI, which are overriding issues for me.
Rutger Bregman @rcbregman
This is a huge opportunity for Europe. Welcome Anthropic with open arms. Roll out the red carpet. Visa for all employees.
Europe already controls the AI hardware bottleneck through ASML. Add the world's leading AI safety lab and you have the foundations of an AI superpower.
Anthropic made it quite clear they are cool with spying in general, just not domestic spying on Americans, and their "no killbots" pledge was asterisked with "because we don't believe the technology is reliable enough for those stakes yet". The implication being that they absolutely would do killbots once they think they can nail the execution (pun intended).
I suppose you could say they're taking the high road relative to their peers, but that's an extremely low bar.
Seems like legally the answer is "no".
But it also seems like practically the answer is "definitely".
LeCun is starting is own thing, I doubt he wants to drop it? He also lives in NYC afaik, he is a professor at NYU.
But, is there a safe haven that'd stand up against the blatant bullying and daily (or more frequent) national threats/trolling (which often stem from social media and sometimes become reality)?
For Americans and international researchers it's easy to get visas there quickly. It's not far at all for Americans to relocate to or visit. Electricity is cheap and clean. Canada has the most college educated adults per capita. The country's commitment to liberalism, and free markets, is also seeming more steadfast than the US at this point in time.
Canada faces obstacles with its much smaller VC ecosystem, its smaller domestic market, and the threat of US economic aggression. Canada's recent trade deals are likely to help there.
I say this all as an American who is loyal to American values first and foremost. If the US wants to move away from its core values I hope other countries, like Canada or the EU, can carry on as successful examples for the US to eventually return to.
Within the Claude Constitution, the words "non-western" do not appear. Where is your quote from?
The right has far more talk of violence, true, but a lot of that is targeted rhetoric to keep voters riled up, and it’s not aimed at American businesses. I’d be surprised if even a third of Republicans supported anything more than not doing business with Anthropic. Even the Nvidia shakedown got a ton of criticism and that’s just money.
As if at this point "the Republicans" have a say or want to have a say in almost anything. They are either scared shitless of who he will come after next or just want the transfer of power to be absolute and are enjoying this unchecked power and want to reap all the benefits. I don't think they want this surreal spectacle of grab and abuse of power to end. So is this a disconnect? Or do people still believe the USA's ruling party and head of state and his select lackeys are doing things by process?
I'm sure this doesn't apply to you since you're not Lord Kelvin. On the other hand, people like Peter Norvig state in a popular AI textbook that, for example, they don't know why similar concepts appear close by in the vector space, so maybe you just know something other people don't.
False, and you've given no argument to the contrary. There's certainly no definition that precludes it. It isn't, currently; there's no reason it can't be, any more than there's reason that Conway's Game of Life can't be, given sufficiently interesting data to process. Any Turing-complete system could simulate AGI. It might not be the most efficient mechanism for doing so, but that's not the question at hand.
Which of the European cultures is "underdeveloped", exactly?
But how do you even begin to discuss that Tweet or this topic without talking about ideology and to contextualize this with other seemingly unrelated things currently going on in the US?
I genuinely don't think I'm conversationally agile enough to both discuss this topic while still able to avoid the political/ideological rabbit-hole.
> Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.
If a commenter who supports the government makes the same argument that the government is making, the guidelines tell us to assume good faith.
My conclusion is that any topic where a commenter might be making a bad faith argument is outside the scope of Hacker News.
This is more for “assume op is not a troll” rather than “assume Donald trump never took part on Epstein’s parties”.
I’ve never taken it to apply to anything other than the interaction with other commenters.
Politics and ideology are not off topic, provided the subject matter is of interest, or "gratifying", to colleagues in the tech/start-up space.
What's important is that we don't use rhetoric, bad faith or argumentation to force our views on others. But expressing our opinions about how policy affects technology and vice versa has always been welcome, in my observation.
So, what do you think about the US government's decision, and why?
Everything is politics and "ideology"
HN likes to pretend otherwise, especially when it's inconvenient.
Our whole society runs on technology. All tech is inherently political.
A "no politics" stance is merely an endorsement of the status quo.
The designation says any contractor, supplier, or partner doing business with the US military can’t conduct any commercial activity with Anthropic. Well, AWS has JWCC. Microsoft has Azure Government. Google has DoD contracts. If that language is enforced broadly, then Claude gets kicked off Bedrock, Vertex, and potentially Azure… which is where all the enterprise revenue lives. Claude cannot survive on $200/mo individual powerusers. The math just doesn’t math.
Anthropic is going to be fine. The DoD is going to walk this back and pretend it never happened to save face.
Anthropic is not even close to too big to fail. And even if this could get settled in court 5 years from now, this can easily throw enough of a wrench into their revenue streams to kill their flywheel.
Think of it this way: each of the hyperscalers have built a handful of data centers specifically for government contracts. A handful each.
Meanwhile, AWS and GCP have dedicated over 50 new data centers solely for Anthropic to train new models, and more were announced today.
My bet is on Anthropic.
Those datacenters are AWS infrastructure that Amazon owns and can repurpose. The equity stake is the only part that’s truly at risk, and $8B is a rounding error on Amazon’s balance sheet.
https://www.businessinsider.com/amazon-ai-bet-anthropic-soar...
It will be interesting to see how they handle this.
Ironically, of all things Trump has done so far, closing Anthropic could set a new record for pissing off the highest number of people globally. Outside of HN with a group of dedicated people who is against it, the whole global software world is already running on CC.
Also $1T is dishonest. DoD spends less than 0.1% of that on cloud services.
Half of that budget gets contracted out to Lockheed, Raytheon, Northrop, Boeing, General Dynamics, etc. Those companies absolutely do spend money on the hyperscalers.
If you're honest with yourself, you'll find the true number.
Look, it’s a very simple question: Amazon has invested $8b into anthropic. Do you think if the DoD disappeared tomorrow that Amazon would lose more than $8b in revenue over the next 5 years?
I think you underestimate how large the DoD budget is and how many times that money changes hands in the pursuit of fulfilling contracts. $20b-$25b in revenue per year across all hyperscalers is a totally reasonable estimate.
Also, AWS has a long history of rolling over when politicians make noise about AWS customers, going back to when Joe Lieberman casually asked Bezos to please stop supporting Wikileaks.
This announcement has made Anthropic toxic in the entire dependency chain because it means years of efforts and tens to hundreds of millions of dollars rearchitecting entire platforms and renegotiating contracts.
The entire cybersecurity industry has a TAM of $208 BILLION [0]
[0] - https://www.bccresearch.com/market-research/information-tech...
This is exactly why this announcement has not made Anthropic toxic. The entire industry knows how ridiculous this move is from Hegseth, and it’s going to be rolled back next week once the adults get back from their weekend.
The downstream effects of this are HUGE.
The designation only applies to projects that touch the federal government, or software developed specifically for the federal government.
Contractors can still use Claude internally in their business, so long as it is not used in government work directly.
A complete ban would be adding Anthropic to the NDAA, which requires congress.
The DoD designation allows the DoD to make contractors certify that Anthropic is not used in the fulfillment of the government work.
" Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic."
Is that just his fantasy or?
https://www.sec.gov/Archives/edgar/data/1018724/000101872423...
https://www.anthropic.com/news/claude-in-amazon-bedrock-fedr...
The supply chain risk directive would come from existing procurement law, which only allows the DoD to require contractors to certify that Anthropic is not used in the fulfillment of any government work.
Which is also separate from Trumps' EO, which being an EO only applies to the federal government directly.
So yeah, banning any contractor, supplier, or partner from any commercial activity with Anthropic is just fantasy without going through congress first.
Anybody with significant contracts with the DOD is not going to use anthropic because they want to keep getting contracts with the DOD.
I work in the enterprise SaaS and cybersecurity industry. There is no way to guarantee that amongst any FedRAMP vendor (which is almost every cybersecurity and enterprise SaaS or on their roadmap).
Almost all FedRAMP products I've built, launched, sold, or funded were the same build as the commerical offering, but with siloed data and network access.
This means the entire security and enterprise SaaS industry will have to shift away from Anthropic unless the DPA is invoked and management is changed.
More likely, I think the DoD/DoW and their vendors will force Anthropic to retrain a sovereign model specifically for the US Gov.
Edit: Can't reply
> This is the core assertion that is not clear nor absolute.
If Walmart can forcibly add verbiage banning AWS from it's vendors and suppliers, the US government absolutely can. At least with Walmart they will accept a segmented environment using GCP+Azure+OCI. Retraining a foundational model to be Gov compliant is a project that would cost billions.
By declaring Anthropic a supply chain risk, it will now be contractually added by everyone becuase no GRC team will allow Anthropic anywhere in a company that even remotely touches FedRAMP and it will be forcibly added into contracts.
No one can guarantee that your codebase was not touched by Claude or a product using Claude in the background, so this will be added contractually.
You can add new language to new contracts. That is not what this is.
This is the core assertion that is not clear nor absolute.
That label forbids contractors on DoD contracts for billing DoD for Anthropic, or including Anthropic as part of their DoD solution.
So - AWS can keep claude on bedrock, but can't provide claude to the DoD under its DoD contracts
I am both dumb and without access to Claude, thus I must ask: My fellow smart HN'ers, what kind of impacts would this likely have on the economy?
Has a lot of money and resources not been pumped into Anthropic (albeit likely less than OpenAI)? I imagine such a decision would not be the ROI that many investors expected.
This is the new McCarthyism. Do what the administration says, or you will be blacklisted, or worse.
This is authoritarian behavior. You're having trouble negotiating a contract, so instead of just canceling it - you basically ban all of F500 from doing business with that firm.
Soon enough the midterms will be effectively cancelled.
Americans remain blissfully unaware.
We are in a bad place right now, that is for certain.
> Soon enough the midterms will be effectively cancelled.
That would be a pretty big leap from where we are. I think it is important to pay attention, and very important to vote, but there is not a particularly plausible route to cancelling any elections. But they can certainly make enough noise that a lot of people may become confused or scared to vote. So we need to remain laser focused on getting everybody to the polls. Like, this should be priority #1 for every citizen who wants to see democracy continue.
Civil unrest very likely in November.
Belarus has elections, technically. Are they legitimate and free from coercion? Not really…
I guess I would support the democratically elected sovereign over the private corporation.
They are now exercising that power in the interest of the people (they believe) that grant that power.
2. It is a principled take on that private companies shouldn't be making the decisions what their tools can and can't be used for in such an important sector.
You could even make a Third Amendment case if you stretched the logic far enough. Does "you can't be forced to quarter soldiers" extend to being forced to provide other forms of support?
California literally does this all the time.
Does what? Place companies on a list of businesses that no supplier to the state government of California is allowed to do business with? I'm unfamiliar with such a list but I suppose anything is possible these days.
We are not talking about soldiers living in Anthropic's offices. We are talking about an office employee being able to generate a PowerPoint about autonomous weapon systems.
Dumber stretches of logic have certainly emerged from SCOTUS. If Wickard v. Filburn makes sense to them, so could this.
Maybe not from the present bench, but perhaps from a hypothetically less-partisan one.
I’m sure the lawyers just got paged, but does this mean the hyperscalers (AWS, GCP) can’t resell Claude anymore to US companies that aren’t doing business with the DoD? That’s rough.
Additionally, every major university will undoubtedly have to terminate the use of Claude. First on the list will be universities that run labs under DOD contracts (e.g. MIT, Princeton, JHU), DOE contracts (Stanford, University of California, UChicago, Texas A&M, etc...), NSF facilities (UIUC, Arizona, CMU/Pitt, Purdue), NASA (Caltech).
Following that it will be just those who accept DOD/DOE/NSF grants.
We do have a recent example with Huawei, and it did fall just like this - and that was just some hardware.
In the recent Supreme Court hearing over the firing of Lisa Cook from the Federal Reserve, the administration is acting like Truth Social posts are official notices.
>Several justices have noted the unusual nature of the case before it, which began with a post by Trump on his social media platform, Truth Social, that said he would fire Cook.
>Jackson wondered why that would be considered sufficient notice: “How is it that we can assume that she’s on social media?”
https://apnews.com/live/supreme-court-lisa-cook-federal-rese...
Generally any machine that touches Supply chain Risk software cannot ship any software to DoD. AWS has separate clouds but software comes from same place.
You got it backwards, can't use claude if you ARE doing business with DoD.
Presumably AWS/GCP don't care, its up to the end customer to comply. Not like GCP KYC asks if you work with DoD.
This is almost certainly going to be rolled back, because I guarantee the DoD isn’t going to stop doing business with the hyper scalers, and the hyper scalers aren’t going to stop doing business with Anthropic.
Oh you tender babes, trying to logic the meaning of what the lieutenant of the biggest crime syndicate in the world means with his words, as if this was a well thought-out strategy... it's a shakedown; it would make more sense to ask "at least if Hegseth is sober..."
It's a bit like how the US Cuba sanctions worked and why they effectively isolated Cuba from everything.
I still don't see any way to read that as saying they could only do business with US customers, whether they give in or not?
The courts have historically been pretty consistent about giving the DoD whatever the fuck they want, going back to WW2 and even longer for the predecessors of the DoD. I agree that the next administration might reverse it, but the thing is, the government will stay irrational longer than Anthropic will remain solvent.
The US government told every American company to stop doing business with Huawei and they all did it overnight, even when it cost them billions. TSMC stopped fabricating for them, Google pulled Android licensing… The machinery of sanctions compliance is extremely well-oiled and companies fold instantly because the outcome of noncompliance is literally getting thrown in prison.
Even if a ton of companies have to switch over to an alternative, it won’t be catastrophic to the economy.
Who's next? OpenAI? Google? What if they refuse to allow the DoD to use AI with zero safeguards and Trump's goons decide they are also a "supply chain risk"?
Look no further than the famous expose by Mark Klein, the former AT&T technician and whistleblower who exposed the NSA's mass surveillance program in 2006, revealing the existence of "Room 641A" in San Francisco. He discovered that AT&T was using a "splitter" to copy and divert internet traffic to the NSA, proving the government was monitoring massive amounts of domestic communication.
Words cannot describe how crazy things were at that time.
I feel like someone will make a movie about it someday.
They are only refusing two narrow, but important categories. Framing this as blanket "refusal to support the DoD" feels like an angry, reactive own goal rather than a careful reading of what they actually said.
So far the march toward dictatorship keep being detoured by sheer incompetence. In any case, is hard to seize power when you can’t organize a group chat...
Which exactly is best changes on almost a weekly basis as different companies tweak their best model. I doubt the military would want to be switching supplier every week.
https://www.mintpressnews.com/pentagon-recruiting-elon-musk-...
https://thehill.com/policy/technology/5758898-altman-backs-a...
Anyone can use Claude afaik?
Now, my guess is in the ensuing lawsuit Antropic's defense will be that that is just not a product they offer, somewhat akin to ordering Ford to build a tank variant of the F150.
On the non-nuclear battlefield, I expect that the goverment wants Claude to green-light attacks on targets that may actually be non-combatants. Such targets might be military but with a risk of being civilian, or they could be civilians that the government wants to target but can't legally attack.
Humans in the loop would get court-martialed or accused of war crimes for making such targeting calls. But by delegating to AI, the government gets to achieve their policy goals while avoiding having any humans be held accountable for them.
They already have that. By definition. If Anthropic has done the work to be able to run on classified networks, then it's already running air-gapped and is not under Anthropic's control.
The thing is, just because you're in a SCIF doesn't (1) mean you can just break laws and (2) Anthropic don't have to support "off-label" applications.
So this is not about what they have and what it can do today - it's about strong-arming anthropic into supporting a bunch of new applications Anthropic don't want to support (and in turn, which Anthropic or it's engineers could then be held legally liable for when a problem happens).
It worked for Porsche ¯\_(ツ)_/¯
https://www.astralcodexten.com/p/the-pentagon-threatens-anth...
Discussed here:
Edit: oops, I misunderstood. This seems to be more about contractual restrictions.
This restriction is apparently "radically woke"
This administration built almost entirely of dunces and conmen has convinced itself/been convinced that chatbots will help them in deciding where to send nukes, and/or they are invested in the incredibly over-leveraged companies engaged in the AI-boom and stand to profit directly by siphoning taxpayer dollars to said companies. My money is on the latter more than the former, but they're also incredibly stupid, so who's to say, maybe they actually think Claude can give strategic points.
The Republicans have abandoned any pretense of actual governance in favor of pulling the copper out of the White House walls to sell as they will have an extremely hard time winning any election ever again since after decades of crowing about the cabal of pedophiles that run the world, we now know not only how true that actually is, but that the vast majority are Conservatives and their billionaire buddies, and the entire foundation and financial backing of what's now called the alt-Right, with some liberals in there for flavor too of course.
If this shit was going down in France, the entire capital would have been burned to the ground twice over by now.
Heard that one before. We'll get a reprieve of 4-8 years and the vote will go to the fascists again. Take that to the bank.
your view of France is severely outdated
Worse, they act like it's virtuous.
- $1,000,000 donation from NVIDIA CORPORATION to the Trump–Vance Inaugural Committee.
- $1,000,000-per-head Mar-a-Lago dinner where Nvidia CEO Jensen Huang attended.
- Jensen Huang’s contribution toward Trump’s "White House ballroom" project. Confirmed, but undisclosed value...lets says at least another $1,000,000?
Also I believe NVIDIA's supposed to pay the US government 15% of its revenues from Chinese sales:
https://www.ft.com/content/cd1a0729-a8ab-41e1-a4d2-8907f4c01...
Which is incredibility short term thinking. You're in strategic competition, and you compromise you position for a bit of cash?
It's also not solely about money, you can get far just knowing how to chum it up with Trump when you get in the room with him. Look at the odd quasi-bromance between him and Mamdani who you'd expect to be enemy #1 but Mamdani knows how to schmooze the exact type of New York Guy Trump is.
It's also potentially an implementation of the foot-in-the-door technique (https://www.simplypsychology.org/compliance.html). It's a common manipulative strategy where you get someone to do a small favor for you which makes them much more likely to do a large favor for you later.
"Trump’s Profiteering Hits $4 Billion" - https://www.newyorker.com/news/a-reporter-at-large/trumps-pr...
"How much money President Trump and his family have made" - https://www.npr.org/2026/01/14/nx-s1-5677024/trump-profits-m...
https://polymarket.com/event/which-party-will-win-the-house-...
Looking forward to a military platoon defying orders and seizing the president, hey, all countries suffer through coups, about time this young democracy go through one!
Did you skip class they day that discussed the Civil War?
"The practical outcomes of a system over the long-term reveal something important of the the true-preferences of the various interests which control that system, and these interests may be very different from the system's stated goals."
I don't see the contradiction here. The purpose of a cancer hospital is to cure as many patients as possible. "What it does" is cure as many patients as possible. The fact that as many patients as possible is currently (presumably) two-thirds is irrelevant. If major advancements in medicine or new types of cancer emerged which changed the percentage of people cured it wouldn't matter at all. "What it does" and "the purpose of the system" is still unchanged.
(Although there is another message, there, too: “the purpose of a system, insofar as it can be said to exist separate from what it actually does, has no weight in justifying the system’s existence or design”.)
You can make a lot of claims and they can match to reality a lot - normally people think of evaluating things in terms of a strict "does this fit or does this not", but it's often the meta-style (why do you keep bringing up that argument in that context?) that's important, even if it's not "logically bulletproof".
The follow-up is slightly better. But still not very convincing, IMO. They get far too stuck on a literal interpretation. Of something that self-describes as a heuristic.
The phrase does not make more sense even if we go all the way back to Beers. I certainly don't feel alone in not understanding how he went from his (fair) observation that "[There's] no point in claiming that the purpose of a system is to do what it constantly fails to do" to his more controversial conclusion: "The purpose of a system is what it does (aka POSIWID)".
Surely, there were many more sensible (but perhaps less quippy) stops between the two.
"No Tax or Duty shall be laid on Articles exported from any State.".
Meanwhile, irrelevant "AI Czar" David Sacks, member of the PayPal mafia alongside known Epstein affiliates Elon Musk and Peter Thiel, is furiously retweeting all the posts from Trump, Hegseth, and other accounts. He is such a coward and anti American:
Aren't all the AI companies saying that AI poses even a greater threat to humanity than nukes?
How can these two not be deeply connected? If a technology poses humanity extinction level of risk of course it will also be a matter of national security - how can it not be?
20-30 years ago eco-terrorists bombed and burned down a number of biological research laboratories and other targets, because of the perceived risks of gene technology.
https://en.wikipedia.org/wiki/Earth_Liberation_Front#Notable...
Given all the current talk (and the famous scifi movies) about the risks of AI, I am a bit puzzled how there are no similar activists groups trying to sabotage AI facilities.
What is it that made the risk from gene manipulation feel so much more real and leading to actions, than the current AI risk? The Terminator movie franchise is more famous than any scifi movies about gene technology. (Edit: I guess Jurassic Park franchise surpasses The Terminator.)
I am not. Anyone who understands the various downside risks and has a basic grasp of how the technology works also understands that compute is fungible and that there's no way to point at a given data center and be sure about whether it's providing search functionality, hosting cat pictures, enabling online shopping, training AI, or keeping planes from falling out of the sky. Even if you receive guidance in a vision that a given data center is bad, how do you deal with the reality of load balancing and the knowledge that the evil computation you hate won't be just hosted on a different server instance?
The Terminator movie franchise
I agree with you in that people probably do understand the existential risks of AI run riot better than many other possibilities due to those movies. But the problem is that the movies all depend on time travel. The unwilling human protagonists are persuaded to undertake drastic life altering criminal action based on information from The Future: both absolutely compelling demonstrations of technology from The Future (to justify the moral decision) and highly specific historical analysis from The Future (providing the operational gameplan).
I don't recall the specific plot crises of every movie, but all of them have well-defined success conditions, such as: ensuring the Terminator is destroyed and Sarah Connor survives; ensuring Cyberdyne Systems and the Terminators are destroyed and John Connor survives; ensuring the bad Terminator is destroyed before it can push the Skynet OS to production on every consumer computer device etc. For every dystopia-advancing use of time travel, there's a good use of time-travel helpfully pinpointing exactly where everything went wrong and what to do about it.
But back in the real world, even if you have absolute moral clarity that the creation of Skynet/the Torment Nexus/the Basilisk is imminent and must be stopped, how exactly do you go about this? I can think of a few people who have tried to attack data centers (for political/ideological reasons) and not only did they end up in federal prison, they also had no operational impact whatsoever. Realistically, we maintain a social status quo despite approximately quarterly assassinations, massacres of schooldren, or similar atrocities; why would any rational actor expect to alter the course of history by targeting a faceless abstraction? Even if the top ten tech CEOs were all simultaneously assassinated tomorrow, would things be substantively different a month later? Once the public freakout subsided, the companies would get new CEOs with much more proactive security details, a bunch of restrictive new laws would be promulgated, and everything would carryon more or less as before.
I dunno, safeguard seems like a weasel word here. It’s just reserving control to one party over another. It’s understandable why the DoD(W) wouldn’t like that.
Here's the term defined in an official context:
https://www.acquisition.gov/dfars/252.239-7018-supply-chain-....
[0]: https://www.acquisition.gov/dfars/252.239-7018-supply-chain-....
[EDIT] Oh man, yours is like that too? WTF.
[EDIT2] If I follow your link, hit the 404 page, then add a period at the end of the URL, it does load. God that's strange.
That gave me a good, actual LOL, thanks for that one.
HN separates trailing dots from URLs, so that you can have working URLs at the end of a sentence. Hence you have to percent-encode trailing dots if they are a necessary part of the actual URL. (Same for some other punctuation characters, probably.)
This behavior is common for auto-hyperlinking of URLs in running text, so it’s bad practice to have such URLs.
https://www.acquisition.gov/dfars/252.239-7018-supply-chain-...
I really wonder what Palantir's role in all this is because domestic surveillance sounds exactly like Palantir and whatever happened during the Maduro raid led to Anthropic asking Palantir questions which the news reports is the snowball that escalated to this.
Recent news reports from February 2026 indicate that a significant rift developed between Anthropic and the Department of War (Pentagon) following the capture of Venezuelan President Nicolás Maduro in January 2026.
According to a report by the Wall Street Journal (referenced by TRT World and others on February 14–15, 2026), the controversy originated when an Anthropic employee contacted a counterpart at Palantir Technologies to inquire about how Claude had been used during the raid. Key Details of the Reports:
* Discovery of Use: Anthropic reportedly became aware that its AI model, Claude, was used in the classified military operation through its existing partnership with Palantir. This was allegedly the first time an Anthropic model was confirmed to be involved in a high-profile, classified kinetic operation.
* The Inquest: The Wall Street Journal and Semafor reported that an Anthropic staff member reached out to Palantir to ask for specifics on Claude's role. This inquiry reportedly "triggered the current crisis" because it signaled to the Pentagon that Anthropic was attempting to monitor or place "ad hoc" limits on how its technology was being used in active missions.
* The Confrontation: During a recent meeting between Anthropic CEO Dario Amodei and Defense Secretary Pete Hegseth, the inquiry to Palantir was a point of contention. Hegseth reportedly claimed Anthropic had raised concerns directly to Palantir about the Caracas raid. Amodei has since denied that the company raised objections to specific operations, characterizing the exchange with Palantir as a routine technical follow-up or a "self-serving characterization" by Palantir.
* Current Status: This friction has escalated into a public showdown. Today, Friday, February 27, 2026, reports indicate that the Trump administration has officially designated Anthropic a "supply chain risk" and ordered federal agencies to cease using Claude after the company refused to remove guardrails related to autonomous weaponry and mass domestic surveillance.
The primary reporting you are likely recalling comes from The Wall Street Journal (approx. February 14, 2026) and was later expanded upon by Semafor regarding the specific communications between Anthropic and Palantir employees.
If we could give Taiwan killbots that would ensure China could never invade, or at least could never occupy Taiwan, would that be good or bad? I have a feeling I know what the Taiwanese would say.
While we're at it, should we also strip out all the machine learning/AI driven targeting systems from weapons? We might feel good about it, but I would bet my life savings that our future adversaries will not do the same.
The world is more nuanced than that.
But to answer your question. No we should not give anyone automatic kill bots. Automatic kill bots shouldn’t even be a thing.
Whether you or I like it or not, automatic kill bots will be a thing. It will only be a question of which countries have them and which do not.
Generally, in war, there are no rules, and someone is going to make automated killbots, and I expect one place to see them quite soon is in the Russia-Ukraine war. And yes, I'm hoping the good guys use them and win over the bad guys. And yes, there are good guys and bad guys in that conflict.
No, thanks, we don't need those "fully automated kill bots". There's absolutely no guarantee that they wouldn't kill the operator (I mean, the one who directs them) or human ally.
We're pretty much fine with drone technology we have.
But for me personally, that's not the most important point. What is more important - and what almost no one in the Western countries seems to realise (no offence, but many of westerners seem to be kind of binary-minded: it's either 0xFFFFFF or 0x000000, no middle ground at all) - is that on the Russian side, soldiers are not "fully automated kill bots" either. Sure, there's a lot of... let's say - war criminals. Yes, for sure. But en masse they are the same young men that you can see on the Ukrainian side. Moreover, many people in Ukraine have relatives in Russia, and there already were the cases where two siblings were in different armies, literally fighting with each other. So in my opinion, "fully automated kill bots" are not an option here. At least unless you deploy them in Moscow and St. Peterburg to neutralize all of the Russian elites, military commandment and other decision-making persons of the current regime.
Personally, I think it'd be great to have the Anthropic people at the table in the creation of such horrors, if only to help curb the excesses and incompetencies of other potential offerings.
They still pay taxes, which fund the US government, which kills innocent human beings around the world...
How is it that high!?
That means that more than 1-in-3 of your countrymen are ride-or-die, and it's just heartbreaking to see that we're going to have to launch that many people into the sun.
The only other thing that the foreign AI companies could do is say no to automated killing bots, which doesn't even seem like that good of an idea considering that your countrymen will most likely have to interact with these robots that can kill without any oversight.
... in the same sense as the two sides of a coin are separate sides maybe.
I guess if we are going this way, we might already start building camps for the undesirables, since our adversaries will surely do that too.
https://www.realclearpolling.com/polls/approval/trump-obama-...
China's AI Safety Governance Framework: https://www.cac.gov.cn/2025-09/15/c_1759653448369123.htm
Most Americans hate AI and it's effectively the ostrich effect where they hope to outright ban it and ignore everything else. Meanwhile, all the evil people are running the show. While Anthropic continues to propagate Sinophobic messaging, DeepSeek and other companies have a much more muted tone.
When we have first politician blown to bits by autonomous AI FPV there will be sheer panic of every politician in the world to put the genie back into the bottle. It will be too late at that point.
Anthropic is correct with its no killbot rule.
Even during the Nagorno-Karabakh war, Azeri loitering munitions were able to suppress Armenian air defenses by hitting them when they rolled out of of concealment. I believe that killchain requires a level of autonomous functionality.
However Russians and Ukrainians are using AI recognition in recon drones, but not yet in FPV. There is strong suspicion that long range one way attack drones are using AI during terminal guidance, but I did not see it confirmed by either side.
Open-source everything. Papers, code, weights, financial records. Do all of your research in the open. Run a 100% transparent organization so that there's nothing to take from you. Level the playing field for good and bad actors alike, otherwise the bad actors will get their hands on it while everyone else is left behind.
Stop comparing AI capabilities to nuclear weapons. A nuke cannot protect against or reverse the damage of another nuke. AI capabilities are not like nukes. Diffuse it as much as possible. Give it to everyone and the good will prevail.
Build a world where millions of AGIs run on millions of gaming PCs, aligned with millions of different individuals. It is a necessary condition for humanity's survival.
Supply-chain risks means "the potential for adversaries to sabotage, subvert, or disrupt the integrity and delivery of defense systems, including software, hardware, and services, to degrade national security".
So now Anthropic is an adversary, because it does not want "fully autonomous weapons" or automated mass surveillance? Sure thing, DoD. Go use Grok or whatever, I'm sure that will go great.
So OpenAI will also be marked as a supply chain risk too, right?
[1]: https://www.axios.com/2026/02/27/altman-openai-anthropic-pen...
Dario is Lando, complaining “We had a deal!” Only to be told, “I’m altering the deal. Pray I don’t alter it any further.”
Statement from Dario Amodei on our discussions with the Department of War - https://news.ycombinator.com/item?id=47173121 - Feb 2026 (1508 comments)
I wish I thought enough Americans had the spine required to stand up to this, and I know for a fact that a lot do... the solution is literally written into your constitution.
Hacking is using a system in a way it was not intended to be used.
Here it is that, but applied to the law.
Hegseth and friends are a bunch of black hat legal hackers.
This administration consistently exploits what were designed to be emergency powers because no such requirement exists. Leave no room for interpretation.
Q: "Is there anything we could do to change your mind?"
A: "Yes! Stand up to the current administration."
Does this mean Azure & AWS will have to stop offering Claude as a model?
AWS Bedrock has deployed Anthropic models under an interesting structure. It is fully hands off - the models are copied into the AWS infrastructure and don't use anything from Anthropic. I think if push came to shove, Anthropic could cut ties with Amazon and AWS could probably still keep serving the models it has with Anthropic forgoing revenue until this is resolved, while asserting they are not "conducting commercial activity" between each other.
All speculation of course.
Edit: I should perhaps clarify I'm more interested in paid users, rather than free. It's harder to tell if free users switching would help them or hurt them... curious if anyone has thoughts on that too.
i told myself if anthropic does not back down on their current stipulations to the DoD, then i’d cancel and switch over to claude
they said there is a line they do not want to cross, and stuck to that stance, at great personal and financial risk to themselves
Come to EU guys, we'll prepare a warm welcome!
TIL Fully automated killbots and mass domestic surveillance are American principles.
I mean, I should have known but there's no clearer sign saying "leave the country now if you don't agree with this admin" than now I guess.
This is why 996 bosses think AI can replace their employees, because they already see the employees as robots, not humans.
[1] https://www.businessinsider.com/996-work-culture-silicon-val...
[0]https://www.ap.org/news-highlights/spotlights/2025/unquestio...
[1]https://www.nytimes.com/2026/02/23/us/politics/judges-contem...
That's not what the government is doing here.
Kesha tried to hug Jerry Seinfeld vibes.
> Anthropic delivered a master class in arrogance and betrayal as well as a textbook case of how not to do business with the United States Government or the Pentagon.
Strange way of saying "this vendor doesn't meet our software requirements".
> they have attempted to strong-arm the United States military into submission
Err... You approached them?
> a cowardly act of corporate virtue-signaling that places Silicon Valley ideology above American lives.
It's an orthogonal point, but "Silicon Valley ideology" has made up a significant portion of the USA's GDP for the last however many years.
> Their true objective is unmistakable: to seize veto power over the operational decisions of the United States military. That is unacceptable.
Again... You approached them?
> I am directing the Department of War to designate Anthropic a Supply-Chain Risk to National Security.
Like most companies in the world I imagine. They just haven't been approached yet.
> to allow for a seamless transition to a better and more patriotic service.
Internally re-framing all the recent "EU moving away from American tech!" articles as "EU builds more patriotic services!"
> This decision is final.
Nothing says "final" like a Tweet. The most uncontroversial and binding mechanism of all communication.
This word effectively means NOTHING, anymore.
Doublespeak this motherfucking wrongthink.
Anthropic folks: I've been a bit salty on HN about bugs in Claude Code, but I feeling pretty warm and fuzzy about sending you my cash this month.
I don't think it will hold, in the end this is mafia behavior, but if it does, we are yet again in uncharted waters.
Oddly, though, it seems like that should solve this problem as well. I'm not sure why the Department of Defense insists on Anthropic's models in particular; one would think one of the other players, at the very least least xAI, would be willing to step in and provide the capability Anthropic doesn't want to provide.
It is an interesting point. What's the difference between this use license and others?
That is what they are doing.
> why would you then [....] ban other companies from making business with Anthropic if they want to be a defense contractor
Because, if it shops with Anthropic code, the DoD becomes subject to the restrictions when they receive the contractor's product. Anthropic's limitation is on the use, not (just) on the product or distribution.
To stop using them requires making the suppliers still using them as well.
Designating them a supply chain risk is unprecedented authoritarian strong-arming.
It almost is parody that a former Fox News host is the SECRETARY OF WAR.
And people here are debating legalese...
But anyway, I guess the question is, will any other big AI companies stand with them? It's what needs to happen, but I am not hopeful.
Government: We will destroy any company that refuses to create the Torment Nexus
Makes for very confusing reading when comments from "1 hour ago" are actually on preceding events from earlier, before TFA news (announcement of designation).
mods: Especially in sensitive and rapidly developing situations like this, please don't mess with timestamps of comments. It's effectively revisionism.
Don't get me wrong i'm glad they are unwilling to do certain things...
but to me it also seems a little ironic that Anthropic literally is partnered with Palantir which already mass surveills the US. Claude was used in the operation in Venezuala.
Their line not to cross seems absurdly thin?
Or there is something mega scary thats already much worse they were asked to do which we dont know about I guess.
To me all of this reads like "we don't trust our models enough yet to not cause domestic havoc, all other is fine, and we don't trust our models enough yet to not vibe-kill people". Key word being "yet".
Their hard lines are:
- no usage of AI to commit murder WITHOUT a human in the loop
- no usage of AI for domestic mass surveillance
Claude: "Are you sure you want me to commit murder?"
User: "Yes"
Or do you mean Human presses button:
Claude: "Do you to commit murder? If so press the button."
User: "I pressed the button"
Claude: "Great! Now lets summarize what we did."
There are many ways to construct HITL UXes. But typically they'd take the form of the first one
I think you're missing the forest for the trees. All Anthropic is saying is that HITL is required before murder, the UX is irrelevant
How going against the most powerful army on Earth is coward?
Because that could be absolutely staggering.
It's more been about the size of the government; that it should do a minimal amount of control (and do it well), but leave a lot of things for "the market to decide".
Having said all that, I think this issue is just tangential to any big/small government ideology. This is a hissy fit about a defence contractor sticking to their agreement where the DoD want to change the agreement in a way that goes against the contractors Mission Statement and/or the US Constitution itself.
The old ideology of the Republicans doesn't mean anything here. This administration is purely about 'give me what I want, now!'.
And it's whims change with the breeze. Do not look for consistency here.
It's magnified because it's right now, but this won't affect midterm results barely a whisker compared to many other daily headlines.
There are no serious enemies to this administration in SV and I can't see this changing that. SV has bent the knee exactly like Anthropic didn't. They're not going to stand up because of this, they've proven they don't have those muscles.
So using Claude Code to write software for the DoD is now a no go, you'd be in breach of procurement directives now.
If they go as far as to convince congress to add Anthropic to the NDAA, that would be a nationwide ban like Huawei making it illegal for any federal contractor to use the tech anywhere in their business.
But for now, even fed contractors can still use Claude in their business, just not directly for government work.
> Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic
Emphasis mine.
And I’m looking at news organizations that presumably have staffs of legal analysts pouring over this stuff, and they also seem to be saying that it can’t be any commercial activity:
> The label means that no contractor or supplier that works with the military can do business with Anthropic.
https://www.nytimes.com/2026/02/27/us/politics/anthropic-mil...
> Secretary Hegseth has implied this designation would restrict anyone who does business with the military from doing business with Anthropic. The Secretary does not have the statutory authority to back up this statement. Legally, a supply chain risk designation under 10 USC 3252 can only extend to the use of Claude as part of Department of War contracts—it cannot affect how contractors use Claude to serve other customers.
https://www.anthropic.com/news/statement-comments-secretary-...
Looks like the NYT might have gotten it wrong…
He called me and he seemed like a nice enough guy, but I realized that he's one of the DOGE/Elon acolytes and he started talking about how he's "fixing" the Treasury and that every engineer is apparently supposed to use Claude for everything.
It would have been a considerable pay downgrade which wouldn't necessarily be a dealbreaker but being managed by DOGE would be, but mostly relevant is that I found it kind of horrifying that we're basically trusting the entire world's bank to be "fixed" with Claude Code. It's one thing when your ad platform or something is broken, but if Claude fucks something up in the Treasury that could literally start a war. We're going to "fix" all the code with a bunch of mediocre code that literally no one on earth actually understands and that realistically no one is auditing [1].
If they're going to "fix" all the Treasury code with stuff generated by Claude, I'm not sure they will have a choice but to stick with it, because very it seems very likely to me that it will be incomprehensible to anything but Claude.
[1] Be honest, a lot of AI generated code is not actually being reviewed by humans; I suspect that a lot of the AI code that's being merged is still basically being rubber-stamped.
it won't be the world's bank for very long
If you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
I have just purchased a chunk of extra usage credit. I encourage my peers to do the same. Let's send a message to those that work forces.
If an employee of the government makes a decision that subsequently turns out to be very very unpopular, that unpopularity is sooner or later going to coalesce and land on them, and the more unpopular it turns out to be the less of a shield legal arguments about immunity or pardons will be because so many people are increasingly out of patience with a system they deem to be corrupt. Being able to offload the political, legal, and personal risks of extremely consequential decisions onto The Bad Computer System is the political equivalent of crack cocaine - you might know that the feeling of freedom and power it provides is wholly illusory, you might know that it's likely to ruin your own and many other lives, you might know that it's a disaster for the health of the body politic...but it also offers the possibility that you can have an absolute blast and get away with it.
My anecdotal experience of being around wealthy and powerful people over the years inclines me to think that not only do our social systems select in favor of people who take big risks for big rewards, but that virtually everyone in that class has a) done a lot of getting away with things legally speaking and b) enjoys using illegal drugs. Even if they've given up recreational drug taking or limit it to strictly defined times and places so as not to interfere with their business/personal success, they like thrills and have confidence about their ability to enjoy them without negative consequences. You need some of that risk-taking, high personal autonomy attitude if you aspire to be a mover and shaker as opposed to a leading figure in risk management or regulatory compliance.
Everyone enjoys the feeling of power without responsibility; it's a fundamental underpinning of games and many other kinds of recreation. Add in significant amounts of money and people think differently about risk, as in the topical case of the experienced Supreme Court litigator who turned out to have have a secret life as a high-stakes poker gambler and eventually started betting against the IRS while filing his taxes (https://www.politico.com/news/2026/02/25/supreme-court-litig...).
Now, if you're in the political-military sphere and you get your thrills by literally redrawing lines and relationships on the map of the world and deciding what the news on TV is going to be for the next day/week/month/year, and you get offered a tool that promises to give a significant edge over other players in this game but which also gives you a versatile and widely accepted excuse for avoiding consequences for the inevitable losing hands, there are massively compelling psychological incentives for using it. And correspondingly, there's going to be massive emotional disruption (and bad decision-making and behavior) if your supply is threatened. You might start labeling the people who are interfering with your good time as cognito-terrorists and telling all your friends and supporters that your formerly trustworthy supplier did you dirty...
> Anthropic said no, and now the admin is trying to destroy the company in retaliation.
From https://bsky.app/profile/bbkogan.bsky.social/post/3mfuuprph5...
LLMs produce output of unknowable and unpredictable accuracy, and as far as we know, this is a mathematically unsolvable problem. This shit should not be within 1000 miles of a weapons system. Why are we even talking about this?
Anthropic could have said "you can use our technology for anything but faster-than-light travel." The military administration would have said "you're not the boss of me," and the outcome would have been exactly the same.
It's a hot-button issue, just like flag burning. Nobody ever really cared about flag burning.
By the way, your "No, stop" was rude and unnecessary, and your comment would have been stronger without it.
Joking aside, this administration clearly cares much less others. They don't care if innocent people are killed.
So do humans. But humans might not follow illegal or immoral orders.
A few arrests and a few in detention centres, will be enough to make them fold and grovel.
They are now categorised as "radical left" and woke.
The elections will be controlled to "prevent the radical left take over of the greatest country on the planet".
edit : The stage is also being set for total media control. My prediction is that the next target will be Google, specifically Youtube. You should start seeing talks about how the radical left is inflitrated youtube.
> 1. No mass domestic surveillance of Americans
> 2. No fully autonomous weapons (kill decisions without a human in the loop)
Surveillance takes place with or without Anthropic, so depriving DoW of Anthropic models doesn't accomplish much (although it does annoy Hegseth).
The models currently used in kill decisions are probably primitive image recognition (using neural nets). Consider a drone circling an area distinguishing civilians from soldiers (by looking for presence of rifles/rpgs).
New AI models can improve identification, thus reducing false positives and increasing the number of actual adversaries targeted. Even though it sounds bad, it could have good outcomes.
Anthropic are taking a moral position which is admirable, but in this case it could actually make people's lives worse (if we assume more false positives and fewer true positives, which is probably a fair assumption given how much better 'modern' AI is compared to the neural net image recognition of just a few years ago).
But there's some irony in this happening to Anthropic after all the constant hawkish fearmongering about the evil Chinese (and open source AI sentiment too).
Sometimes it pays to think even two steps ahead of your most immediate thought…
Nevermind Claude, does that mean Anthropic's offices can't use a power company if that same company happens to supply electricity to a US military base? What about the water, garbage disposal, janitorial services? Fedex? Credit card payments? Insurance companies? Law firms? All the normal boring stuff Anthropic needs that any other business needs.
This is a corporate death penalty. Or corporate internal exile or something, I don't know of a good analogy.
This will mean Grok becomes the defacto US Gov AI provider.
Absolutely insane.
If we don't impeach for this, we might as well surrender to MAGA.
I don't think that Secretary Hegseth is qualified to speak on American principles.
Cheating on multiple spouses[1], being an active alcoholic, and being accused of multiple sexual assaults and paying off the accusers[3] is fundamentally incompatible with being a Secretary of Defense and a good leader.
Also, this violates freedom of speech and will probably get shot down in the courts.
1. https://en.wikipedia.org/wiki/Pete_Hegseth#Marriages
2. https://en.wikipedia.org/wiki/Pete_Hegseth plus multiple recent media pieces
3. https://en.wikipedia.org/wiki/Pete_Hegseth#Abuse_and_sexual_...
https://www.trumpstruth.org/statuses/36981
Don't worry, this is an archive/mirroring site for his account, not the actual TS site.
I'd comment on how wackadoo this all is, but, 1) that applies to almost everything these days, and 2) the post's right there, see for yourself.
With that said: what are the chances, in your opinion, that Donald wrote that himself?
To me it reads too coherent for there to be any chance he wrote or even dictated that.
I do think a lot of the more hot-take type posts (often in response to stuff he’s watching on tv) are either actually him, or he’s dictating to an aide. These larger policy-type ones that he treats as quasi-executive-orders, I think are likely drafted by one or more of his cabinet-level folks, or others roughly as high up. That’s just my speculation based on reading the “tea leaves”, though.
As for official word, it waffles between “all of it’s him” and “oh not that one though, that racist video repost was a staffer who made a mistake”, so that’s little help in sussing out the truth (but I am rather certain they’re not all directly written and posted by him)
Every conservative needs to do some very deep, very serious soul-searching. As for me, as a hyper-progressive, I'm drawing up proposals for nationalizing real estate developers in order to force them to build new houses to sell below cost.
Model collapse making models identify everyone as a potential threat who needs to be eliminated.
Companies should have a right to refuse such requests on moral grounds though.
This stance is vindictive. Just don't use Claude in the military. Extending it to all government agencies is not right. They do great work. Can't deny that.
And here’s the irony: Musk, who claimed only he is virtuous enough to defend us from AI, who insisted he always wanted model labs to be non profit and research focused, will now bring his for profit commercial entity into service to aid in mass domestic censorship and fully autonomous weapons of war.
In fact it won’t surprise me further if NVIDIA is strong armed into providing preference to xAI, in the interest of security, or if the government directly funds capital investments.
Anthropic saves some dignify and they’re the losers today, but we are the losers tomorrow.
One, it’s going to fuck with the AI fundraising market. That includes for IPO. If Trump can do this to Anthropic, a Dem President will do it to xAI. We have no idea where the contagion stops.
Two, Anthropic will win in the long run. In corporate America. Overseas. And with consumers. And, I suspect, with investors.
A lot of corporate America contracts for the military in some capacity (it's a giant piggy bank and if you jump through a few hoops you get to siphon money out of it, so of course they do) and assuming this Tweet is accurate (Jesus, what a world) this will also affect them.
IDK maybe they have corporate structures that avoid letting this kind of thing mess too badly with the parts of their company that don't have contact with the government, or maybe it'll only apply to specifically the work they do for the government, but otherwise I expect it'll be devastating for Anthropic's B2B effort.
And a lot does not, or does so through dedicated subsidiaries so they can work multinationally.
Will the next Democratic President do it to xAI? On what grounds?
The Biden admin negotiated a contract with a supplier with terms which are – to the best of my knowledge – rather unprecedented – do Pentagon contracts normally have terms like this, restricting the government's use of the supplied good or service? Do missile or plane contracts with Boeing or Lockheed Martin contain restrictions on what kind of operations that hardware will be used in? I don't think that's the norm. So the next administration tears up a contract made by the previous admin with unusual terms – nothing unexpected about that. The "hardball" of declaring them a "supply chain risk" is escalating this dispute to a never-before-seen level, but the underlying action of cancelling the contract isn't. I honestly suspect the "supply chain risk" aspect will be suspended by the courts, and/or heavily watered down in the implementation; but the act of cancelling the contract in itself seems legally airtight.
Next Democratic administration inherits a contract with xAI (and quite possibly OpenAI and/or Google too) – with presumably standard terms. I can totally understand the political desire for vengeance. But what's the actual legal justification for it? Facially, the current administration has a politically neutral justification for what they are doing, even if some suspect there is some deeper political motivation. Will the next Democratic administration have such a facial justification for doing the same to xAI?
Plus, Democrats always sell themselves on "we obey norms". They have the structural disadvantage that either they keep their word on that, and can't do the same things back, or they break their word, and risk losing the people who supported them based on that word.
Elon being affiliated with Trump. About the strength of logic that makes Dario woke.
> don't think that's the norm
Norms are different from law or contract. And yes, lots of service providers limit where their civilians can be deployed and under what circumstances.
> can totally understand the political desire for vengeance. But what's the actual legal justification for it?
President has core Constitutional control of the military.
> Democrats always sell themselves on "we obey norms"
That hasn't worked. The American electorate is looking for change. And up-and-coming Democrats are picking up on that.
> risk losing the people who supported them based on that word
The Democrat base absolutely wants vengeance. It doesn't play in swing states. But it probably also doesn't hurt. These are court politics, at the end of the day.
I think you have to distinguish between the official justification and some of the associated political rhetoric.
Official justification: "Previous admin agreed contract with unprecedented terms, we demand those terms be removed, vendor is refusing to renegotiate"
Political rhetoric: "THE UNITED STATES OF AMERICA WILL NEVER ALLOW A RADICAL LEFT, WOKE COMPANY TO DICTATE HOW OUR GREAT MILITARY FIGHTS AND WINS WARS!"
If you forget about the political framing, and look at the official justification in the abstract, it doesn't actually seem facially unreasonable. The escalation to "supply chain risk" is a different story, but the core contract dispute and cancelling the contract as a result of it isn't.
So the question is, can Democrats come up with an equivalent abstract official justification–if so, what will it be? Or do they decide they don't even need that–in which case they aren't just matching Trump, they are going even further down the road to normlessness than he's gone.
> And yes, lots of service providers limit where their civilians can be deployed and under what circumstances.
There's a big difference between contracts for boots-on-the-ground and contracts for hardware/software. There is lots of precedent for contractual limitations on how boots-on-the-ground can be used. I'm not aware of similar precedent for hardware or software.
> That hasn't worked. The American electorate is looking for change. And up-and-coming Democrats are picking up on that.
Are they? Gavin Newsom? Zohran Mamdani? AOC? Do they actually sell themselves as "we see Trump breaking the rules, and we'll break them just as hard, even moreso"?
> The Democrat base absolutely wants vengeance. It doesn't play in swing states. But it probably also doesn't hurt.
It is too early to tell. You can argue in the abstract that X approximately equals Y, so if swing voters will tolerate the GOP doing X, they'll also tolerate Democrats doing Y – but the actual swing voters might not agree with you on that.
> You don’t anthropomorphize your lawnmower, the lawnmower just mows the lawn - you stick your hand in there and it’ll chop it off, the end.
Except this is like two lawnmowers going at it, which would be a sight to behold indeed.
As someone who wants America to win, ripping out Claude and putting in xAI is a terrible idea. Definitely setting us back a few months on capabilities
> Populist nationalism + “infallible” redemptive leader cult
> Scapegoated “enemies”; imprison/murder opposition/minority leaders
> Supremacy of military / paramilitarism; glorify violence as redemptive
> Obsession with national security / nation under attack
TBH could be worse.
And nobody in the administration is concerned at all that the model itself might be somewhat against their own views?
If it was so radically woke, wouldn’t the model, as used in fully autonomous weapons, be potentially harmful to ICE officers that the left considers as a threat to the American people?
Wouldn’t the mass surveillance of Americans be biased against the right?
These people are so dumb.
It's true that Chinese companies are extensions of the state. But they serve the state. And the state has thus far served the citizenry eg raising 800M people out of extreme poverty. China's HSR network of 32,000 miles of track was built in 20 years for ~$900B. That's less than the annual US military budget.
You can look at the relationship between the US government and US companies in one of two ways:
1. US companies serve the government but the government doesn't serve the people. After all, where's our infrastructure, healthcare, housing and education? or
2. The US government serves US corporate interests to enrich the ultra-wealthy.
Either way a handful of people are getting incredibly wealthy and all it takes is for a little corruption. Political donations, jobs after government, positions on boards and so on.
The place to set policies on the use of hammers and police enforcement is not at the counter of the hardware store. “You want a hammer but don’t have a contractors license? Are you in a training program? Oh you just want to hang framed art - can I see your lease, does it allow hammering metal into the walls?”
We govern these things through laws and a democratic process. Police enforce the laws.
I don’t want some overconfident Silicon Valley engineering firm telling me how to use my digital tools, and you shouldn’t either.
Whatever you think of this administration, our military should not have to ask contractors permission for their operations.
To stop mass surveillance and autonomous lethality, pass laws. Asking unelected tech executives to do this is asking for trouble. They have no business doing it.
Trump orders federal agencies to stop using Anthropic AI tech 'immediately'
https://news.ycombinator.com/item?id=47185528
Statement from Dario Amodei on our discussions with the Department of War
"THE UNITED STATES OF AMERICA WILL NEVER ALLOW A RADICAL LEFT, WOKE COMPANY TO DICTATE HOW OUR GREAT MILITARY FIGHTS AND WINS WARS! That decision belongs to YOUR COMMANDER-IN-CHIEF, and the tremendous leaders I appoint to run our Military.
The Leftwing nut jobs at Anthropic have made a DISASTROUS MISTAKE..." - President Donald J. Trump
“You won’t let us use your product unrestricted for military applications? Fuck you, we’re going to stop using it for anything at all across the entire federal government, even if not remotely related to military.”
....................../´¯/)
....................,/¯../
.................../..../
............./´¯/'...'/´¯¯`·¸
........../'/.../..../......./¨¯\
........('(...´...´.... ¯~/'...')
.........\.................'...../
..........''...\.......... _.·´
............\..............(
..............\.............\...
"Our AI is so advanced and dangerous Trump has to beg us to remove our safeguards, and we valiantly said no! Oh but we were already spying on people and letting them use our AIs in weapons as long as a human was there to tick a checkbox"
I just don't buy anything spewing out of the mouths of these sociopathic billionaires, and I trust the current ponzi schemers in the US gov't even less.
Especially given how much astroturfing Anthropic loves doing, and the countless comments in this thread saying things like "Way to go Amodei, I'm subbing to your 200 dollar a month plan now forever!!11".
One thing I know for sure is that these AI degenerates have made me a lot more paranoid of anything I read online.
Things happen.
Appealing to the pragmatic and the "game theory" of complying with authoritarian rule that you don't have power over - because the other party that you don't have any power over will benefit from it - is a zero-sum argument.
The "zero-sum" label is equally off-base. Zero-sum describes a situation where one party's gain is necessarily another's loss, and that is precisely the nature of military capability competition. If an adversary fields unrestricted AI systems and you field restricted ones, the gap is real and the consequences are asymmetric. You don't have to like that reality, but calling it a zero-sum argument as though it's a rhetorical trick misidentifies what's actually a structural condition. The term you seem to be reaching for is something closer to "fear-based reasoning" or "false dilemma," but neither of those applies cleanly here either, because the competitive dynamic being described is well-documented and not hypothetical.
If there's a genuine objection to be made, and there may well be, it has to engage with the specifics: whether the restrictions in question actually matter operationally, whether the transition plan is proportionate, whether the policy creates worse risks than it solves. That's where the real debate is.
[edit:typos]
What I don’t understand is why the two parties couldn’t reach agreement. Surely autonomous murderous robots is something U.S. government has interest in preventing.
The big difference here is that Claude is not military equipment. It's a public, general purpose model. The terms of use/service were part of the contract with the DoD. The DoD is trying to forcibly alter the deal, and Anthropic is 100% in the clear to say "no, a contract is a contract, suck it up buttercup."
We aren't talking about Lockheed here making an F-35 and then telling the DoD "oh, but you can't use our very obvious weapon to kill people."
> Surely autonomous murderous robots is something U.S. government has interest in preventing
After this fiasco, obviously not. It's quite clear the DoD most definitely wants autonomous murder robots, and also wants mass domestic surveillance.
i as a foreign citizen get to have hard to detect influence over the model because it scraped tons and tons of my internet comments.
if youre going to have a supply chain, it needs to include where the trainjng data is sourced from and who can contribute to it
This is just petty.
Consider the government. It’s Hegseth making this decision, and he considers the US military’s adherence to law to be a risk to his plans.
> What I don’t understand is why the two parties couldn’t reach agreement.
Someday we'll have to elect a POTUS who is known for his negotiation and dealmaking skills.
I wouldn't want a bullet manufacturer to hold back on my government based on their own internal sense of ethics (whether I agreed with it or not, it's not their place)
The fuck?
