The sold services to a willing counterparty at mutually agreed upon terms. And now the other side of that deal has recalled that they're Twelve and You're Not My Real Mom You Can't Tell Me What To Do, and so wishes they had agreed to different terms and is throwing a tantrum to attempt to force a change.
And that's Anthropic's fault? That's a risk they should have predicted?
Yeah, and the legal environment that contract was written in, which both parties were aware of during negotiation, defines the means by which those terms can be changed.
> And that's Anthropic's fault? That's a risk they should have predicted?
It is deeply funny to me to imagine that an AI company doing inference at an unprecedented scale could not see this coming.
Go ask Claude how usgov should act if a contractor preemptively refuses to deliver. What are the top five tools they could use to demand compliance?
See this is your confusion. They're not refusing to deliver, they're happy to provide the services agreed to at the rate negotiated. What they're refusing is a change in the terms.
If you contract me to build you a building, and I agree with a stipulation that it won't be used as a slaughterhouse (and that you'll write that into the deed), you can't compel me to continue building if you change your mind on that point six months in. You either break the contract subject to agreed terms, renegotiate to remove that clause, or stop breaching the contract.
Of all American claims to exceptionalism, one that rings closest to true is that the the people AND the government are all bound by the rule of law. Contracting with the government is no different than contracting with any other party.
Your point seems to be "but it's the government clearly they can do whatever they want lol" viz. DPA, Supply-Chain risk, etc. You're right that they have those powers. But accepting/asserting that the capricious, vengeful, use of those extraordinary powers should be an anticipated, normal feature of contracting with the government runs counter to what should be among our highest shared values. We might as well jump directly to the authoritarian logic 'they have the army, they can compel anything they want for $0, so just give them what they ask'.
Furthermore, one presumes Anthropic did see this coming, given no more evidence than that this is playing out in a giant public fracas making their values clear to all their possible customers the world over, instead of over a tense email thread between the assistant to the sub-under-deputy secretary for AI procurement and a half-dozen lawyers in SF.
(addenda: you're going to say we've used the DPA a bunch. I would argue that vanishingly few instances have compelled private enterprise to act in direct opposition to their own interests; an even in those cases they were just being asked to lose money (meatpackers, PG&E suppliers, ...))