From 00179b40209a551b44ec253a6c38d9e1702209ca Mon Sep 17 00:00:00 2001 From: "Zack M. Davis" Date: Fri, 31 Jul 2026 23:11:56 -0700 Subject: [PATCH] drafting summary judgement dispatch Really need to finish this tomorrow for timeliness reasons! --- ...artment-of-war-summary-judgment-hearing.md | 145 ++++++++++++++++++ 1 file changed, 145 insertions(+) create mode 100644 content/drafts/dispatch-from-anthropic-v-department-of-war-summary-judgment-hearing.md diff --git a/content/drafts/dispatch-from-anthropic-v-department-of-war-summary-judgment-hearing.md b/content/drafts/dispatch-from-anthropic-v-department-of-war-summary-judgment-hearing.md new file mode 100644 index 0000000..b4540d4 --- /dev/null +++ b/content/drafts/dispatch-from-anthropic-v-department-of-war-summary-judgment-hearing.md @@ -0,0 +1,145 @@ +Title: Dispatch from Anthropic v. Department of War Summary Judgment Motion Hearing +Date: 2026-07-31 +Status: draft +Category: social science +Tags: court reporting + +Dateline SAN FRANCISCO, Calif., 30 July 2026— A hearing was held on a motion for summary judgment in the case of _Anthropic PBC v. U.S. Department of War et al._ in Courtroom 4 on the 17th floor of the Phillip Burton Federal Building, the Hon. Judge Rita F. Lin presiding. + +The case is not going well for the government. Two days after [the last hearing in March](https://zackmdavis.net/blog/2026/Mar/dispatch-from-anthropic-v-department-of-war-preliminary-injunction-motion-hearing/), Judge Lin [issued a preliminary injunction](https://storage.courtlistener.com/recap/gov.uscourts.cand.465515/gov.uscourts.cand.465515.134.0_1.pdf) halting the implementation of President Donald Trump's [order for federal agencies to stop using Anthropic's technology](https://truthsocial.com/@realDonaldTrump/posts/116144552969293195) and preventing the Department of War from designating Anthropic as a supply chain risk. + +With no factual disputes requiring a jury to decide, the case was scheduled to be decided by Judge Lin on the basis of the written record. Anthropic filed [their argument for why they should win](https://storage.courtlistener.com/recap/gov.uscourts.cand.465515/gov.uscourts.cand.465515.166.0_1.pdf). Perhaps tellingly, the government's [rebuttal explaining why they should win instead](https://storage.courtlistener.com/recap/gov.uscourts.cand.465515/gov.uscourts.cand.465515.215.0.pdf) ends on a section explaining that "only modest relief is warranted" if Anthropic wins—and Judge Lin asked Anthropic to propose [what they think the final judgement should look like](https://storage.courtlistener.com/recap/gov.uscourts.cand.465515/gov.uscourts.cand.465515.240.0.pdf). Meanwhile, in Congress, [next year's defense appropriation bill](https://www.congress.gov/bill/119th-congress/house-bill/8800/text) adds language to the statute on the supply chain risk designation that prohibits designating a domestic company as a supply chain risk for declining contract terms. + +About a dozen spectators (including the present writer) dotted the gallery Thursday as the parties convened to discuss Judge Lin's [homework questions for both sides](https://storage.courtlistener.com/recap/gov.uscourts.cand.465515/gov.uscourts.cand.465515.242.0_2.pdf). Anthropic's contingent of ten people took up the long counsel table in the center of the courtroom, while the government's two lawyers sat in counsel overflow seating on the left. Michael Mongan of WilmerHale spoke for the plaintiff. The defense swapped in Department of Justice attorney James Harlow to speak (replacing Deputy Assistant Attorney General Eric Hamilton, who filled that role at the preliminary injunction hearing). + +[TODO— + +Judge Lin said the record was largely the same as for the PI, in some ways the record got worse for the government: no signs that Anthropic could alter the model after deployment + +the government's position is that Anthropic has inspired just a lack of trust + +"I find that position, if that's really what the government's position is, to be troubling." + +G: depends on the nature of the hypothetical punishment + +Pickering distinguishes between regulartoy or sovereign power, vs. acting as an employer + +in this case, all actions have been about the gov'ts wn IT + +J: what if it's just a contract termination, but the government says they're making an example +G: Pickering would apply +J: even if the purpose is to deter other contractors? +G: the gov't being clear about what it wouldn't tolerate from a contractor, that's not soverign power +J: what about the secondary boycott? +G: our position is that there is no secondary boycott +J: suppose there had been +G: if goverment said Bank of America couldn't use ANthropci on non-DoW work, that would be regularoty +J: but Hegseth's directive—would that go beyond pickering. Assum it's applied without limitation, as Hegseht wrote it + +J: this case is making me think of a lot of hypotheticals, bear with me; I think the shiould be easy if the government said it was terminating contracts for critcism of Trump +G: Yes, Pickering would apply +J: suppose the government say to be clear, this is to prevent unfair criticism of the prez +G: Umber and pickering can handle it (my gloss: it doesn't mean the gov't automatically wins) +It would be hard for the governemnt to meet its burden + +J: "I'm surprised that you can't give a yes to what seem to me to be an easy question." + +J: what if gov't say it's because we can't trust you; I assume Pickering still applies +G: Yes + +G: It would come down to the facts, not just the bare statement + +G says the government is not a monolith; the 1st amendment analysis of the President's action needs to be separate from the + +plantiff response + +A: Picking doesn't apply because providing AI is not comparable to anything a government employee woudl do +teh actions were not day to day maangement; they were a use of soverign power + +each of the chalenged actions was soverign power, natsec authority + +staggering implications of the answer you just heard + +I suspect that the reason weren't got geting a clear answer is that my collegue is a very good lawyer + +J: do you have a view as to wehtehr the actions are separate or as a whole +A: fact-specific, but on this record, the actions are link + +J: would-eviscerate qurstion +G: no, pickering would apply . Risk assessment memos. Frontier AI is a black box, not like a rifle that gets made to specification that you can inspect. Anthriopcii can put its corporate views into its products; DoW needs to trust + +J: in the AI context, public criticism is different? + +G: the leap of trust is greater than with military hardware It's not a one-time deal; the dept needs the latest models +J: is this really different from other defense contracts, like drones + +G: AI is staggerlying opaque + +J: what are the facts that ANthropci might sabotage + +G; the prime contractor was also concerned (my gloss: the Palantir thing) +corporate moral judgement + +and people not cleared at Anthropic talking + +communication shared with the company (Dario's memo) that expressed hosility) +do they see thesmevles as a partner +case law gives deference + +Platiff response— + +A": timeline matters. Mar 2 risk memo can't justify Feb 27 actions + +resist the notion that AI opaqueness obviates 1st A protection + +redlines where there from teh beginning +no indication that Ant took steps to intefere +an actual sabatour wouldn't taken a stand on usage restrictions + +"Despite my colleague's best efforts" + +the central concern was about ANthropic sabotaging the model after the fact, the shifting rationale is powerful evidnece of pretext + +J: windown question + +G: we're not sure what prompted this question. DoW wind-down ongoing, to be completed by Sept. 30 +another agency is moving off +other defendant agencies didn't say; otehrs were only using a pilot program which would expire Aug. 30 + +J: agencies using more Claude? + +G: defendant respectually objects to the question on national security grounds and because it comes after the challenged actions (!!) + +J: the reason I'm asking is bvecause it's inconsistent to expand use of Claude if Ant is untrustworhty +I was a prosecutor; sometimes actions afterwards shed light on motives + +G: I'm not authorized ot give a substantive response at this hearing + +J: maybe I'll issue a written order + +A: we think the question is relevant be we respect national securty; we want redress, if the court thinks importatn +A: two agencies have paused + +J: remand to the department, and sec boycott uinnesccary + +A: as a practical matter, ANt is not plausibly an adversary; as a formal matter, OK + +G: the secretary would retain authoirty to tkae separate action; remand doens't matter + +closing remarks— + +the mere fact that ther'es a regulation doesn't make it a sovering power +DoW was aware of Ant's publci statments and this not about the company's speech + +Re asking for who expanded usage of Claude: would ask for a week, would take work on our end; 1 or 2 days is not sufficent; 2 days or a week doenst' matter + +proposed order lumps together differne agnecies + +we dont' think the facts warrant futher relif +taileor relief to particular actions and agencies + +any proposed order should respect right to not use Anthropic, and there's no basis for a compliance report + +A: re the 2 days vs. 1 week, the gov'ts had these questions since Mondau; we've been suffering unconstituational burden since feb and we would like relief expeditiously; we appreciate the court and defendants moving quickly + +] -- 2.53.0