Thread

1. petcat · 1790353704 · focus · HN ↗

   I still don't fully understand what this was about. It sounds like the DoD said they wanted unrestricted access to Anthropic's models, Anthropic refused, so the Pentagon declared that they wouldn't use them at all.

   Isn't this basically what Anthropic wanted?

   1. tcdent · 1790355688 · focus · HN ↗

      As I understand it cane down to one point: Anthropic wanted the ability to review actions that would have an effect on human life.

      Which sounds sensible on the surface, until you realize that all military organizations have extensive experience with that exact question, and have spent centuries refining their practices. Deferring that to a private company is actually a major regression and a concerning role to pass on to the public regardless.

      FWIW I am not pro military, but it's a prominent part of human culture and not going away anytime soon. I do understand the nature of professions, however, and only one of these groups has made the interpretation of this question their specialization.

      1. daveidol · 1790356307 · focus · HN ↗

         You raise an interesting point. Although wouldn’t Anthropic be the best experts on when a model would be most applicable or likely to falter?

         I guess my opinion would really depend on the exact wording and details of the arrangement. Lots of people jumping in with strong stances based on pretty limited details.

         1. tcdent · 1790356675 · focus · HN ↗

            On faltering: the US Military already has rules around the use of automated systems for this type decision making. This is not a new concept to them by any means, and currently it always involves a human decision.

            1. garciasn · 1790356947 · focus · HN ↗

               But, yet, because the technology has advanced so very rapidly and large bureaucratic organizations cannot keep up, we almost went to war w/China because of it.

               I am going into this eyes open and realize that we did indeed stop it before we hit the button, but when there is immense political pressures over career military experts, we may end up with someone ignoring those HITL stops.

            2. sfink · 1790370443 · focus · HN ↗

               Anthropic agreed to uses that involve a human decision, so this is a non sequitur.

               The whole conflict is specifically about uses that do not involve a human decision. The DoD agreed to a contract that said it wouldn't use Anthropic for that, and then tried to change the terms.

               Incidentally, the DoD used Claude in its decision that resulted in killing 120 children at an elementary school for girls. The Pentagon concluded that there was human review (which would mean that it fell within the contract's terms) but that the review was hollowed out (which could mean that it was a violation of the contract).

      2. hardbass · 1790356572 · focus · HN ↗

         Frankly I would trust Anthropic much more than the US military with human lives.

         1. scottyah · 1790357555 · focus · HN ↗

            Would you still in 10yrs? 30? When all the current employees have left, and anti-Western people make most of the decisions?

            1. hardbass · 1790357754 · focus · HN ↗

               Anti western people are already in the government now. I'd worry about the present than future hypotheticals.
         2. esafak · 1790358838 · focus · HN ↗

            Which part of the military? I would trust the rank and file. The commander in chief?

            1. hardbass · 1790358964 · focus · HN ↗

               Whichever part caused the girl school to be hit. Which has to be multiple levels for it to occur.

               1. myko · 1790360134 · focus · HN ↗

                  The DoD is blaming AI for this (you are of course right though)

                  1. hardbass · 1790360223 · focus · HN ↗

                     Anthropic is vindicated in that you need oversight over AI being used to kill people, the USA military clearly showed itself incapable of wielding that responsibility.
         3. karmasimida · 1790361563 · focus · HN ↗

            I don’t, for profit company you say, nah

            1. Analemma_ · 1790362354 · focus · HN ↗

               The US military has assassinated American citizens with no judicial review, Anthropic has not.
      3. felixgallo · 1790357000 · focus · HN ↗

         you understand wildly incorrectly; please go and do some research.
      4. cubefox · 1790357287 · focus · HN ↗

         > As I understand it cane down to one point: Anthropic wanted the ability to review actions that would have an effect on human life.

         This is false. Anthropic wanted assurance that its technology would not be used for fully autonomous weapons or domestic mass surveillance.

         1. tcdent · 1790357753 · focus · HN ↗

            > [Anthropic] would not permit its technology to power fully autonomous weapons without human oversight over targeting and firing decisions. Defense Secretary Pete Hegseth argued that the Pentagon should have unrestricted access to AI systems for “any lawful purpose” and that a private contractor should not be permitted to dictate how the military can use its technology.

            It's about them maintaining an interpretation of law at all.

            Simpletons read that as: "The DOW wants to kill people with AI."

            <a href="https:&#x2F;&#x2F;www.pearlcohen.com&#x2F;anthropic-sues-department-of-defense-over-supply-chain-risk-designation&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.pearlcohen.com&#x2F;anthropic-sues-department-of-defe...

            1. simianwords · 1790361722 · focus · HN ↗

               well if you were Anthropic and you didn&#x27;t want the government using your technology in these ways, how would you phrase it? What else would you have done?
            2. cubefox · 1790362077 · focus · HN ↗

               &gt; Simpletons read that as: &quot;The DOW wants to kill people with AI.&quot;

               The DoW actually used AI in the Iran war, which led to them attacking a school and killing more than a hundred childrem. That&#x27;s exactly what Anthropic wanted them to prevent. <a href="https:&#x2F;&#x2F;www.bloomberg.com&#x2F;graphics&#x2F;2026-iran-school-attack&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.bloomberg.com&#x2F;graphics&#x2F;2026-iran-school-attack&#x2F;

               1. tcdent · 1790370529 · focus · HN ↗

                  The DoD has been using AI for engagement since the 80s. Maven, the system used in your example has been active since 2017.

                  None of those systems pull the trigger. They assemble data for human review.

                  1. noahbp · 1790374596 · focus · HN ↗

                     You seem to be misunderstanding the facts. Claude was already used by the US Military, under an existing contract between Anthropic and the US Department of War.

                     The core dispute was the US Military deciding it needed to renegotiate this contract, but without Anthropic&#x27;s two red lines: No fully autonomous weapons (i.e., zero humans in the loop), and no mass domestic surveillance.

      5. SillyUsername · 1790358035 · focus · HN ↗

         Yes, that&#x27;s exactly why they used AI to blow up a school murdering a load of kids.

         That kind of extensive experience.

         Anthropic did the morally right thing and are being punished for it. The case in point justifies their position.

         As they say, no good deed goes unpunished.

      6. everforward · 1790360081 · focus · HN ↗

         Militaries have a lot of experience, and an exceptionally poor track record.

         I don’t remember the last war that didn’t have credible evidence of war crimes occurring. Were bombing civilian infrastructure in Iran, Iraq had Abu Ghraib among all the Collateral Damage stuff, the Highway of Death in the Gulf was probably a war crime, Vietnam had My Lai, WWII was the advent of carpet bombing civilian infrastructure. I can’t think of any for Korea, but I also know very little so that doesn’t say much.

         We still haven’t charged anyone for the second strike on that fishing boat in South America, and I haven’t heard a single rationale for why that’s not a war crime other than “fog of war”.

         The US doesn’t even really have a meaningful system for finding and prosecuting these, because we aren’t signatories for the ICC and have a bill saying we’ll invade if they charge one of our service members with a war crime. We aren’t basically the furthest thing from having any experience prosecuting war crimes. I can probably count on my fingers the number of cases we’ve tried. We rarely charge our own service members, and we usually kill foreign combatants rather than capture and charge them.

         1. tcdent · 1790361030 · focus · HN ↗

            A poor track record compared to who? The Romans?

            War includes death by design. The trajectory is trending toward less unjustified death.

         2. daedrdev · 1790378853 · focus · HN ↗

            The US could carpet bomb their enemies like they did in WW2, instead they pursue one of the lowest civilian casualty methods of in the history of the world, and nobody even cares because somehow that’s not good enough

            1. whatisthiseven · 1790408995 · focus · HN ↗

               What makes you think these targeted strikes are any better? Did the school girls say, &quot;o, but at least our entire city wasn&#x27;t carpet bombed like an obvious war crime, instead we were drone strikes in such a morally ambiguous way. O well&quot;

               No.

               And why did America have so many enemies???

               Maybe because we keep finding excuses to invent them, and then does hundreds of &quot;morally grey&quot; bombings. But at least it wasn&#x27;t another Vietnam, right?

               What a low bar.

            2. everforward · 1790460043 · focus · HN ↗

               Bombing civilian infrastructure is a war crime because the end state is the same as bombing civilians. Killing power shuts down hospitals and emergency responders (generators run out eventually) and desalination plants. The military is largely unphased, they’re the first ones to get gas for generators.

               Also, this is carpet bombing. Carpet bombing is the targeting of civilian infrastructure with effectively willful ignorance of the collateral damage.

               We started a war knowing the only way to win was either boots on the ground or war-criming our adversary into submission. We don’t get to pretend our hands are tied and we have to send out the bombers. We don’t have to do this because Iraq forces us to, we have to do this because we elected a man who can barely read Post It notes and ignored half a century of military intelligence. We are at fault for every dollar of damage caused to infrastructure and every life lost.

      7. folocitoan · 1790370458 · focus · HN ↗

         &gt;Anthropic wanted the ability to review actions that would have an effect on human life.

         They wanted assurances that it would not be used for mass surveillance or autonomous weapons. They did not request to be &quot;in the loop&quot;.

         Furthermore, even if Anthropic had requested that, it would be a reason for the DoD to cancel the contract because the terms were unacceptable, not to declare Anthropic a supply chain risk. Declaring them a supply chain risk requires proving that they intended to sabotage their own product, adding malicious functionality. There is no evidence whatsoever that they planned to do that.

         <a href="https:&#x2F;&#x2F;www.cbsnews.com&#x2F;news&#x2F;anthropic-pentagon-supply-chain-risk-lawsuit&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.cbsnews.com&#x2F;news&#x2F;anthropic-pentagon-supply-chain...

      8. tastyface · 1790373490 · focus · HN ↗

         120 Iranian schoolgirls would have loved to hear this lecture.
      9. noahbp · 1790374303 · focus · HN ↗

         &gt;As I understand it cane down to one point: Anthropic wanted the ability to review actions that would have an effect on human life.

         This is not at all what happened. The existing contract that Anthropic had with the government had two red lines: No autonomous weapons (i.e., do not build Skynet with Claude), and no mass domestic surveillance.

         Anthropic did not, and does not want, any case-by-case veto power over the US Military&#x27;s use of their AI models. The red lines were negotiated ahead of time.

      10. adastra22 · 1790405608 · focus · HN ↗

          Downvoted for misinformation, likely out of ignorance, despite the reasonable argument made on top of it. Anthropic was not asking for the ability to review and block particular uses. Those are classified and wouldn’t even be running on Anthropic’s servers. They just wanted the government to promise not to use it for killing people without a human in the loop.
      11. trashface · 1790469629 · focus · HN ↗

          &gt; Which sounds sensible on the surface, until you realize that all military organizations have extensive experience with that exact question

          Tell that to a certain school in Iran

Unofficial Hacker News client; not affiliated with Y Combinator.