20
votes
Peter Thiel's new model army. The Palantirisation of the UK military is a national security disaster.
Link information
This data is scraped automatically and may be incorrect.
- Title
- Peter Thiel's New Model Army
- Authors
- Carole Cadwalladr
- Word count
- 1911 words
I was really disappointed by this one. For a piece that opens by criticizing pro-palantir puff pieces, it was disappointing to see that it's basically an anti-palantir puff piece. It's rage bait, not a serious reflection.
This is literally the full argument. It doesn't mention even once what Palantir will actually be used for, what the terms of the agreement were, etc. It's not really clear to me how this entails "the palantirization" of the UK military.
In fairness, the details of the deal are linked from the article in the line: "The UK Ministry of Defence has just signed a new £240 million contract with Palantir."
From that linked article:
Maybe that could've been included more clearly, but it seems like the focus was more on supporting the argument that Palantir are a threat than on diving into exactly what they're being contracted for. The Tesla analogy seems reasonable, no? We've seen numerous examples of tech companies bricking devices and remotely locking out users for any number of reasons, and the last time it came up (a guy locked out of his Apple ID due to a fraudulent gift card sold to him by a legitimate retailer) an awful lot of people were saying he shouldered a decent amount of the blame for trusting an inherently untrustworthy corporate entity in the first place.
Now scale that up to the nation state level, and magnify the risk to the point that people are literally shooting at each other when it comes into play. I wouldn't want the government to blithely allow Yandex or Huawei to be part of the MoD's "live operational decision making" pipeline either. It seems like don't integrate critical services from a potentially adversarial company under the jurisdiction of a potentially adversarial government and don't give that company, and by extension that government, deep access to sensitive data should almost go without saying? A lot of us old school nerds are more careful than that about our mundane day-to-day digital lives...
[Edit] For anyone who thinks this is still a hypothetical concern: the UK government, along with Germany, are currently leading plans to deploy more troops to Greenland, to deter US aggression (yes I am aware how insane that sentence sounds). US links, including information sharing with Palantir absolutely will form part of Germany and Denmark's risk assessment when it comes to classified information sharing with the UK. The act of signing the deal alone creates a conflict of interest in operations with reliable allies.
Trump has already imposed sanctions on judges from the International Criminal Court in the Hague, essentially freezing them out of modern society thanks to the global reliance on US financial and tech firms. If that's the punishment for upholding human rights law, I genuinely cannot think of a reasonable argument to suggest that private US companies won't be leveraged as a threat against any person, organisation, or nation going against Trump's whims. Whether or not it follows through to direct action, what possible benefit can there be from allowing the US to hold that card in negotiations?
For what it's worth, the headline there was really misleading. The plan is to deploy NATO troops to show that NATO is securing the artic because Trump's argument has been that NATO isn't taking the artic seriously. It is not to secure the arctic from the US.
This is getting way into the weeds, but I think there's a fundamental misunderstanding among most people -- inlcuding the author of the original article, given the Tesla analogy -- as to how services like this work. The UK isn't giving Palantir their data. They're licensing software from Palantir to analyze data they already have. Palantir can't actually see the data. It's roughly the difference between using chatgpt (your data goes to openAI's servers) and using a local LLM that you control (you didn't make the model, but your data doesn't go anywhere.)
Now, I understand your argument about backdoors. I don't think that's super relevant here because the UK is already part of the 5 Eyes and the US's closest ally, but I get where you're coming from.
[Edit: ignore this first bit, I should've left it out - it's debatable either way, and that's distracting from the actual focus of the thread]
Agreed that it's not a great headline, but I do also think it's important to keep the subtext clear: this is showing "NATO securing the arctic" in the same way that carrier groups just happen to sometimes choose to do "navigation exercises" off the coast of Taiwan. It's a diplomatically acceptable pretext that allows some level of preparation in case the worst happens - and some level of posturing - without locking any side into overt escalation.Statements like "I'd love to make a deal with them, it's easier. But one way or the other, we're gonna have Greenland.", and "If it affects NATO, it affects NATO. But they need us much more than we need them." are a pretty clear statement of intent, after all - the only question is how much it's bluster.
Fair to push back on the details, but I can't agree with the analogy or the conclusion. Major enterprise contracts like this come with incredibly deep integration between organisations, significant access to existing software deployments, bespoke build work to integrate the new tools with the old, and usually a generous handful of full time staff from the supplier being deployed to work full time at the client organisation for the duration of the contract.
This isn't lightweight stuff running on a few thousand laptops, either, it's a huge centralised installation which as far as I can tell is going to be deployed as part of the MoD's £400M Google Cloud contract they signed a few months ago. I was looking for a link to back that one, since I was already aware of it and wanted to make sure I got the facts right, and ironically the MoD's own press release paints it very heavily as a US-UK technical integration - they actually make it sound much worse than I was going to!
So yes, it's technically a software licensing deal where the MoD maintains control of the data. But the software is going to be integrated, customised, and managed by Palantir staff, who will necessarily have access to other MoD systems for that integration, and the data will be stored and processed on servers run by Google on the MoD's behalf. Since the theoretical adversary is by far the most advanced state level actor on the planet, with a recent track record of massive data exfiltration, including attempted access from Russian IPs, and introducing unauthorised network connections into secure environments, along with absolutely no respect for the law, I think it's fair to be wary of data running through those systems even if they are nominally controlled by the MoD.
We're off the edge of the map here - you're absolutely right to mention 5 Eyes, and NATO in general, because in any normal time those would be enough to say I'm worrying about nothing. But these aren't normal times, and if the call comes to put the US on an information diet, it's going to be a lot harder to do so (and to be sure we've succeeded in doing so) when they've provided the support staff and the underlying infrastructure that the data processing is running on. It's going to be a lot harder to negotiate when "do what we say, or that critical decision making infrastructure stops working" is a viable threat. And it's going to be a lot harder to refute statements like "But they need us much more than we need them." when the MoD are rolling out tenders that explicitly say only Palantir can do the job they're asking for.
It is not the same as the posturing that the US does when it puts a carrier in a contentious region. When the US does it, there is an implied threat. There's the subtlety of teenagers at prom when they do it. There is zero, absolutely zero, threat implied to the US by the proposed European forces. It's currently popular from leftists to frame these European moves as the continent standing up to the US; I saw other similarly bad headlines the other day. The real situation is that Europe is trying to appease Trump on the security posturing that they frankly were warned about two decades ago but did not prepare for, hence the scrambling. They are not implying use of force against the US because it would be the hollowest of threats given their security situation. The only credible threat Europe has are diplomatic and economic.
I don't really know how else to respond here other than "no, it's not."
I don't want to defend Trump here. What he's doing is reprehensible. But it's, totally separately from that, false that the UK and Germany are posturing with force against the US.
Edit: I totally agree with your edited comment!
You say that as if the argument isn't substantial.
Do you believe that it's not true or just that it doesn't matter?
Well, I'd disagree that it's substantial. I guess maybe a better way to put it is that I was expecting a piece by an author experienced in national security. Here's what I mean -- I think it's reasonable to argue that you don't really own a Tesla or don't really own any other subscription-based thing you can't fix yourself. That's what right-to-repair is all about, right?
But if you don't own your Tesla, that means you don't own all things Car in your life. The Tesla is the whole category of Car for you. If it stops working, you're screwed. It's an absurd analogy for a military contracting a service with a software company that focuses on intelligence analysis. If Palantir stops working, the UK military still owns all their data. And, obviously, all the actual guns and planes will still work. It's much more like a business contracting out garbage removal services or something. If it suddenly stops, that's bad and garbage is going to pile up, but the business doesn't just freeze and die.
I'm not saying there's no conceivable way to make the analogy work, but "think about Tesla! [taps head]" isn't sufficient. You know? If the author was more experienced with what these contracts look like or what these companies actually do for militaries, I think the piece would've looked very different.
Okay, so the article outlines a bunch of worrying things that should not be things at all, but for some reason they are:
You can't seriously be of the opinion that the Tesla analogy was "literally the full argument" of this article, can you?
(I didn't read the solutions part yet but I'm glad to see there is one.)
Of course I can. :) it was a little hyperbolic, but not much. You responded in a couple of comments but I'm going to roll my reply into one.
Almost everything else is tangentially related at best. Starmer doesn't call the special military operation in Venezuela a violation of international law - what does that have to do with palantirization of the military? The noncompetitive tender is not at all unusual for a unique product offering, which Palantir is - a more experienced analyst (I'm referring to the journalist here, not you) would know that. Same thing with the whole "it's more than a contract, it's a strategic initiative" thing - I laughed at that. "New strategic initiative" is boilerplate language, it's not real.
Ironically this is what I was hoping the actual article would be about. Kudos to you for grasping the meat of the matter, seriously. This is the key question, yeah.
The point is this: there may be serious questions about using Palantir. You yourself raised at least two good ones. I'm disappointed by the article because the article doesn't raise any at all. With respect, you obviously hate Thiel and Palantir, and so you already agree with the author, which I think is why this article is more convincing to you.
Thanks for elaborating. Full disclosure, I didn't read the article so I'm just commenting on your comment. :D
And I'm definitely not an expert. My layperson brain just goes:
'Peter Thiel: crook. Software: fallible, obscure, incomplete, not fully auditable, exploitable. -> Crook will exploit software for personal gain if he can.'
Are you experienced yourself (just asking for context)? Does it matter what the contract says if the software comes with a built-in mechanism that lets it do rouge stuff without anyone noticing? Alternatively, in the event of a major breach of trust between nations when contracts stop mattering, how does the UK know what actions can be taken against them using this software, without them being able to do anything about it?
Would a journalist with experience on military contracts have realistic, up-to-date information on such matters? My intuition is that they might not, but the threats may nevertheless be real.
If I, as a layperson, were in charge, I would carefully consider the source of software purchases that pertain to national security and it would matter to me whether or not the supplying company was built and ran by an obviously morally compromised, callous and power-hungry individual. Not saying that finding morally uncompromised suppliers is easy - but still, this particular person and company? Isn't every sane person completely convinced to not try to collaborate with them? Sure, the article could be poorly worded and researched and annoying, and if so, I get the frustration, but the point is still worth stating.
(I'm curious if the journalist mentions what could be done to avoid this alleged "palantirisation" of their military, so I guess I'm going to have to go and read the thing.)
It's not even about Palantir. Maybe if you thought it is, that explains your disappointment in the article?
Rage bait is an attempt to instill anger by untrue claims or unjustified fear mongering. I didn't see any such things here, and the article ends with a long section that specifically attempts to give people hope.