Louis Mosley
642 posts

Louis Mosley
@louismosley
Leading @palantirtech in UK & Europe



How tech giant Palantir was recruited by the police, NHS and military The US data analytics company is helping to fight crime, cut hospital waiting lists and bolster asymmetrical warfare. But is it a force for good, asks @RSylvester1 bit.ly/4bV4wLn

Domination @nvidia ♥️ @PalantirTech ♥️ @Dell




Forgive me. I am going to begin by quoting two prominent left-wing Londoners – and agreeing with one. ✍️ Louis Mosley spectator.com/article/ai-is-…



“When we’re presented with the biggest challenges that the NHS faces today, having a Palantir engineer working side by side with us in partnership means that we know that we’ll come up with the solution.” Rebecca Llewellyn, Director of Data Management and Transformation, @NHSEngland. We're proud to serve the UK's most cherished institution on its path to a fully connected NHS. With 150 Trusts across onboard and 80,000 additional operations already delivered, The NHS Federated Data Platform is harnessing technology to deliver better patient care.


🗣️ "Zack Polanski is putting ideology over patients" Palantir's @louismosley speaks to @WilfredFrost about the criticism the company has faced after working with the NHS trib.al/mbeVpLj 📺 Sky 501, Virgin 602, Freeview 233 and YouTube

“Ukraine will soon deploy a new generation of domestically produced air-defense interceptors, powered by artificial intelligence, that could allow the country to fight on indefinitely.” How Ukraine plans to stop Russia’s Shahed drone terror — read by @washingtonpost: washingtonpost.com/opinions/2026/…

We are launching Brave1 Dataroom with @PalantirTech @louismosley. AI based on real war data will help intercept enemy drones and protect Ukrainian airspace. Ukraine is developing autonomous air defense solutions that are already delivering results.



Palantir replied 👇to my post on their French intelligence agency contract with, as expected, a lot of lies and gaslighting. Their main argument seems to be "there's no issue and for proof the French are so satisfied they keep renewing." This is completely false. The fact is that Palantir was always meant to be a "transitional" solution before a French alternative could be developed. That's according to the DGSI themselves as quoted in Les Échos, France's leading financial newspaper (lesechos.fr/tech-medias/hi…). Guillaume Poupard, the head of France's National Agency for Security of Information Systems, also confirmed this in an audition to France's National Assembly (assemblee-nationale.fr/dyn/opendata/R…): "regarding Palantir, there is an overall will to create a trusted French alternative". Subtext: Palantir isn't trusted. The background is that Palantir was initially adopted as an emergency short-term measure in the wake of the November 2015 terrorist attacks, the deadliest in the history of France. There was, understandably, panic at the time to get fast solutions in place to prevent the next attack, and Palantir was the only option so the DGSI *reluctantly* adopted it. Since then, there have been multiple attempt to get rid of the dependency. These include: - Programme Artemis (Ministry of Defense): launched to create a sovereign French architecture for massive data processing, involving Thales, Sopra Steria, Atos and Capgemini. Problem: they're insanely slow, they're saying it'll take them 15 years to develop a solution (shs.cairn.info/revue-defense-…) - The "GICAT" project: a consortium of 22 French companies coordinated by GICAT (defense industry group). Problem: this is a very fragmented offering where they try to replicate Palantir with several companies that each offer part of the features and have little interoperability (shs.cairn.info/revue-defense-…) - OTDH Tender (2022): the most serious attempt - a formal procurement process launched by the DGSI. Started with 9 candidates, narrowed to 3 finalists: Athea (Atos-Thales alliance), ChapsVision, and Blueway. A migration was even planned before the Paris 2024 Olympics. For some obscure reason, it didn't happen but, according to ChapsVision, "we're nearly there" (franceinfo.fr/vrai-ou-fake/c…) So no, the truth is that the DGSI is very much NOT satisfied with Palantir and has been actively trying to get rid of it. So much for the satisfied customer story Palantir is trying to sell in their response... Also, on a technical level, their main argument is "but the data stays in France". That's not the main issue at all. The main threat with Palantir, and we've seen the U.S. use this repeatedly in their warfare (for instance against China or Russia), is a coercion scenario whereby, if France ever took a position seriously adverse to US interests, the US could pressure Palantir to stop providing updates/support, withdraw personnel or even operate a kill switch to shut down the software. And poof, suddenly the software that constitutes the "central architecture" of France's intelligence services stops working. That's not far fetched at all, ask Russia which saw U.S. software companies pull the plug overnight in the wake of the war in Ukraine. Also, France might control the data but they surely don't control the code, and that's pretty damn important for sovereignty. Yes, on an air-gapped system, real-time data exfiltration is implausible. But the more insidious threat isn't what the software sends out, it's what it chooses not to show. An algorithm can be designed to deprioritize certain results, miss certain connections, fail to surface certain targets. If Palantir's software is quietly blind to threats the US doesn't want France to see - say US-protected assets - how would France ever detect it? Gaps in results are invisible and completely deniable. This requires no internet, leaves no trace, and is unprovable without source code access, which France undoubtedly does not have. So all in all, this is an insane vulnerability for a country's core intelligence infrastructure and France knows it. You don't launch three separate programs to replace a software you're happy with. France isn't renewing because they're satisfied, they're renewing because they haven't managed to get rid of Palantir yet. Palantir says "pragmatism, not ideology, is what will keep France safe." I completely agree. And the pragmatic thing to do here is not to run your intelligence services on code you can't see, from a country you can't trust. That's not "technically illiterate national chauvinism," it's simply common sense.


This is frankly a complete dereliction of duty. Not only have France's intelligence services been using an American company to analyze some of the country's most sensitive data for a decade, but - in the current context - they're insane enough to re-sign for another 3 years. Try to square this circle: - The Americans, in their National Security Strategy, openly say that one of their key strategic priorities is to "Cultivate resistance to Europe’s current trajectory within European nations," i.e. foreign interference aimed at regime change - Europeans are starting to recognize the problem, with Germany's Merz now essentially saying that the Americans are an adversary (x.com/RnaudBertrand/…), that they're "pursuing their interests very, very aggressively" and that Europeans can "only respond" by also doing so. - Yet France's own intelligence services are handing some of their most sensitive data to an American company. And not in a small way: according to French media (lessentieldeleco.fr/4824-palantir-…) Palantir now constitutes the "central software architecture" of the DGSI. You couldn't make it up. Even Israel, despite being joined at the hip with Washington, won't let Palantir near its core intelligence systems - Unit 8200 and the Israel Security Agency (Shin Bet) rejected the system precisely because of sovereignty issues (en.globes.co.il/en/article-wha…) France, meanwhile, is like "sure, help yourselves and let's renew for another 3 years - we're confident you'll only use this data to protect us, not to 'cultivate resistance' to our government." At some point it's so absurd and strategically incoherent that the Americans can save themselves the trouble of "cultivating resistance", this level of incompetence does the job all by itself.