Palantir’s ‘Progressive’ Manifesto: an AI Death Drive?

On Saturday, AI company Palantir released its vision of a new world. Except, it turns out to be a pretty old one – a vision that the distinctly moral idea of ‘progress’ once reassured us we were accelerating away from.

Not so for the hobbits of targeted-strike-accelerator Palantir, for whom progress is just a technological notion, one measured in ‘clean’ kills. In a lengthy, 22-point manifesto published on X, the company summarised a 2025 book co-written by Palantir co-founder and weirdo-in-chief Alex Karp. It made for hard reading back then, 18 months into the world’s first AI-assisted genocide being carried out by Palantir-pal Israel in Gaza.

It makes for equally troubling reading now, even in summary, as a deal to end the US’s illegal war on Iran looks increasingly shaky. In that context, the timing of the post, of course, is no accident. Clearly, Palantir’s services to the US military include not just AI-hit pros, but AI-shit prose too.

Last week, the baby-faced Louis Mosley, head of Palantir’s UK operations, appeared on BBC Politics Live, where he tried to defend the UK government’s insistence that Palantir get a foothold in our NHS. Projecting both ‘beta-blocker’ and ‘I’m a real boy’, Mosley went to bat for a company that would have made his grandfather (yeah, that one) proud.

“Lavender and Gospel,” he said, referencing two systems that have been used by the Israeli military to target Palestinians in both Gaza and the occupied West Bank, “We have nothing to do with.”

Despite celebrating a mid-genocide contract with Israel in early 2024, that’s been Palantir’s consistent line. In response to accusations of assisting war crimes made by Amnesty International, which the company denies, it’s still “proud to support Israeli defence and national security missions in other programs and contexts”. Karp himself has elsewhere appeared to speculate on the riches to be made from legalising such crimes.

It’s no surprise that people connect Palantir with Lavender and Gospel. These are pattern-finding technologies that use data such as mobile phone records and GPS histories to prime drones. And Palantir has built its brand on finding patterns.

You want a maximally efficient chicken nugget supply line to supermarkets and fast-food franchises across a continent? Pump in your massive dataset of customer purchase patterns and let the AI set the wheres and whens. You want to turn human beings into chicken nuggets? Do exactly the same.

It takes a certain philosophy to build a business servicing the supply chain demand for animal slaughter identically to the kill chain demand for human slaughter. And that philosophy was unashamedly on display this weekend.

“The question is not whether AI weapons will be built,” was point five of the company’s post, “It is who will build them and for what purpose.” That’s the logic of an arms race, and if it’s not cold-war enough for you, Palantir’s vision of ‘the enemy’ is familiarly propagandist. “Our adversaries will not pause to indulge in theatrical debates about the merits of developing technologies with critical military and national security applications,” the company warns, “They will proceed.”

“One age of deterrence, the atomic age, is ending,” the company went on to say in point 12, “A new era of deterrence built on AI is set to begin.”

It’s worth pointing out, as US negotiators head to Islamabad for further talks to end a war begun – very allegedly – over Iran’s nuclear capabilities, that the same country pursuing peace through illegal war is also the only one ever to have used nuclear weapons. Far from deterring the threat of civilisation-collapsing weaponry, America’s carnage in Nagasaki and Hiroshima made the proliferation of insta-mass murder technology a defence necessity.

Palantir’s insta-mass murder technology 2.0 has already been on display in Iran. When almost two hundred people – mostly girls under the age of thirteen – were killed in a US strike on a primary school, it was the company’s Maven system, deeply embedded in the US military, that is alleged to have picked the target. The aerial slaughter (again, illegal) of hundreds of Iranian government and scientific figures – including the country’s head of state – is said to have been facilitated by the same system, which, according to one source, allows soldiers to select up to eighty targets an hour, in just “four clicks”.

All of this is justified, according to Palantir, by supremacy. “Some cultures have produced vital advances,” it asserts in point 21, “Others remain dysfunctional and regressive.”

If you’re not sure which, according to Palantir, is which, it goes on. “No other country in the history of the world,” it asserts, “Has advanced progressive values more than this one.” But there’s little progress to be found in Palantir’s vision. Instead, it’s rooted in a pretty old idea, one espoused by US president Donald Trump’s deputy chief of staff Stephen Miller earlier this year: might makes right.

Generations of schoolchildren have been taught that it required a world war to defeat that idea. But some philosophies, I guess, never truly die.