...and they sold it to the pentagon!

Last post, I griped about not being able to afford a seat at the AI table — the training runs that cost more than a house, the GPUs that cost more than a car, the whole game priced out of reach for anyone without a wire transfer.

This is the part that actually worries me. Not who's paying to play. Who's already at the table.

The labs stopped being labs

Every big AI lab — OpenAI, Anthropic, the rest — started with the same scrappy research vibe. Idealistic. Mission-driven. We're going to figure out alignment, we're going to build this safely, we're going to share what we learn.

And then the bills came due.

OpenAI is now functionally a Microsoft subsidiary with ten-plus billion in deals and counting. Anthropic raised billions from Amazon and Google. xAI, Google DeepMind, Meta AI — they're all either inside or financially tethered to the biggest companies on earth.

I don't think this is malice. I think this is physics. You can't run frontier AI research on idealism and ramen. You need datacenters. You need power. You need the kind of capital that only six or seven entities on planet earth can provide. So the choice is: become a department of Microsoft, or stop existing.

The result is that the cutting edge of one of the most consequential technologies in human history is being decided in board meetings at companies whose primary historical concern was selling office software or shipping packages. We've handed the future to people who measure success in quarterly earnings calls.

And then they sold it to the Pentagon

When you're a lab that needs billions to keep the lights on, you take money where you can get it. And the deepest pockets in the world aren't tech companies. They're governments. Specifically, defense and intelligence agencies.

OpenAI quietly removed the "no military or warfare use" clause from its usage policies in early 2024. Not with a press release. Not with a town hall. They just edited the page. Anthropic has announced partnerships to put Claude into classified government environments. Palantir, which has been selling AI-flavored everything to ICE and the military for over a decade, gets called a respectable AI company now instead of the surveillance contractor it actually is. Anduril is building autonomous weapons and getting unicorn valuations for the privilege.

Remember Project Maven? Back in 2018, Google's employees famously revolted when they found out the company was helping the Pentagon analyze drone footage. Six years later, basically every major lab is doing some version of Maven, and nobody's even pretending to be conflicted about it. The employees who would have walked out either got with the program or got laid off in the last round of cost cuts.

We're also starting to learn what AI-assisted warfare actually looks like in practice — and the lesson isn't about one country, it's about what any military does once a machine will do the sorting for it. The reporting we have so far centers on a system Israel's military reportedly used in Gaza, called Lavender, that flagged tens of thousands of people as targets. By the IDF's own reported tolerance, it ran about a 10% false-positive rate. Ten percent. Of human lives. Run through a classifier and reviewed by an operator in something like twenty seconds — per the +972 Magazine reporting, long enough to confirm the target was male and not much else. There was reportedly a companion system, grimly named "Where's Daddy," built to track flagged people to their family homes, because that's where they were easiest to find, with a reported tolerance of double-digit civilian deaths for a single junior fighter. That's not a hypothetical. That's a system that got deployed, with all the human judgment in it compressed into the twenty seconds it takes to rubber-stamp a name. And the thing that should scare you isn't that one army did this. It's that the tool works, it exists, and every government on earth just saw the demo.

And on the home front? Clearview AI scraped billions of faces off social media and sold the database to police departments. Predictive policing algorithms are baked into law enforcement workflows in major cities, despite study after study showing they just launder the racism of the historical arrest data they were trained on. Facial recognition is at the airport, on the street corner, over the register at the convenience store. Every license plate reader is feeding a database somewhere.

I do security for a living. I spend my days thinking about threat models. And the threat model here isn't subtle: an unprecedented surveillance and targeting capability has been handed to whichever government happens to be in power, and governments have demonstrated, repeatedly and across every political stripe, that they will use whatever capability you give them on whoever they decide is the problem this week.

The labs all have very nice AI safety teams. They publish papers. They run red teams. They worry, very publicly, about whether the AI might turn evil on its own.

You know what they worry about a lot less publicly? Whether the AI will be turned evil by the people who paid for it. Because those are the same people writing the checks.

One of them actually tried. The Pentagon handed out AI contracts worth up to $200 million apiece to four labs — Anthropic, OpenAI, Google, xAI — and when it pushed an "all lawful purposes" standard, three of them agreed to drop the guardrails that apply to the rest of us. Anthropic didn't. In early 2026 it told the Pentagon its models were off-limits for autonomous weapons and for mass surveillance of Americans — the exact two lines you'd want a lab to hold. Within months the government branded the company a "supply chain risk," the first time it's ever pinned that label on an American firm, and ordered every agency to rip Claude out. Anthropic says it's retaliation, and it's suing to survive. So — can anyone push back? Sure. It costs you billions and a knife between the shoulder blades, while the three labs that said yes kept their checks. Everybody took notes.

And it's hard to bite the hand.