OpenAI Took a Classified Pentagon Deal and the Safety Nerd Quit on Cue

OpenAI signed a classified Pentagon deal and a senior robotics researcher basically hit “rage quit” like this was a busted Call of Duty lobby.

The agreement is a “deployment” contract with the Department of Defense for military AI “guardrails.”

The wild part isn’t that the DoD wants AI — the DoD wants everything, including your kidneys if they could file the paperwork. The wild part is the timing: a senior person on OpenAI’s robotics team resigned immediately and publicly said the safeguards weren’t good enough.

And because it’s classified, you don’t get the fun details like scope, money, or where it gets deployed. You just get the press-friendly idea that OpenAI is there to add safety bumpers.

Meanwhile, every company in AI is sprinting toward government cash like it’s the last chopper out of Saigon, because consumer AI is expensive, competitive, and allergic to profits. Defense money is the one place you can light billions on fire and call it “national security.”

If you’re wondering who benefits, it’s the same cast as always: contractors, executives, and whichever committee chair gets to sound serious on camera.

If you’re wondering who gets the risk, it’s everyone who lives in the real world where “oops” has consequences.