Well, it looks like OpenAI will be working with the Pentagon: https://www.axios.com/2026/02/27/pentagon-openai-safety-red-...

My personal guess is that Sam Altman said he'd let policy violations go without a complaint and Dario Amodei said he wouldn't.