Earlier this month, Elon Musk wrote about X Formerly Twitter’s integrated chatbot: “We have improved @Grok significantly, You should notice a difference when you ask Grok questions.” People did, in fact, notice a difference. Soon thereafter, Grok started spewing antisemitic content and eventually began to refer to itself as “MechaHitler”. That doesn’t seem like an improvement to me, but then, I’m not a neo-Nazi.

Anyway, sure, what could possibly go wrong with the U.S. Department of Defense giving a contract worth up to $200 million to xAI, to “develop agentic AI workflows across a variety of mission areas”?