This week, Anthropic researcher Jacob Coxon quit his job, warning that AI could “kill us all by the end of the decade.” Cue the headlines, the conference invitations, and fresh arguments for billions more in AI safety, regulation and research. Every so often, someone close to AI makes another extraordinary prediction about how little time humanity has left. I’m sure the concern is sincere, and the risks deserve serious study. But my cynical side can’t help noticing that predicting the end of humanity is a surprisingly effective way to keep both yourself and AI very well funded.
Last Updated on September 11, 2026 | Published: September 11, 2026