AI’s impact on work is genuinely contested: one study finds no job losses, another sees the market splitting in two, a tech founder predicts labour shortages. Every measure differs, and the loudest voices oversimplify.
The paper Agents of Chaos let AI agents loose in a live environment for two weeks, revealing failures of accountability that isolated, sandboxed tests would never surface, raising urgent governance questions.
In June 2026, the US government forced Anthropic to disable its newest frontier model three days after launch. What looked like a security measure revealed something more consequential: the limits of state power over cloud-distributed AI, and the cost borne by the humans caught in between.