The messengers changed their story. The data didn't.

Since this report was published, the CEOs of OpenAI and Anthropic have softened the AI job-loss warnings we built on. We think you should know that before you read on — so we've written up what changed, what didn't, and why we still stand by the argument.