Why Dario Amodei's AI job bloodbath hasn't arrived yet
Anthropic's head of economics says AI has not materially raised US unemployment, even as CEO Dario Amodei has warned of a white-collar bloodbath. Peter McCrory's data-driven X essay cites full-employment metrics and no job-loss spike in AI-exposed roles—yet early hiring cracks for younger workers remain.
Key Takeaways
- Peter McCrory says 18 months of Anthropic research shows no material AI-driven rise in US unemployment.
- CEO Dario Amodei has warned AI could wipe out half of entry-level white-collar jobs and push unemployment to 10–20%.
- June's 4.2% US jobless rate still matches the Fed's full-employment band, with prime-age employment near multi-decade highs.
- Hiring has softened for younger workers in high AI-exposure roles, an early signal McCrory flagged.
- Anthropic is committing $200 million via its Economic Futures Research Fund to study how societies adapt.
Has Dario Amodei's white-collar bloodbath arrived?
Not according to the latest public assessment from Anthropic's own economics lead. In a lengthy X essay summarized by Fortune and Storyboard18, Peter McCrory argued AI has caused no meaningful jump in US unemployment—a sharp contrast with CEO Dario Amodei's repeated disruption warnings.
McCrory said Anthropic's view rests on about 18 months of company economic research. He pointed to a June unemployment rate of 4.2%, still consistent with what the Federal Reserve treats as full employment. Job openings, he added, remain broadly in line with the number of unemployed workers, while prime-age employment sits near multi-decade highs.
Why does McCrory say jobs are still holding up?
Updated analysis using recent Bureau of Labor Statistics data, McCrory said, found no meaningful unemployment rise among workers whose jobs are more exposed to Claude's automation capabilities versus less-exposed occupations.
"I don't expect unemployment to be noticeably higher a year from now, at least not because of AI," he wrote. He attributed resilience to AI's uneven capabilities: no occupation in the US Department of Labor's O*NET database has every task fully handled by Claude, so human supervision still matters.
McCrory also said Claude is more often used as a collaborative tool than a replacement. Users with stronger domain knowledge tend to get better results and catch AI errors—suggesting today's systems complement skilled work rather than substitute for it at scale.
Where does Dario Amodei's warning still land?
The tension is real. In May 2025, Amodei told Axios AI could eliminate half of entry-level white-collar jobs and push unemployment to between 10% and 20% within one to five years. His January 2026 essay, The Adolescence of Technology, cast AI as a "general labor substitute for humans."
By May 2026 he framed AI more as a productivity multiplier, invoking the Jevons paradox—automating most of a job could expand remaining human work. In June he again warned lasting job losses could be inherent to AI and floated policies such as wage insurance and universal basic income.
McCrory acknowledged early caution flags: weaker hiring for younger workers in high AI-exposure occupations, consistent with Stanford research, plus BLS projections of slower growth through 2034 for roles such as technical writers, data entry workers, and customer support representatives.
Separately, Anthropic outlined a $200 million Economic Futures Research Fund agenda to study workplace design, retraining, income support, and worker stakes in AI-driven growth—signaling the company still expects disruption worth preparing for. For more on AI's economic stakes, see BlasterPost's Future Tech & AI Wonders coverage.