What happened
Anthropic published new survey evidence from two August 2026 surveys — a nationally representative sample of US adults and Anthropic's own staff — comparing how each group expects AI capabilities, adoption, and labor-market impact to evolve. The headline finding: "39% of US adults and 91% of staff expect that AI will be able to make a Nobel-level discovery by 2030." On their own jobs, "over half of Anthropic staff expect AI to be able to do at least 80% of their current work by 2030, compared with 5% of US workers" (staff average 75% vs. 28% for US workers). The note also finds the two groups largely agree on today's capabilities but diverge sharply on the pace of progress, and that both groups reason about unemployment in the same way at a given level of expected automation — so the disagreement is about capabilities, not about labor-market logic.
Why it matters
This is the clearest available measurement of the 'expert vs. public' expectation gap on AI capability timelines — directly relevant to boards and executives calibrating workforce and scenario-planning assumptions against what AI builders themselves believe.
Action needed
Use the gap between public (39%) and frontier-lab (91%) expectations as a calibration input in board scenario planning and workforce-reskilling assumptions.