A Method Is Explained

The city of Ledgerville! The robot study's grades came from an AI. Mayor Pennywhistle: robots grading robots?! Dash: not robots, a language model rated the 7,594 tasks. Zap: four tiers, fenced cell up to open street. Pop: half the work needs the cell; street-ready, under two percent (1.9%). Baron Von Bottomline: a model-derived exposure taxonomy! Dash translates: guesses by a model, not robots watched at work, as the authors themselves note. And so, a method is explained.

The robot study's numbers were produced by an AI: Anthropic's economists had Claude rate 7,594 physical tasks from the Labor Department's job database on four tiers by where a robot can do them, from a fenced factory cell to an open street; half of physical work sits in the fenced-cell tier and only 1.9% in the open-street tier, and the authors note the ratings come from a model rather than observed deployments, so AI-driven robots could leapfrog the scale and make their 40-year projection far too slow.

Starring

🎙️The AnnouncerMayor PennywhistleDashPopZapBaron Von Bottomline