Jack Clark
Jack Clark is the Anthropic co-founder interviewed in Live: Anthropic co-founder on AI and jobs during Planet Money’s book tour. The episode presents him as both a frontier-AI operator and speculative writer: his fiction is used to think through AI-controlled production systems, moral responsibility inside an AI lab, and the possibility that capitalism needs redesign when machine output becomes much larger than human labor output.
Clark’s most concrete forecast in the source is that by April 2027, AI systems may be able to complete tasks that would take a person roughly 150 hours. He connects that to software, research, circuitry, and source synthesis, then extends the discussion into AI Automation Redistribution, AI Cyber-Defense Utility, and education built around preserving curiosity.
Key Claims
- AI progress should be evaluated as a distribution problem, not only a productivity or software-market problem.
- Taxing robots or AI companies could become necessary if a machine economy generates large returns that do not flow naturally to the human economy.
- Cyber-capable Claude systems should be used defensively, but their social value may require utility-like access and incentives.
- Education may need to teach people to remain curious when AI can answer routine questions quickly.
Connections
- Anthropic - company Clark co-founded and represents in the live interview.
- Claude - model family discussed as increasingly capable at long tasks and cyber work.
- AI Automation Redistribution - Clark’s robot-tax and machine-economy redistribution frame.
- AI Cyber-Defense Utility - Clark’s argument that some AI security capability should be broadly available.
- AI Economic Diffusion, Digital Employees, and AI Backlash Politics - adjacent wiki branches affected by Clark’s forecast.