Weekend Listen: Anthropic's Co-Founder and Top Economist on Doing Research at the AI Frontier
Summary
Jill Weisenthal and Tracy Alloway interview Anthropic co-founder Jack Clark and economist Peter McCrory on the rapid evolution of AI. They discuss "recursive self-improvement," noting that engineers now write eight times more code, and analyze why macroeconomic impacts remain muted despite productivity gains. The conversation covers the "bitter lesson" of scaling compute, the shift toward hiring senior experts for direction-setting, and the need for transparent, third-party AI regulation.
Topics discussed
Intro and AI hype vs reality
Guests: Jack Clark and Peter McCrory
AI progress and the 'bitter lesson'
Why AI impact feels slow in the economy
AI in research and social science
AI coding, automation, and IBM case study
Sponsors: Public and Chase Sapphire
AI regulation, safety, and testing
Measuring AI's impact on productivity
Anthropic Institute and public advocacy
Sponsors: LPL Financial and Public
AI hiring trends and labor market effects
AI adoption in startups vs large firms
AI alignment, safety, and extinction risk
Model behavior and anthropomorphism
The race for AGI and safety incentives
Outro and final thoughts on AI future
Closing ads and credits
Listen ad-free on Castria