What does it look like to do economics from inside the AI revolution with access to data nobody else has, and questions nobody has figured out how to answer yet?

In this episode, Ben sits down with Ronnie Chatterji, chief economist at OpenAI, to pull back the curtain on how economic research actually gets done at a frontier AI lab.

Topics covered:

• Why the right time horizon for AI economics is a mix of real-time, medium-term, and scenario planning
• Why you can't afford to be late to recursive self-improvement even when it feels far away
• What OpenAI's own labor data is showing
• The four-category job framework: which roles are at risk, which will be reorganized, which will expand as prices fall, and which are largely insulated
• The Codex paper: what OpenAI's internal data on agentic coding tool a ...