AI’s third era: the rise of persistent AI coworkers | Tara Seshan (OpenAI’s product lead)
Summary
Tara Sa, OpenAI’s product lead for Codex and ChatGPT, discusses the shift from theoretical planning to empirical, rapid testing in AI product management. She emphasizes that PMs must act as "steers" guiding agents, fostering ambition, and building for model capabilities 2-3 months ahead. The episode covers the evolution from chat to persistent AI coworkers, the importance of tight feedback loops, and how OpenAI’s founder-led culture drives fast iteration and user-centric innovation.
Topics discussed
Introduction: The 3 eras of AI products and OpenAI's culture
Working at OpenAI: Speed, market proximity, and lack of secrecy
Sponsor: WorkOS for enterprise-ready apps
Product Management in AI: Empirical testing and the 'Eigen Question'
The Future of Work: Steering agents vs. rowing tasks
Persistent AI Co-workers and the need for system access
Elevating Ambition: Using AI to expand personal capabilities
Building for the Future: Avoiding the 'too early' trap
ChatGPT vs. Codex: Merging interfaces for seamless work
Balancing Innovation: Managing a billion-user product
Sponsor: Mercury Spend for expense management
Team Dynamics: Blurring roles and focusing on outcomes
Human Value: Creativity, expression, and team dynamics
Codex Sites and Visualizations: Rapid prototyping tools
Writing as Thinking: Changing how we document ideas
Venture Capital Lessons: Product-market fit and marketing
Knowledge Work: Showing the AI's reasoning process
Recommendations: Anna Karenina, The Power Broker, and Oppenheimer
Personal Projects: Gats, Thiel Fellowship, and final thoughts
Listen ad-free on Castria