The Future of Everything Best of: The future of AI coaching
10 snips
Dec 26, 2025 In a captivating discussion, James Landay, a computer science professor at Stanford known for his work on human-centered AI, dives into how large language models can revolutionize coaching in fitness and education. He explores the intriguing concept of motivational interviewing within AI, the integration of wearable tech for tailored recommendations, and the importance of long-term interactions. Landay also touches on the balance of social features in coaching, the innovative use of mobile AR for learning, and the philosophy of human-centered AI to maximize positive impact.
AI Snips
Chapters
Books
Transcript
Episode notes
LLMs Can Act As Conversational Coaches
- Large language models can simulate a conversational coach to address qualitative barriers that quantitative fitness apps miss.
- James Landay's GPT Coach uses motivational interviewing and prior sensor data to create personalized fitness plans.
Layer Agents And Expert Guidance
- Combine multiple model agents and domain guidance to keep an LLM-based coach on track and faithful to protocol.
- Use expert manuals and validation checks rather than vanilla out-of-the-box LLM behavior when building coaching systems.
Embed Expert Protocols And Run Staged Studies
- Involve domain experts and encode their training manuals into the system to match professional coaching styles.
- Run short trials to refine interfaces then conduct longer multi-month studies to measure behavior change.
