Ask a Cycling Coach Podcast - Presented by TrainerRoad

TrainerRoad AI Q&A with CEO Nate Pearson | Ask a Cycling Coach Podcast 569

34 snips
Jan 22, 2026
Guest
Nate Pearson
Nate Pearson, CEO of TrainerRoad and a driving force behind their innovative AI systems, delves into the cutting-edge use of machine learning in cycling training. He explains how TrainerRoad's custom AI predicts workouts and FTP, emphasizing the importance of accuracy over traditional benchmarks. Nate debunks common misconceptions about FTP testing, discusses why more training doesn't always lead to better results, and highlights the dynamic nature of AI-adjusted plans that evolve in real-time based on user performance.
Ask episode
AI Snips
Chapters
Books
Transcript
Episode notes
INSIGHT

Purpose-Built ML Beats Chatbot Approaches

  • TrainerRoad built a bespoke ML model trained on tens of millions of workouts and hundreds of millions of rides to predict athlete performance and select workouts.
  • This model is purpose-built for training and is not a general LLM or chatbot and avoids “AI slop.”
INSIGHT

AI Uses Raw Metrics, Not FTP As A Concept

  • TrainerRoad AI doesn't 'know' FTP conceptually; it models watts, duration, and heart rate to predict performance and pick workouts.
  • The system back-calculates an FTP-like benchmark from predicted performance to place you into progression levels.
INSIGHT

FTP Is A Functional, Not Absolute, Benchmark

  • FTP is a functional benchmark, not a precise physiological constant, and people misattribute exact meanings to it.
  • TrainerRoad treats FTP as a consistent, precise scale to calibrate workouts rather than a perfect hourly maximum.
Get the Snipd Podcast app to discover more snips from this episode
Get the app