Click here to watch a video of this episode.
Drug development punishes optimism, and predictivity is one of the few defenses we have against translational risk. In this episode we connect translational science and regulatory science to a simple operating idea: treat attrition as data, not embarrassment. The tension is real. Do you want a model that explains the past beautifully, or one that helps you be less wrong before the next big spend? We unpack why “validation” after the fact is not enough, and how forecast accuracy, calibration, and honest miss rates should travel across programs instead of dying in slide decks. When predictivity is measured, learning becomes infrastructure. When it isn’t, failure stays expensive and repetitive. Takeaway: start tracking misses across your portfolio and make calibration a standing agenda item, not a post-mortem ritual.

• (00:00) - Attrition Framing and Episode Purpose
• (01:00) - Shots on Goal and Fast Fail
• (03:00) - Learning from Patients and Preclinical Gaps
• (04:00) - Opening the Black Box of Attrition
• (06:00) - Learning Fast and Digital Iteration
• (09:00) - Intentional Learning and Structured Registries
• (11:00) - Preventative Technology and Early Detection
• (16:00) - Disease Continuum and Early Modulation
• (20:00) - Chief Learning Officer and Incentive Shift
• (23:00) - Redefining Disease and Closing Reflections