Could AI Transform How Science Gets Done? with Jungwon Byun | Dream Machines
Summary
In this episode of Dream Machines, hosts Alexis Madrigal and Robin Sloan interview Jungwon Byun, co-founder of the AI research tool Elicit. Byun discusses her journey from New York to the Bay Area startup scene and how Elicit helps scientists navigate complex literature with rigorous, cited sources to combat hallucinations. The conversation explores the potential for AI to accelerate drug discovery, the challenges of physical-world friction, and the emotional toll of leading a mission-driven company amidst rapid industry consolidation.
Topics discussed
Sponsors and intro to Dream Machines podcast
Introduction to Elicit and guest Jungwon Byun
Jungwon's move from New York to San Francisco
Early AI experiences and the 'North Beach moment'
Elicit's mission: AI for rigorous scientific research
Case study: Mapping oncology drugs for pharma
Combating hallucinations and ensuring citation accuracy
The future of AI in science and 'powerful search'
Autonomous science and physical world limitations
Defining 'elicitation' and model consistency
Social anxiety, job displacement, and dystopia vs utopia
Founder psychology and the Oakland vs SF dynamic
Talent competition and recent 'holy shit' moments
Credits, production team, and final sponsors
Listen ad-free on Castria