Hard Drugs Will AI solve medicine?
Oct 29, 2025
Artificial intelligence is reshaping drug discovery, but it faces significant challenges. The hosts dive into the unpredictability of biology and the limits of AI in advancing medicines. They discuss the potential of drug repurposing and the need for better data on rare diseases. The complexity of human clinical trials and ethical dilemmas surrounding them are explored. Manufacturing hurdles and the necessity for skilled professionals in healthcare delivery are highlighted, along with the importance of ambition in tackling health issues.
AI Snips
Chapters
Books
Transcript
Episode notes
Progress Is Narrow, Biology Is Vast
- AI progress looks dramatic in narrow problems but biology remains far more complex than early successes imply.
- Solving protein folding didn't mean we can predict whole cells, organs, or organism behavior yet.
Discovery Often Preceded Understanding
- Edward Jenner and Tu Youyou found effective treatments without a full mechanistic understanding by observing and testing widely.
- Repurposing a shelved drug led to a modern schizophrenia treatment after combining it with trospium to reduce gut side effects.
Data Gaps Hide Real Causes
- Better human sequencing and exposure data would unlock many discoveries for rare and environmental diseases.
- Case investigations (like the Swiss ALS cluster linked to wild mushrooms) reveal causes absent from standard datasets.
