The Genetics Podcast
The Genetics Podcast
Sano Genetics
EP 264: What AI’s math breakthroughs mean for drug development and genetics
41 minutes Posted Nov 5, 2026 at 12:00 pm.
Intro to The Genetics Podcast00:58 What the OpenAI math breakthroughs could mean for biology02:57 AI’s recent frontier math results and how mathematicians reacted09:12 Lean as a proof checker and why it matters for math10:44 Four conditions behind AI’s math progress and how biology compares16:22 Why verification speed decides where AI can accelerate biology19:35 How lab automation shortens feedback loops for variants and antibodies23:13 How validated biomarkers let early trials predict clinical outcomes26:11 Why validated surrogate endpoints could speed up drug development31:58 How startups and big pharma are building self-driving labs33:40 Virtual cell models built from perturbation data34:50 Genetics-driven drug discovery and the Phase II success rate question37:42 Verification speed and surrogate endpoints as biology’s biggest gaps40:38 Closing remarks
0:00
41:31
Download MP3
Show notes
This week on The Genetics Podcast, Patrick goes solo to look at what recent AI breakthroughs in mathematics could mean for biology. He discusses the OpenAI results on the Navier-Stokes equation and hundreds of open math problems, what makes math and code so easy for AI to improve on, how self-driving labs and surrogate endpoints could speed up drug development, and where the biggest gaps and opportunities remain.Show Notes0:00 Intro to The Genetics Podcast00:58 What the OpenAI math breakthroughs could mean for biology02:57 AI’s recent frontier math results and how mathematicians reacted09:12 Lean as a proof checker and why it matters for math10:44 Four conditions behind AI’s math progress and how biology compares16:22 Why verification speed decides where AI can accelerate biology19:35 How lab automation shortens feedback loops for variants and antibodies23:13 How validated biomarkers let early trials predict clinical outcomes26:11 Why validated surrogate endpoints could speed up drug development31:58 How startups and big pharma are building self-driving labs33:40 Virtual cell models built from perturbation data34:50 Genetics-driven drug discovery and the Phase II success rate question37:42 Verification speed and surrogate endpoints as biology’s biggest gaps40:38 Closing remarks