
#24 [Guest Appearance] Can AI Make Biomedical Research 10x Faster? (Stephan on the Tech & Drugs Podcast)
Notes de l’épisode
Hey Fit For Science listeners! Today we are sharing Stephan’s guest appearance on the Tech & Drugs podcast with host Dr. Thibault Geoui. Since our conversation explores the intersection of AI for science, wet-lab bottlenecks, and the future of biomedical discovery, we wanted to share the full episode directly with you here. If you enjoy conversations at the intersection of artificial intelligence, biotech, and drug development, be sure to check out and subscribe to his podcast on YouTube, Spotify or Apple Podcasts!
📝Summary
In this special guest appearance on the Tech & Drugs podcast, Stephan joins host Dr. Thibault Geoui to explore the promise and practical limits of accelerating biomedical research with AI. We trace Stephan's non-linear path from technical mathematics and biomedical engineering at TU Wien to bioinformatics at Roche and multi-omics PhD research at CeMM and the Medical University of Vienna. Unpacking Stephan’s recent study, they discuss why irreducible wet-lab processes like cell growth remain global bottlenecks even when cognitive tasks like literature synthesis and coding are massively accelerated. They examine how researchers can bypass physical constraints using predictive biomarkers, public datasets, and tools like AlphaFold or Stephan's MrBiomics project. We discuss the connection between structured chain-of-thought prompting in our SimulateGPT work with the current test-time inference paradigm. Finally, the conversation covers cultural resistance to AI in pharma, Big Tech’s push into neglected diseases, and how scientific alignment will guide the agent-augmented labs of the future.
⏳Chapters
00:00:00 Teaser: Bottlenecks, biomarkers, and guardians of science
00:02:20 Introduction: Welcome to Tech & Drugs and science communication
00:05:00 Stephan’s Journey: From math and coding to bioinformatics
00:09:11 The Hybrid Scientist: Merging wet-lab and dry-lab training
00:12:06 Math vs. Biology: Integrating engineering with biomedicine
00:13:30 Limits of 10x Acceleration: Cognitive vs. physical bottlenecks
00:18:40 Challenging Irreducibility: Visual AI, biomarkers, and public data
00:24:05 From Crystallization to AlphaFold: A 1,000x leap in structural biology
00:29:00 Invisible AI vs. Co-Scientists: Automating research workflows
00:34:28 SimulateGPT & Software 2.0: Chain-of-thought reasoning in biology
00:39:04 Cultural Resistance in Pharma: Adoption hurdles and opportunity costs
00:46:06 Institutional Agility: Small teams and CeMM’s Data Governor
00:48:52 Big Tech Enters Biotech: Anthropic, Alphabet, and Insilico Medicine
00:56:30 Scientific Alignment: Goodhart’s law and high-impact priorities
01:04:05 The Lab of 2035: Agent-augmented researchers and flat hierarchies
01:10:18 Closing Initiatives: Accelerate Europe and Fit For Science podcast
📚Resources
Potential and limitations of AI in biomedical research | MedUni Vienna
GPT-4 as a biomedical simulator (SimulateGPT)
"ChatGPT" for biomedical simulations | MedUni Vienna
MrBiomics: Composable modules and recipes automate bioinformatics for multi-omics analyses | GitHub
…There is more: complete show notes here
🎙️About
Fit For Science is a deep-dive podcast hosted by two biological data scientists, Rob and Stephan, exploring the intersection of research, health tech, and data-driven lifestyle design. The hosts provide evidence-based systems, layered with practical "N=2" personal experimentation, to cut through the noise and enable everyone to become their best N-of-1.
Learn more and subscribe on your favorite platforms:
⚠️Disclaimer: This podcast represents our own opinions and is for informational purposes only. It does not constitute medical or financial advice or a professional relationship.