Podcast Episode Details
Back to Podcast Episodes
#30 Soon you'll take a drug designed by AI | Alex Zhavoronkov
Episode 30
As you may have heard, AI-designed medicines have crossed a historic line. In this episode, Alex Zhavoronkov - CEO of Insilico Medicine and founder of ARDD walks us through how Insilico’s rentosertib became the first AI-generated small molecule with peer-reviewed clinical efficacy, while arguing against AI hype and reminding us that biology still moves at “the speed of traffic.” That duality runs through the whole conversation. On one side: a pragmatic operator obsessed with credible science, biomarkers, and clinical benchmarks; on the other: an AI visionary investing in cryonics, sketching “pharmaceutical superintelligence,” and thinking in decades, not quarters.
We start in Basel, home to Roche and Novartis, where ARDD was born, then trace how the conference morphed into a ”high-signal filter for longevity” - packed with startups (who also fund it), hard data, and mainstream pharma.
Alex looks back at his 2014 Nvidia talk (”Can Nvidia solve aging?”) and explains why Insilico trains its AI to learn age first - so it actually grasps biology. Years of problem-solving with pharma turned into their Pharma.AI toolkit (Biology42, Chemistry42, Medicine42, Science42).
Insilico now runs 40+ programs and in an early Phase 2 study for idiopathic pulmonary fibrosis (IPF), their drug rentosertib showed a dose-dependent boost in lung capacity.
Compared with the old path - often $150–200M and ~5 years just to pick a lead molecule - Insilico says it can often reach that point for under $3M or even less. Still, Alex is cautious: no matter how smart the AI gets, real-world testing and regulation won’t speed up overnight.
Also in this episode:
What made Alex cry.
Why he wouldn’t give his own drug to patients - yet.
How a mirror on a conference poster led to a proposal.
How ARDD became the “WEF of longevity”.
Why internal “kill teams” try to stop their own drug candidates.
Why labeling aging a disease helps - but won’t shortcut approvals.
Why he writes to “feed AI”.
How Nvidia threads through the story - from free GPUs to Jensen’s video.
🚀 Special offer for our LEVITY audience: Join Vitalism today and receive a 30% discount on your membership using the code LEVITY at checkout. https://www.vitalism.io/membership
Show notes for this episode will be available soon after this airs. Sign up for the LEVITY newsletter to get them straight to your inbox: reachlevity.com
LEVITY is co-hosted by Patrick Linden, philosopher and author, and Peter Ottsjö, journalist and author.
CHAPTERS
00:00 – Teaser
03:08 – Introduction to Alex Zhavoronkov
06:11 – Alex talks ARDD
15:15 – Big-pharma starting to embrace ARDD
17:31 – The proposal story
24:52 – Why Alex decided to fight aging
27:44 – Neuralink, humanoids and the brain-aging bottleneck
30:52 – Keeping ARDD pharma-credible
32:02 – The path to Insilico
48:03 – The Zhavoronkov crystal ball
57:29 – The Insilico platform
1:07:16 – The rentosertib story
1:16:42 – What made Alex cry
1:17:44 – Aging-as-disease: rhetoric vs. regulation (GLP-1 analogy)
1:26:53 – Culture check: Middle East momentum, China’s stance
1:34:28 – Costs & timelines: what AI compresses - and what it can’t
1:41:30 – Insilico’s fully automated lab
1:47:21 – “I respect Demis, but…”
1:51:10 – Why even superintelligence won’t skip clinical validation
1:52:15 – Cryonics as plan B: organ preservation, TimeShift, use-cases
2:00:56 – Writing Forever AI & the roadmap to “pharma superintelligence”
2:06:42 – Book recommendations
Hosted on Acast. See acast.com/privacy for more information.
Published on 9 hours ago