AI in drug discovery: a 10-year arc, not 2
David Ricks estimates meaningful AI-driven acceleration of drug discovery is roughly a decade away, not the 2–3 years many tech boosters claim.
Snapshot · The Prof G Pod with Scott Galloway
Why People Are Losing Faith in Healthcare
Explore episode
Jun 4, 2026
Where this was said
AI in Drug Discovery: Overhyped for Now, Transformative in a Decade
At 41:30 · chapter starts 36:55
Jensen Huang's vision of AI as the engine of a 'great age of discovery' in biotech gets a reality check from the CEO of the world's most valuable pharma company. Ricks is not dismissive — Lilly scientists use OpenAI and Anthropic tools daily and he's working with NVIDIA to build better biological datasets. But he draws a sharp distinction: LLMs excel at synthesizing the totality of human written knowledge, and they're good at it because that knowledge base is nearly complete. Biology isn't. Humans have mapped perhaps 10–20% of how the human body actually works, which means training an AI on it produces models that are 'terrible at predicting' outcomes in the remaining 80%. The second limitation: drug discovery ultimately requires physical experimentation. Tirzepatide was the 700th formulation tested — and the 20 of roughly 1,000 preclinical steps that Lilly has automated with AI or machine learning still leave 980 to go. The human experimentation phase — clinical trials — can't be shortcut by AI for the foreseeable future. His estimate: five to seven years before AI meaningfully reduces laboratory experiments, and a full ten-year arc before it genuinely accelerates drug discovery at scale.
Why AI in Drug Discovery Is Overhyped — For Now
AI excels at synthesizing existing knowledge, but drug discovery requires creating new knowledge. Humans have mapped only 10–20% of human biology, so AI models trained on that incomplete dataset produce poor predictions. Meaningful AI acceleration in pharma R&D is a 10-year arc, not a 2-year one.
Only 10–20% of human biology is understood
David Ricks estimates that humans have discovered only 10–20% of all knowledge about the human body system, making AI-driven drug discovery extremely difficult today.
Tirzepatide was the 700th GLP-1 variant tested
Tirzepatide, the active ingredient in Mounjaro and Zepbound, was physically the 700th version of a GLP-1 inhibitor Lilly formulated and tested.
Similar snapshots
Ad monetization for games
Ad-based monetization works well for game apps where users spend extended time in-session, as seen with Grid and Wordle.
Ads don't work for tool apps
Tool-focused apps like PuffCount are poor candidates for ad monetization because users don't stay in-session long enough.
Hard paywall is unskippable
A hard paywall is a screen that blocks all app features unless the user pays or starts a free trial — it cannot be dismissed.
Two core monetization models
Mobile apps are primarily monetized through either ads (best for games) or in-app purchases/subscriptions (best for tools).
YouTube beats every other platform
According to the episode, YouTube outperforms every other social platform for building trust and driving SaaS conversions.
Most users came from YouTube
Vasco stated that the majority of his app's user base came directly from his YouTube channel.
Boost My Domain Rating button
SEO Bot features a 'Boost My Domain Rating' button that routes users directly to Listing Bot, an example of in-product cross-selling.
All products connected to AI
The founder's entire product portfolio is AI-related, making it easier to package products attractively for directories.
AI coding viability debate as hook
The founder attached their SaaS demo to the trending debate about whether AI coding is actually good enough to build a full SaaS product.
