Podcast Summary:
a16z Podcast — “Mark Zuckerberg & Priscilla Chan: How AI Will Cure All Disease”
Date: November 6, 2025
Episode Overview
In this a16z Podcast episode, hosts at Andreessen Horowitz sit down with Mark Zuckerberg and Priscilla Chan to explore the ambitious vision of the Chan Zuckerberg Initiative (CZI): using artificial intelligence and large-scale collaborative science to cure, prevent, or manage all disease by the end of this century. The conversation traces their 10-year journey, the evolution of CZI’s work, the role of open-source data sets and virtual cell models, and the formation of their new centralized Biohub, which aims to pair frontier biology with state-of-the-art AI. The discussion highlights the transformative potential of integrating engineering, biology, and computation, as well as practical lessons learned about scientific tools, collaborative infrastructure, and risk-taking for global health breakthroughs.
Key Discussion Points & Insights
1. Why the Mission to Cure All Disease?
- Priscilla Chan shares her motivation from her time as a pediatrician, seeing patients with unsolvable illnesses and recognizing the limitations of medical knowledge.
- “I wanted to improve people’s lives... I met a lot of patients... for which we just had no idea what the problem was.” (02:20, Chan)
- The couple set a bold goal to cure or prevent disease by 2100, initially met with skepticism from scientists.
- "Most scientists couldn't look at us with a straight face . . . there was no pathway to that being true." (02:23, Chan)
2. The Role of Tools in Scientific Discovery
- Mark Zuckerberg emphasizes that transformative scientific progress is typically enabled by new tools (e.g., the microscope for biology, telescopes for astronomy)—not solely by more grant funding.
- “Most major breakthroughs are basically preceded by the invention of a new tool to observe phenomena in a new way.” (03:31, Zuckerberg)
- CZI has focused efforts on building and open-sourcing such tools and infrastructure for biology, such as data standards and annotation software.
3. The “Cell Atlas” and Network Effects in Biotech
- Their “Cell by Gene” project and the Cell Atlas began as an annotation tool to help standardize and share single-cell data, which unexpectedly became an industry standard and expanded via community “network effects.”
- “Cell by Gene is like almost an accident though.” (15:25, Chan)
- “Now after 10 years, there are millions of cells that have been built into this shared resource . . . 75% came from the broader community saying this is useful and there’s an easy way for us to standardize.” (16:18, Chan)
- Open-source data and software greatly accelerate research, enabling thousands of scientists and startups.
4. Pairing Frontier Biology with Frontier AI
- CZI’s current focus is on integrating advanced AI with top-tier biological research—where they see unique, untapped value.
- “So far there hasn't been anyone who's tried to do both of those at once . . . you can produce specific data sets for the purpose of training AI models to build virtual cells.” (07:28, Zuckerberg)
- New models (like LLMs for biology) will allow for not just pattern recognition, but scientific reasoning and hypothesis generation within biology data.
5. Biohub Vision & Organizational Model
- CZI’s Biohubs in SF, Chicago, and NY each tackle different biological frontiers (deep imaging, tissue engineering, cell engineering). They intentionally co-locate cross-disciplinary scientists to foster unexpected breakthroughs.
- “The locations are not by accident . . . we sort of choose the grand challenge and locations based on partner universities.” (09:50, Chan)
- Unifying these efforts under Alex Reeves as a centralized "operating philanthropy" is their latest step for efficiency, speed, and impact.
6. Virtual Cell Models: Risks, Promise & Practice
- AI-powered “virtual cells” allow scientists to simulate and test biological hypotheses in silico (on computers) before committing costly time/resources to wet lab work.
- “If you had a virtual cell model where you could simulate really high quality biology, you could actually then start testing and tinkering . . . and ask riskier questions.” (21:41, Chan)
- The goal isn’t perfect fidelity, but useful insight.
- “All models are wrong, some are useful.” (23:57, VC Host)
- “I don’t think it needs to be 100% accurate to be useful . . . The more you de-risk, the more efficient it gets, obviously.” (23:16, Chan)
- Such models could become biology’s equivalent of “model organisms” (e.g., fruit flies in genetics)—but scalable and personalized.
7. Open Data, Compute as Lab Space, and Collaboration
- CZI is investing heavily in scalable compute infrastructure (thousands to tens of thousands of GPUs) and sharing access with researchers to empower large-scale AI for science.
- “We were the first to really build a large scale compute cluster, a thousand. Now we have plans to move to the 10,000 range...” (39:19, Chan)
- They purposely lower barriers to entry (user interfaces, open-source tools) so diverse fields and backgrounds can contribute.
- “The user interface was intentionally designed to not need to have a computational or really a deep biological background . . . that’s intentional.” (33:34, Chan)
8. Design Principles & Philosophical Takeaways
- Double-down on approaches that deliver impact: feedback from community use of tools and unexpected positive results guide where to invest further.
- “Not only did we achieve what we said we were going to do . . . it actually delivered more than we thought . . . we can really continue doubling down.” (41:07, Chan)
- Patient with long timelines but impatient for iteration (“be willing to have a long time horizon but be impatient at the same time”).
- Organizing interdisciplinary scientists in close proximity is a powerful unlock—though decentralized efforts have value too.
Notable Quotes & Memorable Moments
-
On Ambition and Skepticism:
- “[Most] scientists couldn't look at us with a straight face.” (02:23, Chan)
- "The biology folks... looked at it as if it were crazy ambitious. And then the AI folks are like, 'Well, that's kind of boring. That's just automatically going to happen.' There's something in between..." (00:45 & 06:47, Zuckerberg)
-
On Tools & Impact:
- “Most major breakthroughs are basically preceded by the invention of a new tool to observe phenomena in a new way.” (03:31, Zuckerberg)
- “The reason to do it together is then we can actually complete the flywheel . . . these people need to be sort of working shoulder to shoulder and shaping each other's work.” (31:29, Chan)
-
On the Cell Atlas Becoming the Standard:
- “It’s kind of this crazy thing that we’re, you know, here in, you know, 2025, and there’s not the kind of periodic table of elements equivalent for biology.” (00:00 & 15:47, Zuckerberg)
-
On Open Collaboration:
- “We’ve only funded 25% of it. 75% came from the broader community saying this is useful . . .” (16:18, Chan)
- “It's like an interesting, what you'd call a network effect, right?” (18:00, Zuckerberg)
-
On Virtual Biology:
- “If you had a virtual cell model . . . you could actually then start testing at and tinkering on the computational side and like ask riskier questions...” (21:41, Chan)
- “All models are wrong, some are useful.” (23:57, VC Host)
- “You just want to be able to de-risk the idea on the front end a little bit.” (23:16, Chan)
-
On Organizational Design:
- “You can fix so many organizational questions . . . just by having two teams sit together.” (36:50, Zuckerberg)
-
On Compute as Lab Space:
- “We are not expanding, like, a lot of square footage per se, but we're expanding our computer... in a sense, that's new lab space. It's much more expensive than wet lab space.” (38:48, Chan)
- “The researchers, they don't want employees working for them. They don't want space. They just want GPUs.” (38:48, Zuckerberg)
Timestamps for Important Segments
- [00:00–04:13] — Origins of CZI and philosophy of tooling in science
- [09:31–11:29] — 10–15 year horizon for grand scientific challenges & Biohub structure
- [12:06–15:25] — Open data, impact on rare/common diseases, Cell Atlas
- [16:18–18:31] — Community adoption and network effects for cell atlasing
- [18:31–26:14] — The vision and current state of virtual cell models, simulation and reasoning
- [30:15–34:50] — Centralizing teams, importance of user interface, interdisciplinary problem-solving
- [39:19–40:18] — Investing in compute, enabling outside scientific collaborations
- [41:07–42:20] — Learning from feedback, persevering through ambiguity
Conclusion
Mark Zuckerberg and Priscilla Chan share how their vision for CZI—and especially now the Biohub—evolved from a bold, (once-)laughable ambition to a highly collaborative, tool-centric engine for accelerating biomedical science. Bridging biology and AI, they focus on democratizing data, building community infrastructure, and catalyzing both basic research and translational breakthroughs. Their commitment is to adapt and double down on high-impact approaches, learning from practical feedback, and seeding a new era where precision medicine and rapid discovery become not just possible, but inevitable.