scripod.com

Mark Zuckerberg & Priscilla Chan: How AI Will Cure All Disease

Priscilla Chan and Mark Zuckerberg join a16z leaders to discuss a bold scientific vision: leveraging AI and long-term investment to transform how we understand and treat disease. Their conversation centers on reimagining the infrastructure of biological research, not through incremental improvements, but by building foundational tools that redefine what’s possible in medicine.
The Chan Zuckerberg Initiative (CZI) is accelerating the fight against all diseases by investing in large-scale, long-term scientific tools that traditional funding bodies like the NIH cannot support. Central to this effort is the Biohub, which integrates frontier biology with AI to build virtual cell models—digital simulations that allow scientists to test hypotheses in silico before costly lab work. A key milestone is the Cell Atlas, an open-source catalog of millions of cells that serves as biology’s missing 'periodic table,' enabling researchers to map cellular behavior with unprecedented precision. By unifying CZI and Biohub under AI leadership, the initiative fosters deep collaboration between biologists and engineers, prioritizing compute over physical labs. Virtual models reduce risk in drug discovery by simulating rare cell states and causal mechanisms, while user-friendly interfaces democratize access. Success hinges on cross-functional teams, proprietary datasets, and domain-specific AI models that outperform general ones. The goal is to enable personalized therapies by treating most diseases as rare conditions shaped by individual biology, ultimately compressing the timeline from basic science to real-world treatments.
05:56
05:56
AI is both overestimated and underestimated, similar to the early internet
07:21
07:21
Few have combined frontier biology with frontier AI, creating an opportunity for transformative science.
09:05
09:05
The 10-15 year horizon is concrete and achievable for solving grand challenges
09:43
09:43
Integration of LLMs and AI has become an interesting development for data use
11:24
11:24
Large language models are prompting new questions about therapeutic outcomes.
11:35
11:35
Most diseases should be thought of as rare due to individual biological differences
13:38
13:38
Biology lacks a 'periodic table of elements' equivalent in 2025
15:47
15:47
75% of the Cell Atlas data comes from the broader scientific community.
19:31
19:31
Researchers from Meta join the Biohub to advance protein folding models
21:18
21:18
Virtual cell models allow scientists to test riskier drug development ideas with lower costs and higher efficiency.
21:56
21:56
Even imperfect virtual cell models can de-risk experimental hypotheses
28:19
28:19
CZI can uniquely tie together Frontier AI and Frontier biology
30:32
30:32
Closing the loop between data generation and AI modeling drives breakthroughs in cell biology
33:37
33:37
User interfaces are designed to allow broad participation in scientific problem-solving
38:23
38:23
The Biohub has enabled academic labs to access large-scale compute resources they couldn't otherwise afford.
40:41
40:41
AI can bring huge leverage to the pipeline from basic science to novel therapies and public health.