Mark Zuckerberg & Priscilla Chan: How AI Will Help Cure Disease...
a16z PodcastFull Title
Mark Zuckerberg & Priscilla Chan: How AI Will Help Cure Disease
Summary
Mark Zuckerberg and Priscilla Chan discuss their Chan Zuckerberg Initiative's ambitious goal to cure and prevent disease by leveraging AI and building foundational tools for biology.
They highlight the creation of biohubs, advanced data sets, and virtual cell models as key strategies to accelerate scientific discovery and enable precision medicine.
Key Points
- The Chan Zuckerberg Initiative's (CZI) mission to cure and prevent all disease by the end of the century initially seemed overly ambitious to many, including scientists.
- Their strategy focuses on accelerating basic science by building foundational tools for biologists, rather than directly funding individual therapies.
- Historically, major scientific breakthroughs are often preceded by the invention of new observational tools, a principle CZI is applying to biology.
- Traditional government funding models (e.g., NIH grants) often favor smaller, near-term projects, leaving a gap for developing large-scale, long-term tools like AI-driven virtual cell models.
- CZI views the intersection of frontier biology and frontier AI as a critical area, noting that few organizations are actively pursuing both simultaneously.
- They are developing comprehensive cell atlases and virtual cell models to provide researchers with the infrastructure and data necessary for advanced biological understanding and discovery.
- The development of CZI's tools, such as the Cell By Gene annotation tool, has led to a significant expansion of the cell atlas through community contributions, demonstrating a powerful network effect.
- Virtual cell models are seen as a crucial tool for hypothesis generation, de-risking scientific research, and enabling more efficient drug discovery and personalized medicine.
- CZI is consolidating its efforts into a unified Biohub, an operating philanthropy focused on advancing biology at the intersection of AI, led by Alex Reeves.
- The focus on building domain-specific AI models, combined with open-source data access, is seen as a strategy to accelerate progress in biology, mirroring successes in other AI applications.
- The importance of intuitive user interfaces and lowering the barrier to entry for scientists is emphasized to foster true interdisciplinary collaboration.
- CZI views its role in the scientific ecosystem as unique, filling a void in building foundational tools and data resources that empower the broader scientific community.
Conclusion
AI offers immense leverage in accelerating biological discovery and the fight against disease.
Building foundational tools, robust data sets, and accessible platforms is crucial for democratizing scientific progress.
The Chan Zuckerberg Biohub is uniquely positioned to bridge the gap between AI and biology, fostering collaboration and enabling breakthroughs for a healthier future.
Discussion Topics
- How can we better bridge the gap between the rapid advancements in AI and the foundational needs of biological research?
- What are the most critical "tools" or "infrastructure" needed in biology to accelerate discovery at the same pace as other scientific fields?
- How can organizations like CZI Biohub foster true interdisciplinary collaboration to tackle complex diseases and accelerate medical breakthroughs?
Key Terms
- Cell Atlas
- A comprehensive map or dataset detailing the types, functions, and relationships of cells in a biological system.
- Virtual Cell Models
- Computational simulations that replicate the behavior and functions of biological cells, used for research and experimentation.
- Biohub
- A research center or institute focused on advancing specific areas of biomedical science, often through interdisciplinary collaboration.
- Transcriptomics
- The study of the complete set of RNA transcripts produced by an organism or cell under specific conditions.
- Cryo-EM (Cryo-electron microscopy)
- A technique used to determine the atomic and molecular structure of biological molecules and their complexes.
- Large Language Models (LLMs)
- Advanced AI models trained on vast amounts of text data, capable of understanding, generating, and manipulating human language.
- CRISPR
- A family of DNA editing technologies that enables scientists to make precise changes to the genome of living organisms.
- RFA (Request for Applications)
- A document issued by a funding agency to solicit proposals for research projects.
- NIH (National Institutes of Health)
- A primary agency of the U.S. government responsible for biomedical and public health research.
- Idiopathic Pulmonary Fibrosis (IPF)
- A lung disease characterized by progressive scarring of lung tissue, with an unknown cause.
- Variant of Unknown Significance (VUS)
- A genetic variant that is not yet clearly linked to a specific disease or health condition.
- Fertilitiy
- The ability to conceive children. (Note: This term appeared in the transcript in a way that seemed like a mishearing or unrelated tangent, and has been omitted as it's not pivotal to the key points).
- GPUs (Graphics Processing Units)
- Specialized processors designed for parallel processing, commonly used in AI and machine learning for training complex models.
Timeline
CZI's mission to cure and prevent all disease by the end of the century was initially met with skepticism.
The strategy to cure disease is to empower scientists by accelerating the pace of basic science, often through new tools.
CZI aims to accelerate scientific progress by building tools that allow researchers to observe phenomena in new ways.
CZI's philanthropic approach focuses on empowering scientists and entrepreneurs rather than seeking direct credit.
The initiative faced initial disbelief from biologists and a lack of urgency from AI experts, highlighting the need to bridge this gap.
CZI's work at Biohub is described as frontier biology paired with frontier AI, aiming to connect leading-edge research in both fields.
The development of AI models like AlphaFold highlights the importance of data sets, but CZI aims to create specific data sets for training AI models.
Over 10 years, CZI has found that science research has yielded the biggest returns, leading to a doubled-down focus on the Biohub.
Grand challenges are identified with a 10-15 year horizon, requiring a credible pathway, a capable leader, and significant risk.
CZI Biohubs in New York, Chicago, and San Francisco focus on cell engineering, tissue communication, and deep imaging/transcriptomics, respectively.
Success in the therapeutic realm is seen as enabling an "explosion of a community" to deploy precision medicine based on individual biology.
Understanding mutation to protein expression and predicting off-target effects is key to developing targeted therapies.
CZI's open-source tools and data sets, like cell hygiene atlases, are enabling innovation in the startup and pharma communities.
The Cell Atlas aims to create a "periodic table of elements equivalent for biology" by standardizing and aggregating data.
CZI funded the methodology for single-cell work and built annotation tools to address bottlenecks in data analysis.
Virtual cell models are considered a critical tool for generating hypotheses, enabling precision medicine, and accelerating scientific work.
Evolutionary Scale researchers are joining a Biohub to lead the science program, focusing on AI's role in understanding biology.
The Biohub network aims to build tools that generate novel data sets, which are then used to build increasingly general AI models of virtual cells.
Virtual biology simulation allows for testing riskier ideas and asking more ambitious questions computationally before investing in wet lab experiments.
Early reasoning models over biology are being developed to go beyond correlations and understand the "why" behind biological processes.
A hierarchical approach, starting with protein models and building up to cellular and system-level models (like a virtual immune system), is being employed.
CZI is unifying its efforts into a single Biohub operating philanthropy under Alex Reeves' leadership, focusing on advancing biology at the AI intersection.
Bringing AI and biology research together under one roof allows for a complete feedback loop, enabling the models to inform data set generation and vice versa.
Domain-specific AI models, when combined with a deep understanding of the problem (like biology), yield better results than general-purpose models.
Communication and user interfaces are critical for making complex scientific data and models accessible to a broader range of scientists.
Organizational structures, like having interdisciplinary teams sit together, can solve complex problems and foster collaboration.
Science itself is viewed as a portfolio, and CZI aims to be additive by focusing on underrepresented areas and encouraging collaboration.
The first Biohub fostered collaboration between different universities and brought biologists and engineers together physically.
CZI is expanding its compute resources, recognizing that large-scale AI models require significant computational power beyond individual lab capabilities.
CZI's approach combines tolerating early ambiguity with impatience for progress, learning from feedback and iterating on successful tools.
The creation of data sets to leverage AI and large language models is a direct result of years of foundational work.
The pipeline from accelerating basic science to developing therapies and improving public health requires foundational tools and AI.
Episode Details
- Podcast
- a16z Podcast
- Episode
- Mark Zuckerberg & Priscilla Chan: How AI Will Help Cure Disease
- Official Link
- https://a16z.com/podcasts/a16z-podcast/
- Published
- July 9, 2026