Curing All Human Diseases & the Future of Health & Technology | Mark Zuckerberg & Dr. Priscilla Chan

KEYPOINTS

• 01

  CZI's goal: help science cure, prevent or manage all disease by 2100

  Chan, a physician and educator, frames the Chan Zuckerberg Initiative (launched in 2015) as a bet on basic science rather than on any single disease. CZI does three things: funds scientists, builds software tools and runs its own science through 'Biohubs'. Both stress that CZI won't cure disease itself. Its goal is to speed up the whole field, and its funding is a small fraction of what the NIH spends.

• 02

  New tools usually come before big discoveries

  Zuckerberg says their reading of the history of science is that major breakthroughs tend to follow a new way of seeing, such as the telescope or microscope. He argues the next ten years should mainly go to measuring and observing human biology, comparing it to debugging code: you can't fix a codebase you can't step through, and at Meta big projects start with instrumentation.

• 03

  The cell sits between a genetic typo and the disease

  Chan's analogy is a recipe with a typo that produces a bad cake: we know the typo (the mutation) and the result (the disease), but not how the chef and oven, the cell's mRNA, turned one into the other. Single-cell work studies how each of the body's roughly 37 trillion cells reads the same DNA differently in health and in disease.

• 04

  Cell atlases plus AI could produce a 'virtual cell'

  Since CZI funded methods work in 2017, near-complete single-cell atlases of humans, flies and mice have emerged, far too much data for people to interpret. Large language models (based on the Transformer architecture) are pattern-recognition machines that work on more than words. Trained on cell-atlas data, a model given a cell's state could predict its next states, much as AlphaFold did with protein folding.

• 05

  In-silico experiments still have to be checked in humans

  Chan welcomes cheap, fast computer experiments but warns that 'we've basically cured every single disease in mice'. Model systems are relevant but not one-to-one with humans, so predictions must be confirmed in human biology.

• 06

  Biohubs require collaboration across institutions

  Proposals had to involve at least three institutions and several disciplines. The call drew 57 proposals from over 150 institutions. Each Biohub has about 50 dedicated staff, independent of the universities, while letting faculty contribute without leaving their posts. CZI runs in five-year cycles: try many models and double down on what works.

• 07

  Chicago Biohub: sensors inside engineered tissue to catch inflammation early

  The University of Illinois Urbana-Champaign, University of Chicago and Northwestern are embedding ultra-thin sensors in lab-engineered skin to read what cells secrete and when inflammation starts. Chan notes inflammation drives about 50% of all deaths. They're also studying the neuromuscular junction, whose slowing with age helps explain why older people fall.

• 08

  New York Biohub: immune cells engineered to act as internal microscopes

  Zuckerberg calls it a 'cellular endoscope'. The plan is to engineer immune cells that travel to tissues such as coronary arteries, record whether plaque is present and report back, then eventually engineer cells to clean it up. Targets include places the immune system rarely reaches, such as ovarian and pancreatic cancer and neurodegenerative disease. They describe it as a 10–15-year project.

• 09

  CZI leaves drug development to others but backs rare-disease groups

  CZI doesn't develop therapeutics itself, though some Biohub teams have spun out startups. Its Rare As One program funds patient advocates to build registries and recruit researchers and drug developers. Chan stresses that 'rare diseases aren't rare': there are over 7,000, and each offers a window into how normal biology works.

• 10

  Social media is best for connection, worst as negative passive scrolling

  Zuckerberg separates interacting with people, which he associates with well-being, from passive consumption, which is at best entertainment and at worst relentlessly negative news. Meta defaults under-16 accounts to private and offers parental tools, time-use tools and nudges when teens get stuck on one type of content. He says Meta has at times been 'too paternalistic' about deciding what content is good for people.