On Tuesday, Priscilla Chan and Mark Zuckerberg unveiled their visionary quest:
The eradication of human disease by 2100. Their audacious plan centers on developing a powerful computing system that researchers can leverage alongside AI to meticulously catalog cells and predict their behavior when afflicted by diseases. The insights gleaned from this data could potentially lead to groundbreaking discoveries that could fundamentally eliminate human diseases, as articulated in their statement.
While the announcement has generated excitement in the scientific community, experts caution that the project is still in its nascent stages. Anne Carpenter, a researcher at the Broad Institute of Harvard and MIT, emphasized that the endeavor is not an immediate solution but rather a forthcoming resource for biologists to create novel models.
The Chan Zuckerberg Initiative, the couple’s LLC, intends to have its innovative product operational by 2024, though the financial investment required remains undisclosed. The endeavor’s costs are expected to be substantial, given the scarcity of the computer components it seeks.
Here’s how the proposed system could function:
The project envisions constructing a computer system equipped with 1,000 graphics processing units (GPUs) specifically designed for image processing. These GPUs possess the capability to analyze both healthy and diseased cells from extensive databases. Following this, the initiative aims to democratize the tool by making it accessible to the broader research community, fostering collaborative research endeavors.
One of the promising aspects of this proposal is its potential to level the playing field for laboratories across the United States. Currently, not all research facilities have equal access to the computing power required for cutting-edge biotechnology research. If the project materializes as envisioned, it could empower researchers who employ AI in the search for innovative therapeutics.
It’s worth noting that while the Chan Zuckerberg Initiative primarily supported local San Francisco community and education initiatives since its inception in 2015, Mark Zuckerberg’s other ventures have ventured into biotechnology territory before. In 2022, Meta, previously known as Facebook, introduced an AI tool called ESMFold, which shares similarities with the AI processing techniques proposed for their new database.
ESMFold essentially expedites protein folding predictions, a longstanding scientific challenge, using AI. By training the model on known proteins, the AI system can predict the folding and structure of previously unknown proteins. This achievement sparked innovation and potential applications in the realm of biology.
The AI methods behind ESMFold and AlphaFold, a similar program recognized for its breakthroughs in protein structure prediction, align with the large language model (LLM) framework, a type of AI trained on extensive online datasets. This same LLM framework is the cornerstone of the CZI project.
Large language models, such as the one powering this project and ChatGPT, have demonstrated their effectiveness in enhancing our understanding of complex structures like protein and cell structures. By harnessing the capabilities of AI and innovative computing systems, Chan and Zuckerberg’s initiative holds the promise of revolutionizing disease research and taking substantial steps toward a future free of human disease.