The Gordon and Betty Moore Foundation
In The You On AI Field Guide
The grants were not intended to enable AI. They were intended to support open science and reproducible computational research. The foundation's Data-Driven Discovery Initiative, launched in 2014, funded open-source tools because Moore and his team believed that scientific progress depended on researchers being able to share methods as well as results. Jupyter provided the notebook interface that let researchers combine code, narrative, and results in a single document. NumPy provided the numerical array operations that underlie almost all scientific computation in Python.
These tools, developed initially for physics, genomics, and climate science, became the default infrastructure of machine learning research. TensorFlow, PyTorch, and the Jupyter-based workflows that run virtually every AI training pipeline descend from the infrastructure the Moore Foundation's grants supported. The Colab notebooks that enable experimentation at scale, the Kaggle competitions that train a generation of ML practitioners, the research papers that cite computational methods — all run on tools traceable to philanthropic investments Moore made without knowing what they would eventually enable.
The chain from the 1965 paper to the current AI moment runs through this foundation. Moore drew a line on a graph and organized an industry. The industry produced chips. The foundation funded the software tools that eventually harnessed those chips to train models that learned to speak human language. No single act of planning produced this outcome; it was the consequence of values — measurement, open access, the diffusion of capability — applied consistently over decades, generating, through the compounding logic of the exponential, outcomes that exceeded any individual's foresight. This is, in Moore's framework, the characteristic shape of engineering at the exponential frontier: consequences compound beyond what any single decision could have intended.
Origin
The Gordon and Betty Moore Foundation was established in November 2000. The Data-Driven Discovery Initiative, which funded Jupyter and NumPy, was launched in 2014. Grant histories and project reports are publicly available through the foundation's website and through the annual reports of the NumFOCUS organization that administered much of the open-source funding.
Key Ideas
Philanthropic infrastructure for open science. The foundation funded the computational substrate of modern research, predating and enabling the AI era.
Jupyter and NumPy. The specific tools funded became, a decade later, the default infrastructure of machine learning research worldwide.
Unplanned consequences. Moore did not fund these tools to enable AI; he funded them to support open science, and the AI consequence compounded through the same exponential logic his law identified.
Values over planning. The foundation's work embodied values — measurement, open access, diffusion of capability — that Moore's career consistently demonstrated across both commercial and philanthropic domains.
Endowment scale. With an endowment exceeding five billion dollars, the foundation operated at a scale that made infrastructure-level contributions possible.