A Glossary for the Perplexed. It Is Overdue. The Terms Are Not Optional Anymore.

The Boston Globe published a beginner's glossary covering core AI terms including machine learning, deep learning, distillation, and compute. Deep learning is defined as brain-inspired neural networks that find patterns in massive datasets. Distillation is described as using a powerful model to teach a smaller one, which the piece notes is controversial because it can leverage another developer's investment. Compute is defined as the processing power needed to train and run AI models, with related reference to massive data centers.

The underlying principle is what I would call vocabulary as gatekeeping. You cannot evaluate what you cannot name. The term distillation, for instance, is not just a technical curiosity. It is the center of a legal and ethical dispute about whether smaller AI companies are effectively stealing multi-million dollar investments from larger ones. Compute is not an abstract concept. It is why AI companies are building enormous data centers and consuming staggering amounts of electricity. The terms map directly to the stakes.

The Boston Globe published the glossary for readers trying to follow business and political debates around AI. The guide covers terms including deep learning, distillation, and compute.