On the same holiday weekend that TechCrunch dropped its definitive AI glossary, America was busy debating its own foundational lexicon. Words like 'providence,' 'freedom,' and 'democracy' were being stress-tested at bicentennial celebrations while artists, historians, and politicians all made competing claims to define what 250 years of nationhood actually means. The throughline is embarrassingly obvious once you see it: naming is power.
Hallucinations, Founding Myths, and the Authoritative Definition
The AI glossary project is, at its core, an act of institutional authority. Who decides what 'hallucination' means in a machine learning context? The same dynamic governs The Atlantic's excavation of how the simultaneous deaths of Adams and Jefferson on July 4th, 1826, cemented providential American exceptionalism as near-theological fact. A useful myth becomes a glossary entry becomes an unchallengeable truth. The British Museum's new show recasting America's founding through Indigenous eyes is essentially a competing glossary, a counter-taxonomy of the same 250 years. It insists on different definitions: whose land, whose liberty, whose declaration.
Defining Culture in Real Time
The AI terminology problem is especially acute because the field moves faster than institutions can ratify meaning. 'Agent,' 'alignment,' 'grounding': each term arrives pre-loaded with ideological freight that its definers often don't acknowledge. Kyle Chayka's work on algorithmic homogenization maps the same territory from the cultural side: when platforms define what 'relevant' means, they're writing a glossary nobody voted for. The Atlantic's 30 objects exercise attempts the opposite, letting artifacts do the defining rather than institutions. Objects resist spin in ways that press releases don't. The lesson for both AI and national mythology: whoever publishes the glossary first tends to win, at least for a generation.