Twenty-five years in development

The dominant synthetic-persona method begins with categories such as age, gender, income, region, occupation, ideology, or personality type. A language model then generates biographical and psychological detail beneath those labels.

The result can read like a person without containing a sufficiently deep model of one. Increasing model size produces richer narratives, but it does not, by itself, create the underlying cognitive structure those narratives appear to describe.

QBC reverses the direction of construction. It begins beneath demographic and psychographic labels, with computational cognitive units called Quanti. It models how interactions among those units generate motivations, meanings, identity commitments, trust relationships, internal tensions, resistance, and change. Demographic and behavioral characteristics can then emerge within the modeled persona rather than predetermining that persona from above.

QBC renders these modeled individuals as queryable Persona Voices, organized into Human Terrain Maps. Each terrain represents both the modeled internal cognitive structure of its Persona Voices and the relationships operating across the population.

Users can question the terrain, examine differences among its voices, inspect the modeled causes of a response, introduce new evidence, and compare predicted responses with observed outcomes. The terrain can therefore be audited, challenged, updated, and progressively validated rather than accepted as an opaque AI answer.