CONCEPT
Threshold Model of Collective Behavior
Granovetter's 1978 model explaining why similar groups produce dramatically different outcomes — and why AI adoption cascades in some communities and stalls in others.
A threshold is the proportion of a group that must act before a given individual will act. Granovetter's 1978 model demonstrated that groups with identical average thresholds can produce dramatically different collective outcomes, depending on the distribution of thresholds across members. A group with evenly distributed thresholds cascades into full adoption; a group with thresholds clustered around the middle stalls after the first mover. Applied to the AI transition, the model explains why the orange pill spread through developer communities in weeks while other professional communities resisted for years — not because individuals were less capable, but because threshold distributions created different cascade dynamics.
In The You On AI Field Guide
The critical insight is distributional, not averaged. Two communities with identical median thresholds can produce opposite outcomes if the variance around that median differs. In developer communities, thresholds were heavily skewed toward low values: professional identity was aligned with frontier-seeking, weak ties provided extensive exposure to others crossing the threshold, and cascades