Preferential Attachment

Preferential attachment is the mechanism that produces the hub-and-spoke structure of most real-world networks. When new nodes join a growing network, they preferentially connect to nodes that are already well-connected, producing a power-law distribution in which a few hubs accumulate disproportionate centrality while the vast majority of nodes remain sparsely connected. The mechanism was formalized by Albert-László Barabási and Réka Albert in 1999 but had been recognized in various forms long before — Robert K. Merton's Matthew effect in science, Herbert Simon's work on the distribution of word frequencies, and the economist Pareto's observations on wealth distribution all pointed at the same dynamic. In the AI economy, preferential attachment explains why capability expansion produces hub concentration rather than uniform distribution: each expansion of capability compounds the advantages of the best-positioned users, producing new hubs at the same rate the old ones are displaced.