Attention Scarcity Migration

In The You On AI Field Guide

The attention economy's empirical emergence traces to the late 1990s and early 2000s. As internet distribution costs approached zero, the economic question shifted from who can reach the audience (distribution) to who can hold the audience (attention). The companies that captured the largest economic returns of the internet era — Google, Facebook, Amazon — were those that captured attention at scale, often through algorithmic curation optimized for engagement.

Andreessen's articulation of the pattern, developed across Andreessen Horowitz's investment memoranda and public writing in the 2010s, emphasized its structural character. The migration from distribution to attention was not a contingent feature of specific companies but a predictable consequence of what happens when one scarcity is resolved. The pattern could be used to predict where value would accumulate in the information economy, and the firm's investment thesis was explicitly organized around this prediction.

The AI transition introduces a second migration. As execution costs collapse through language models and related tools, the binding constraint moves from producing output to evaluating output. This second migration, the Andreessen — On AI volume argues, is the most consequential shift in the economics of cognitive work since the distribution-to-attention migration that defined the previous generation of internet companies.

The framework has specific implications for labor markets, professional education, and corporate strategy. Labor that was previously valued for its execution capacity is being repriced to reflect the value of its evaluative capacity — a repricing that produces winners and losers whose identity depends on which capacity an individual has cultivated. Professional education organized around execution training faces obsolescence questions it has not confronted in previous generations. Corporate strategy must now account for the possibility that execution-heavy business models face structural repricing, while judgment-heavy models face new competitive advantages.

The framework's limits are increasingly visible. Judgment is not a single capacity; it decomposes into forms that develop differently. The attention-to-judgment migration may itself be incomplete — some forms of judgment are being absorbed into AI systems through RLHF and related techniques, suggesting the next scarcity migration may already be beginning.

Origin

The distribution-to-attention component of the framework crystallized in Andreessen's thinking through the 2000s as he observed the economic patterns of the commercial internet. The execution-to-judgment extension developed in 2023–2025 as the specific capabilities of frontier AI systems became visible. The full framework in its current form appears in the Andreessen — On AI volume and in Andreessen Horowitz's AI investment thesis.

Key Ideas

Structural migration pattern. Binding constraints in information economies do not disappear when technology resolves them; they relocate to whichever resource remains scarce.

Distribution-to-attention phase. The internet's collapse of distribution costs produced attention as the new binding constraint — the structural foundation of the attention economy.

Execution-to-judgment phase. AI's collapse of execution costs is producing judgment as the new binding constraint — the structural transition the current moment represents.

Labor market repricing. Each migration produces specific repricing of labor categories, with the new scarce resource commanding a rising premium over the previously scarce resource.

Recursive possibility. The pattern may continue — as AI absorbs specific forms of judgment, scarcity may migrate again to forms of evaluative capacity that remain difficult to automate.