Productive Capacity Framework Claim: Modeling AI’s impact on and integration with humanity’s productive elements enables us to see and shape outcomes from humanity’s most potent force.

THE FRAMEWORK

• The productive frontier is the set of producers (usually corporations) that produce the first 90% of all goods within a category — these are referred to as the productive set.
• The cardinality of the frontier is the number of entities in each industry’s frontier set.
• The concentration of the frontier is 1 over its size, expressed as a percentage, ranging from 0% to 100% concentration, where one entity produces more than 90%.
• The skew is a measure of how quickly production drops off in each industry.
• The production binning of capacity splits a productive element into a fixed number of bins, where each bin represents the same production output.

• The churn matrix represents the average number of producers in one bin that transition into another bin over a specified time window (e.g., the annual 100-bin churn matrix).

• Hyperscalers are the organizations pushing the frontier of AI capacity.
• Hyperscaler Concentration is 1 over its cardinality expressed as a percent.

Claim: AI-driven automation will accelerate trends that concentrate the productive frontier.

TL;DR. The productive frontier its driven forward by capital and labor, AI shifts that balance towards capital as the determining factor. ???

Claim: AI-driven innovation will dramatically accelerate the concentration of the productive frontier.

TL;DR. Disruptions to the productive set are usually the result of innovation in the means of production. Incumbents have both the capital and assets required to drive these innovations. Disruptions occur when an external group innovates in a novel way not considered by the incumbents. As AI takes over the innovator’s role, will these disruptions happen rarely to any incumbent with a strong connection with a hyper-scaler?

Claim: Nation-state diversity provides ecosystem stabilization.

TL;DR. AI-driven innovation tends to favor frontier concentration and capital-driven decision-making as labor-related concerns fall away. Nation-states, by contrast, prioritize their power and the well-being of their citizens, even in cases where this conflicts with profit maximization. Presently, nation-states have significant control over producers and hyperscalers within their borders, as well as control of domestic purchasing from producers outside their borders. Producer-concentrations that might be beneficial for one nation-state will be staunchly opposed by other nation-states, creating a headwind that can help encourage domestic productive competitors.

Paradoxically, while transnational collaborations on AI safety may improve stability, transnational corporate AIs likely have the opposite effect. Indeed, such transnational corporations threaten to undermine the primary lever humanity has against the inexorable pressure to maximize profit, national self-interest.