The productivity surplus, and how it pays for the field

Where the surplus comes from

When generative and agentic AI takes over cognitive task categories that previously required salaried human time — drafting, review, analysis, scheduling, routing, basic adjudication, code generation, customer service, document synthesis — the marginal cost of producing the same output drops sharply. The size of the drop is the subject of active estimation; the McKinsey 2023 forecast of $2.6–4.4 trillion in annual economic value from generative AI is one data point in a wide distribution.

What matters for this argument is not the precise magnitude but the existence and direction of the surplus. The surplus accrues, by default, to the owners of the AI systems and the firms that deploy them — the gap between previous payroll and current operating cost.

Where the surplus goes by default

By default, the surplus accrues as profit, capital appreciation, and shareholder return — concentrated. Some fraction is returned to consumers as lower prices; some fraction is captured by governments as tax revenue. The displaced human labor receives, by default, none of it. That is the structural shape of every prior automation wave.

The standard policy response — sovereign-wealth shareholding in AI firms, AI-specific corporate taxes, UBI funded by either — treats the displaced labor as the residual recipient: a transfer payment to people who no longer have economically valued work.

The reallocation proposal

The proposal of this site is not a transfer payment. It is an explicit wage-for-work reallocation of the AI productivity surplus into a labor-intensive sector that exists already, is undersupplied, and produces a tangible public good: food forest and orchard agriculture.

A wage subsidy or direct-employment program funded by AI surplus, scaled at the magnitude of the displacement, pays orchard, fruit-tree, perennial polyculture, and small-scale market garden labor at a living wage. The food produced enters the food supply at price points the market alone does not currently support, because the alternative — heavily mechanised monoculture — has long since outcompeted small-scale labor-intensive agriculture on dollars per calorie.

The point of the subsidy is not to make food cheaper. The point is to fund the labor whose value is the labor itself — the body using itself in the environment it evolved for, producing the nutrient-dense food the orchard delivers as a byproduct.

Three open questions

• Magnitude. What fraction of the AI surplus is required to absorb a labor force of the scale displaced? This depends on the displacement rate, the wage floor, and the labor absorption per acre of the agricultural systems in question. The mid-range estimates suggest the surplus is sufficient; the precise calculation is provisional.

• Mechanism. Sovereign-fund stake in AI firms (Alaska Permanent Fund pattern), a direct AI-output tax, a federal labor-corps employment program, regional cooperatives, or some combination. Each has different administrative and political profiles.

• Political tractability. The proposal is structurally compatible with both left (labor-share, food sovereignty, public-investment) and right (work requirement, family-scale agriculture, anti-dependency) frames. Whether that crossover survives translation into actual legislation is an open question.

The mechanism is contestable. The structural claim — that AI surplus exists, is concentrating, and could be deployed for embodied labor at scale — is the load-bearing part.