The paradox inside abundance

McKinsey Global Institute’s The AI Economy: Interconnected Forces, Feedback Loops, and Speeds of Change invokes a familiar economic phenomenon in considering AI: the Jevons paradox. Improvements that make a resource more efficient do not necessarily reduce its total consumption. Lower cost can stimulate new uses, expand demand, and ultimately increase aggregate consumption.

The observation becomes particularly interesting when the resource becoming cheaper is intelligence.

If reasoning becomes cheaper, we may not simply perform the same amount of reasoning at lower cost. We may reason more. If software becomes dramatically cheaper to produce, we may create vastly more software. If analysis becomes inexpensive, organizations may perform more analysis. If decisions can be generated continuously, the number of decisions may expand. If autonomous agents reduce the cost of execution, organizations may initiate activities that would previously have been uneconomic.

AI abundance therefore does not merely make existing activity more efficient. It can create entirely new quantities of activity.

That changes the meaning of efficiency.

Efficiency is a ratio, not a destination

Much of industrial and managerial thinking has been built around improving ratios: less energy per unit, less labor per transaction, less time per process, lower cost per decision. These improvements matter. But an efficient unit multiplied by an exploding number of units can still increase the total burden placed upon the system.

The same reasoning applies beyond energy and compute.

Suppose AI reduces the cost of forming companies. We may create more companies. Suppose it reduces the cost of producing personalized products. We may produce more products. Suppose it reduces the cost of persuasion. We may generate more persuasion. Suppose it reduces the cost of decisions. We may surround ourselves with more decisions than human institutions can meaningfully absorb.

At some point the optimization problem changes. It is no longer sufficient to ask whether each unit has become cheaper or more efficient. We have to ask what the expanding system is consuming, what it is degrading, and what must remain capable of supporting another cycle.

That is where the idea of regeneration begins to matter.

From efficiency to regenerative capacity

Efficiency asks how little we can consume to produce a unit of output.

Regeneration asks a different question: What condition do we leave behind after producing it?

That difference is profound.

A system can be extraordinarily efficient while progressively depleting the conditions on which its future depends. Soil can be made productive while losing biological capacity. Organizations can extract more output from people while diminishing institutional resilience. Markets can optimize transactions while weakening the relationships and trust that allow markets to function. AI can accelerate knowledge production while simultaneously overwhelming our capacity to establish provenance, meaning, and judgment.

The consequence of abundance, then, is not automatically prosperity.

Prosperity depends upon whether the systems generating abundance reproduce the capacities they consume.

This is why regenerative thinking increasingly feels relevant far beyond agriculture or ecology. Regeneration may be an operating principle for any system in which increasing efficiency creates increasing throughput. It asks whether each cycle leaves the underlying system less capable, equally capable, or more capable of supporting what comes next.

A different objective function

The AI economy may eventually force us to reconsider the objective function itself.

If intelligence becomes abundant, maximizing intelligence is not particularly interesting. If creation becomes abundant, maximizing creation is insufficient. If execution becomes abundant, maximizing activity may become actively destructive.

The scarce resource may instead become the capacity to determine which activity deserves to exist, what consequences it creates, and whether the system from which it draws remains capable of renewal.

This connects two conversations that are usually kept apart: artificial intelligence and regenerative systems.

One is commonly framed around abundance.

The other begins with limits, interdependence, renewal, and consequence.

Perhaps they belong together.

The question I want to keep

The Jevons paradox asks us to notice that efficiency can increase consumption.

The regenerative question goes one step further:

After abundance has multiplied what we can do, is the system that made it possible more capable of supporting the next cycle—or less?

That may become one of the more consequential measures of prosperity in an age of abundant intelligence.

Abundance does not abolish consequence.

It makes consequence easier to multiply.