An AI budget can buy more capacity while leaving the people who need it waiting for permission to try something useful.

In his October 25 essay, “Cognitive Mercantilism,” Venkatesh Rao argues that much AI debate treats intelligence as a scarce resource to accumulate and control. Drawing on his reading of historian Joyce Appleby, he offers another possibility: cheaper access may let people discover projects and ambitions they could not previously afford to pursue.

This is an argument about demand, not a measured forecast of employment or proof that AI benefits will reach everyone. Rao also acknowledges that freedom creates risks and that powerful autonomous systems need constraints. The useful strategic question is narrower: what becomes possible for the person who gains access?

For an enterprise, that question can change the design of an adoption pilot. Imagine a service team with a backlog of small reporting requests that never justified a software project. Giving the team an approved way to explore those requests tests a different proposition from asking it to complete the existing report faster. The first experiment looks for useful work that was previously left undone. The second measures efficiency within a known task.

Keep those outcomes separate. For the exploratory pilot, record who proposed the work, whether someone used the result, what review it required, and whether its value justified the effort. An increase in generated reports alone would not answer those questions. Neither would a busy usage dashboard.

This interpretation of Rao’s essay does not require accepting his larger historical analogy. It does require hearing from people outside the group that originally selected the tools. Otherwise, a pilot can discover only the uses its designers already imagined.

Start with one neglected request and an accountable reviewer. The result worth keeping is a useful new capability someone can actually use.