When AI takes on more of the execution, knowledge workers function more like art directors. They curate the references that shape the work and determine whether the resulting work meets their standard.
We tend to describe these skills in engineering language: Provide context, write specifications, build evals. But these technical skills will not get you very far without an underlying point of view—dare I say "taste"—about what makes for a good reference or finished product.
Fortunately, there is a type of institution that excels at training exactly that: art school.
First, art school teaches artistic literacy. Students do not encounter art as a stream of decontextualized images on Pinterest; they learn who made the work, what influenced them, what came after it, and why exactly it was successful. Over time, they build a library of references, learning to reach beyond the obvious examples available to everyone else.
Second, art school teaches critique. Students present their work to panels of experts, who tear it apart using exacting criteria. Think about the judges on Project Runway: They do not merely say, "Something's off here." They say, "The draping is beautiful, but the volume around the hips throws off her proportions." Having your work repeatedly critiqued in this manner improves your own taste and discernment, and teaches you to evaluate work similarly.
You might say it teaches you to write evals.
The most in-demand talent will excel on both fronts: They'll have good instincts around which references to pull in and when, plus the ability to clearly communicate standards of excellence to people and robots alike.
Tomorrow's tools may be technical, but the people who wield them most effectively will be trained like artists.
