The Cup of Coffee Theory of AI

As the author Ryan Holiday’s research assistant, I’m worried about AI replacing me. 

By most of the common psychological tests of intelligence, AI is smarter than me. AI can brainstorm more ideas than I can. AI can read a book, find information, fact check, and create content faster than I can.

So I’ve considered going back to working in coffee shops.

After I graduated from college, I chased winter. I lived and skied all over Colorado for the Northern Hemisphere’s winter months, and for the southern hemisphere’s, Australia and New Zealand.

To fund this year-round skiing, wherever I went, I worked part-time as a barista. When I think about my time working in coffee shops around the world, I find reasons to be optimistic about my job security.

Coffee shops around the world all have the same powerful machines. They’ve science’d most parts of the process—how the beans are grown, the mineral composition of the water, the milks and the drinkware. So, unlike not that long ago, you can get a great latte, cappuccino, or flat white basically anywhere. 

Yet many people still love their drip coffee with half and half. Many love their cheap store-bought beans brewed by their countertop coffee pot. And many don’t care—they just want their shot of caffeine.

Coffee preferences are personal and abundant. The world is big enough for espresso machines and French presses. And which machine the barista uses depends on who they’re serving. 

Art preferences are similar. So when I hear talk about how AI is going to replace artists, I think to myself, the world is big enough for both. Or when I see tweets about how AI wrote this “great” article or produced this “great” image, I ask myself, “great”—according to whom?

Countless authors, painters, and inventors who were considered ordinary in their own time are revered in ours. And anyone who has made and released creative work has experienced something analogous to making a delicious espresso for an audience that prefers a French press.

This is the perennial challenge of the artist: finding the middle of the Venn diagram where one circle is the artist’s tastes and the other is the audience’s tastes. 

For the past two years, we’ve tried to hire another research assistant. We’ve trialed dozens of smart, creative, ambitious, book-loving, speedy fact-checking individuals. We always give them a version of the same assignment: read this book (usually a biography) and pull out two good anecdotes and two good quotes. What they come back with is not in the middle of that Venn diagram. 

In a word, we’ve struggled for two years to find someone with taste—with an eye for what stands the chance of being interesting, entertaining, or useful to an audience. 

We’re not alone. I’ve talked to authors, podcasters, documentarians, late-night TV writers, YouTubers and TikTokers, and they all share this problem. Most of these creators are actively or passively looking for people—whether researchers, writers, animators, video editors, producers, or production managers—with what Jerry Seinfeld identified as the ultimate skill of the artist: “taste and discernment.”  

“It’s one thing to create,” Seinfeld said. “The other is you have to choose. ‘What are we going to do, and what are we not going to do?’ This is a gigantic aspect of [artistic] survival. It’s kind of unseen—what’s picked and what is discarded—but mastering that is how you stay alive.”