Why We Like Things—Asterisk

What Silicon Valley gets wrong about taste.

When I was 22, I met, and later fell in love with, a woman who had impeccable taste. She dressed well; that was the first thing I noticed. She always knew when to eschew the latest Michelin-starred restaurant for an obscure diner in her neighborhood. She could tell you what art openings to go to and what albums to listen to. I admired her; I wanted to be her and be with her.<br>For the first few years of our relationship, I remained over-awed at how easily she moved through the world, automatically affiliating herself with the right people, ideas, experiences.  Later on, I began to see that it wasn’t automatic: There was an art to knowing what was cool and seeming not to care. But it was an art she excelled at. Everything she did — the books she read, the music she listened to, the theorists and takes she referenced — seemed to exhibit that quality known as taste.<br>Good taste, as I understood it, meant that you were consuming the right things, from commodities to cultural works. But it was also, as I began to realize in my professional life, something that made people like, trust, and value you more. A few years into our relationship, I began working for a tech startup where the CEO encouraged us to watch French New Wave cinema. The most tenured designers on my team, dressed in Issey Miyake, were trusted for their taste as much as their work ethic. Informal meetings often took place at the Togo chairs scattered around the office, where people peered at laptops together and drafted PRDs. And all of this — the lunchtime conversations about Lemaire, the office that looked like a Design Within Reach showroom, the Heath ceramics in the kitchenettes — was understood to be, in some way, the reason for the company’s success.<br>The way we spoke about our work, too, leaned on evocative metaphors and instincts. We were distinguishing ourselves from the previous generation of tech companies, who often eschewed intuition for data: then-Google executive Marissa Mayer famously insisted that her team A/B test 41 different shades of blue. Skip ahead a few years though, and you’re more likely to hear founders evoke the ineffable. “We are extremely data-informed,” the co-founder of Shopify, Tobi Lütke, said in a podcast, but “comfortable with the unquantifiable things, such as tastes, quality, passion, love, hate.”<br>Tech’s preoccupation with taste has intensified in the last two years; it is now, the critic Kyle Chayka observed, “as much of a tech-world cliché as ‘disruption’ was in the 2010s.” And it has much to do with the industry’s embrace of generative AI, which is shifting the focus of knowledge work from production to selection and curation. “I can easily imagine a world,” OpenAI’s CEO Sam Altman said to Die Welt, “where 30, 40% of the tasks that happen in the economy today get done by AI.”<br>Taste has become the catch-all term for what’s left. “In the AI age,” the Y Combinator cofounder Paul Graham tweeted in February, “taste will become even more important. When anyone can make anything, the big differentiator is what you choose to make.” Greg Brockman, the president of OpenAI, was even more succinct: “Taste,” he wrote, “is a new core skill.”<br>Taste is also what the music producer Rick Rubin — whose bestselling The Creative Act gave him a second career as Silicon Valley’s self-help guru — cites as the reason for his success. “I have no technical ability,” Rubin said in 2023. “So what,” his interviewer asked, “are you being paid for?” Rubin’s response, which has since been clipped and reshared thousands of times on YouTube and X, was simple: “The confidence I have,” he said slowly, “in my taste.” It’s this confidence that founders and funders alike (including the venture capitalist Joshua Kushner, who sought Rubin’s advice on “intuitions…that I cannot explain to anyone”) are hoping for.<br>But those who invoke taste often fail to explain what it is. And curiously, although many in Silicon Valley are bullish on AI’s potential to write novels, solve mathematical proofs, and generate feature-length films — to create outputs, in other words, that combine taste and technique — they seem convinced that taste itself is not replaceable. To the literary scholar Leif Weatherby, the idea that taste “is the last preserve of the human” seems flawed. LLMs, after all, already incorporate “judgments of taste,” from literary-critical texts to Likert scale assessments of response quality, “in a direct and infrastructural way.” Taste, he suggested to me in a phone call, “is a misnomer.” The idea that “this is the last preserve of the human, the ability to make judgments… it’s anxiety on the part of industry leaders…because they’re deskilling, right? They’re saying, If Codex writes the code, what’s your job?” If all code is fungible, if all software features can be commoditized, you need any competitive edge you can get. The fear goes beyond individual job loss; whole...