The Quiet Luxury of Language

🔒 On Friday, Premium Subscribers received the latest batch from After Office, my serialized book about the future of work, cities, and companies.

🚀 The first cohort of Hype-Free AI is starting later this month. Get a practical introduction to the tools and trends shaping your industry, business, and career. Click here to learn more and sign up.

🗞️ I spoke to Vox's Rani Molla about the changing nature of work and the implications for offices and cities. Read here.

🎧 The audio version of this article is available on Spotify, Apple Podcasts, and beyond.

It's easier to be the King of England than an office worker. When Charles was coronated last week, he looked ridiculous. But he also looked just as a king should. The protocol for that specific job is clearly defined; essentially, anyone can do it. Just follow the rules, and you'll do fine.

For other elite jobs, it's trickier. As The Nation points out, technology and globalization made it hard to distinguish between the ultra-rich and the rest of us:

One of the great disappointments of contemporary life is that in times of hopelessly vast income inequality, society’s arch-capitalists—and many very rich people in general—are boring. The zenith of their boringness? How they dress. Look, for example, at Meta CEO Mark Zuckerburg. Extremely rich guy. If you saw him on the street, walking around in a Patagonia vest and a quarter-zip pullover paired with chinos, you’d easily mistake him for one of your dad’s friends.

Of course, this is only true on the surface. In reality, the ultra-rich have not given up on signaling their wealth; they just do so in more subtle ways. These ways often eschew the unsophisticated observer. It's called Quiet Luxury, or QL for short. Glamour explains:

...dressing subtly, simply, and in a way that exudes luxury without the obvious markers, such as excessive logos, obvious name brands, or trending shapes and styles that everyone else is wearing. Instead, quiet luxury leans into classic, tailored silhouettes typically in neutral shades like black, white, beige, gray, and navy. That means no low-rise baggy jeans and crop tops, head-to-toe neon, or sheer metallic cutout dresses. Classic Italian cashmere brands like Loro Piana and Brunello Cuchinelli and minimal designers such as Jil Sander, The Row, and Max Mara are being cited as labels to achieve the look.

The beauty of this approach is that you seem "normal" to most people, but other wealthy people can tell you're one of them. QL is not new, but HBO's hit show, Succession, triggered a barrage of articles and how-to guides on this topic. In the show, a group of billionaire heirs and their friends dress in wardrobes that "exude class and simplicity," epitomized by Kendall’s signature baseball cap. As Fashion put it:

"To the untrained eye, it looks like Ken, played by Jeremy Strong, might have picked up a multi-pack of yawn-inducing navy baseball caps at Costco... But to someone adept at sniffing out the quiet luxury on Succession, that fitted topper is a marker of the Roy family’s stealth wealth: a cashmere Loro Piana baseball cap that’ll run you upwards of $605 USD. Indeed, money talks and wealth whispers.

As class boundaries become harder to discern, the middle class becomes more anxious. This has been going on since at least the 1960s when counterculture undermined traditional middle-class norms. Things got even more confusing when the hippies joined the workforce and ascended into management positions. California was once again the center of the universe, but this time it wasn't exporting bell-bottom pants and flower shirts; it was exporting khakis and polo shirts.

"Welcome to the confusing world of business casual," The Chicago Tribune declared in 1995, as Silicon Valley's office culture began to infect corporate America. One clothing retailer described how lawyers were "hysterical" and had "to be educated" once they were allowed to wear anything other than "pin-striped suits" for one day a week.

In 2002, CNET observed that employees struggle to figure out what's appropriate and are anxious about appearing like someone of a lower class:

Employees accustomed to putting on a dark suit, dress shirt and lace shoes (for men) or suits, skirts, blazers and high heels (for women) suddenly have to worry about their attire.

"Now it's what kind of slacks, what kind of shirt, what's acceptable and what's not. The ranges and limits are tested, and they are a lot more ambiguous," said one ambivalent convert to casual wear.

In other words, it takes considerable thought and confidence to dress well casually.

Then there is the question of how to distinguish yourself in a workplace where you want to be casual but don't want to look like the person two floors below who makes one-third your salary.

The intricacies of Quiet Luxury trickled down to the middle class. It became harder to figure out who really belonged. Immigrants and those who made their way up from the working class could no longer follow a simple dress code and fit in. Those who knew — knew. Those who didn't know, thought they were dressing like everyone else but, in reality, they indicated their recent arrival in a variety of subtle ways.

The democratization of office attire and luxury clothes made it harder for outsiders to fit in. It created a semblance of equality but, in reality, created a whole new stratum of premium products and premium people.

The democratization of "intelligence" might do something similar. It will affect the type of work humans do in person — and the distribution of rewards from such work.

As you may have heard, ChatGPT and other Large-Language Models (LLMs) enable anyone to write intelligently about almost anything. Below is a simple example. I asked ChatGPT to write an email that pitches a story to The New York Times. You can click to enlarge the image, but reading it is unnecessary.

Then, I asked ChatGPT to write the first two paragraphs of the actual story:

As you can see, ChatGPT does a solid job of writing in proper English and sounding credible. But it has no character or signs of true intelligence. The output can be improved through iteration and more complex prompts. You can also ask it to imitate the style of a specific writer to give it more character:

But the bottom line is that ChatGPT's output is quite plain. It might seem excellent and correct to a non-native speaker or to an unsophisticated reader. But an actual NYT editor could easily tell this isn't the right stuff.

Just like in the fashion industry, cheap substitutes can only fool some people. But unlike fast fashion, we can expect AI's capabilities to improve exponentially — making it harder to spot mass-manufactured text.

And yet, I suspect that as machines become better at sounding like sophisticated humans, the most sophisticated humans will adopt even more nuanced, coded, and complex ways of speaking that are harder to imitate. Consider what happened in financial markets: As machines become better at predicting price movements, markets necessarily become more complex and harder to predict in a variety of new ways. Every new strategy has diminishing returns; once everyone begins to behave in a certain way, the edge is pushed further out. Yesterday's alpha is tomorrow's beta. In parallel, new behaviors emerge in an effort to beat the market. Language, fashion, and other forms of social signaling are complex adaptive systems that behave in a similar manner.

This is already happening. And Succession offers some great examples of that, too. The show is full of cultural, historical, and business references that only certain people can make sense of, for example (and pardon the language):

• "If a deal collapses in the woods and no one hears it, is it still an SEC violation?"
• "You don’t hear much about syphilis these days. Very much the MySpace of STDs."
• “They’re not all crypto-fascists and right-wing nut jobs. We also have some venture capital dems and centrist ghouls.”
• “It’s an incredibly evolved, ruthlessly segregated city you’ve built on this geological fault here.”

Netflix, and the internet in general, are full of shows that might as well have been written by AI. But Succession was clearly written by a team of humans who are masters at their craft.

And it won't end there. Language is likely to become even more segregated and exclusive. Consider the trajectory of human language from the invention of print until today. Printing was the first era of language abundance — suddenly, it was possible to take all of the world's knowledge and put it into everyone's hands. But printing didn't just spread information; it also changed language itself. Ben Thompson summarizes the beginning of this story:

Meanwhile, the economics of printing books was fundamentally different from the economics of copying by hand. The latter was purely an operational expense: output was strictly determined by the input of labor. The former, though, was mostly a capital expense: first, to construct the printing press, and second, to set the type for a book. The best way to pay for these significant up-front expenses was to produce as many copies of a particular book that could be sold.

How, then, to maximize the number of copies that could be sold? The answer was to print using the most widely used dialect of a particular language, which in turn incentivized people to adopt that dialect, standardizing language across Europe.

Print didn't just make it possible and profitable to print millions of copies of the same texts; it made it necessary to do so. The economics of print dictate mass production and the adoption of a unified language that as many people as possible should understand.

The economics of artificially-produced language are very different. At a minimum, tools like ChatGPT make creating millions of unique texts possible and profitable. They break the trade-off between personalization and mass production. To signal their uniqueness and intelligence, humans will have to develop and adopt narrower and stranger forms of communication. To signal their belonging, members of different groups will have to increasingly differentiate themselves from other groups (and machines).

I am trying to say that printing pushed the world towards a unified language. Artificial intelligence might move it in the opposite direction. It will likely make English even more popular, but it will create narrower niches within English that only make sense to certain people. To some extent, the internet is already doing precisely that: It draws more people into the English-speaking culture, but it pushes the boundaries of that culture further out, making a lot of dialogue utterly unintelligible to people who formally speak the language. AI will intensify these dynamics.

The fashion industry illustrates how this happens. Mass production "pushed" the ultra-wealthy to seek more artisanal items in order to differentiate themselves. Production technology constantly got better at copying their style and driving down the cost of their favorite materials and methods. The ultra-wealthy pushed further into rarer materials and techniques — and more subtle ways of signaling that are hard to copy because they are hard to understand and describe. This created a world in which the ultra-rich understand each other's signals while the same signals are invisible to everyone else.

The mass production of "premium" goods resulted in a world where "money talks and wealth whispers." The mass production of "premium" content will give rise to a world of Quiet Intelligence — everyone will think they sound smart, but those who are really smart (or "in") will communicate at a whole different level.

Best,

P. S.

This is not just a story of economic elites. There will be multiple groups, positioned in various ways, who adopt their own speech forms in order to signal belonging and authenticity. Humans have been doing this for a long time, but now they are able and incentivized to do it more intensively and at a higher level of granularity.

P. P. S.

If you're interested in AI's impact on your industry and career, check out my upcoming course. We'll have a fun group of people from across the world, 100% online.