A few years back, around the time of the financial crisis, I remember hearing about a guy with a startup idea for a news site. The main expense any online publisher has is people, right? People who stubbornly insist on being paid actual money.

But the Internet meant reporters could churn out content from anywhere. So rather than pay 23-year-olds $24,000 a year to live in a closet in Brooklyn, he wanted to buy up some land in Mexico — somewhere in the Yucatán, on a beach — and build a little Content Village. Rather than Williamsburg bars, he’d tempt young journalists with great surfing, a sandals-based dress code, and cheap cerveza. And because the cost of living was so much lower in Mexico, he could pay them crazy low wages — like half the American going rate.

I don’t believe his little experiment in journalism arbitrage ever went anywhere, but it fit into a sort of thinking common in certain circles in earlier days of the web: The power of the Internet was to eliminate barriers to entry, and geography was going to be one of the first to go. The “Publish” button in your CMS worked from anywhere, after all.

But like so many optimistic narratives from that era of online media, the reality’s turned out different. Rather than disperse the news business around the country, the Internet has concentrated it more firmly than ever in New York and a few other major cities. And that has real impacts on the kind of news we get.

“The Internet doesn’t spread things apart — it pushes them together,” Richard Florida, the urban studies theorist, told me. “You’re seeing more of these winner-take-all effects.”

An artificial distribution

Let’s start by thinking of the pre-web news business. Physical distribution of newspapers and over-the-air distribution of TV signals meant location was all-important for daily news. Journalistic talent was arrayed to match, with substantial newsrooms in every city.

Digital changed that. I took a look recently at the job openings posted on JournalismJobs.com to see how many of them were based in New York City or Washington, D.C. Among television jobs, only 8.8 percent were based in one of the two news capitals. (The states with the most jobs? Texas, Ohio, and Florida.) For newspaper jobs, that number was 10.5 percent. But among digital media and startup jobs, nearly 4 in 10 — 38.9 percent — were located in New York, D.C., or their suburbs.

When The Washington Post looked at Bureau of Labor Statistics data last year, it found that the share of American reporting jobs that were in New York, Washington, and Los Angeles went from 1 in 8 in 2004 to 1 in 5 in 2014.

Think of the most prominent digital-native news companies, like Vice Media, BuzzFeed, Business Insider, Gawker Media, Mashable, Vox Media — all of them are in New York or D.C. (Vice adds a sort of geographic diversity by being in Brooklyn instead of Manhattan, I suppose. But you could still visit a dozen of them without your Uber bill climbing too high.) There are smaller hubs in the Bay Area (for tech reporting), Los Angeles (all about video), and even Miami (for Spanish-language and Hispanic-targeting media), but the increase in concentration is unmistakable. Journalism jobs are leaving the middle of the country and heading for the coasts.

This won’t come as a shock to anyone who’s studied cluster theory, the idea that industries naturally tend toward concentration in one or a few places — think autos in Detroit, oil in Houston, or music in Nashville. Small geographic advantages start to snowball; companies that want to work with the big players naturally want to be near them, and talented people know that, if they want to do interesting work, they’d best go where the innovation is happening. The dispersed nature of the 20th-century American news business was the exception, not the rule; the world is spiky, not flat.

“These highly creative, high-velocity, high-metabolism industries really gravitate to really big cities,” Florida told me. “It’s where other companies are, it’s where the talent is, it’s where the innovation is happening the fastest.”

An unrepresentative media

So if the news business is becoming even more centered in New York, what sort of impacts would that have on our news?

For one thing, you’d expect it to make the media more liberal — culturally and politically. Journalists don’t like it when conservatives point out that they, as a group, lean farther left than the country as a whole. But you don’t need to be a conspiracy theorist to believe it: College-educated liberal arts grads who live in cities — a group most American journalists fit into — are more liberal as a group than the American median. And those who live in New York or San Francisco are going to be more liberal as a group than those in Cincinnati or Knoxville.

You can argue about whether that’s a good thing or a bad thing. But one element of Donald Trump’s rise is a backlash against the sort of cultural cosmopolitanism that lots of people who’ve never taken the Acela feel is on the rise. When trust in the media is at an all-time low, this shift could make it harder to bridge those divides.

Clustering could also change the sort of people who can enter the business. Some people don’t want to live in New York. Twenty years ago, they might have taken a good job at a newspaper in Indiana and built a career there. Those jobs aren’t all gone yet, but a lot of the ambitious college journalists I talk to these days are all about landing something, anything that’ll get them in at the ground floor of an outlet in New York.

Does this trendline only move in one direction, toward greater concentration? There are a few forces working the other way. Tools like Slack make it easier for individual journalists to work outside the main office, which has let some reporters for whom location is unimportant to their work get out of the city. One could imagine a point at which the cost of living in New York becomes too limiting a factor for growth and smaller cities become more appealing. (A version of this is happening now in the tech industry, where the insane cost of housing in Silicon Valley is helping cities like Austin and Seattle attract workers and companies.)

But the larger shift will be with us for years to come. There’s little reason to believe news jobs will ever again be distributed as evenly around the country as they were a decade ago — the market forces are too strong. It’ll increasingly be up to non-market forces — nonprofits and public media — to fill the local voids.

You see this pattern over and over again in digital news: What was once pitched as opening up a space has led instead to greater concentration of power in the hands of a few. The web and linking were supposed to expand routes of distribution — but over time, we’ve seen that power become clustered in Facebook, Google, Twitter, and a handful of other tech companies. Online advertising was supposed to let a thousand media flowers bloom, supporting independents and small, high-quality publishers. Instead, it’s led to a generation of digital publishers — all those guys in New York — chasing scale and racing to be as big as possible.

America is a big, highly distributed place. Our democracy is structured around cities and counties and congressional districts and states. Our media used to be too. As an industry, it’s our responsibility to make sure we don’t become too myopically focused on a few square blocks in Midtown Manhattan.