News Diet

If news was a restaurant, what kind of restaurant would it be?

Would we be a neon-lit fast-food joint, a hushed Michelin-starred restaurant, an impossible-to-reserve omakase sushi counter, a friendly neighborhood watering hole with a familiar bartender, or an anonymous prepared food delivery service? And what does that mean for our mission and our sustainability? And more fundamentally, for who we are and what we do?

Bear with me. It may sound a little insane as an analogy, but, hey, you sat through an entire Taylor Swift-inspired song on news infrastructure; how much worse can this be?

(Well, OK, it is long. But there are payoffs, I swear.)

This all came to me the other day as I was musing about a smart Ezra Klein column in the New York Times — you should read it — about how AI might, in fact, not spark a global jobs apocalypse. He riffs off a post by U of Chicago economist Alex Imas that looks at how, as once-rare products become commodities, demand shifts to what is newly scarce. And in the context of AI systems increasingly automating work that people used to do, that means we’ll likely gravitate to products and services that feature the personal — human — touch. As Ezra notes:

The better my A.I. has gotten, the more I’ve wanted from the human beings around me — and from myself.

Which could be an argument — and is in fact the one that many journalists make, with perhaps more faith than evidence — that human-made journalism will be the lifeboat that keeps the industry (and us) afloat and even, perhaps, thriving, in a sea of efficient, machine-generated news.

Which brings us to food. If calories are scarce, you’ll eat anything and everything to stay alive. But once we have an abundance of things to eat, our choices shift, towards experience and connection rather than just the quality of the food. Do we want the fine-dining option, with the attentive waiters and snooty sommelier, or to outsource our choices to the famous sushi chef and whatever’s on the omakase menu that day, or pour our heart out to the bartender who knows exactly what kind of scotch you like (Islay, for the record)? The post by Imas is instructive: Starbucks rolled back automation and shifted back to handwritten names on cups as it realized the real business was in building more human connection with its customers. (Apparently misspelling and mispronouncing names can do that.)

And part of our choices are driven by other people’s choices. Imas spells this out:

René Girard called it mimetic desire: the idea that we don’t desire objects only for their intrinsic properties, but because other people desire them as well. We want what others want, and we want it even more when they can’t have it—for status, social capital, reputation, etc. Desire is not just a relationship between a person and an object; it is also a function of what other people desire.

We want it, not just because other people want it; we want it to show we can have it when other people can’t. We’ll want human connection because, in a machine-abundant world, human connection will be scarce. The value is in the scarcity. (Hang on to that thought; we’re coming back to it.)

There’s much to unpack in this analogy, but some parts are clear: High-end investigative work and deeply reported features are the luxury dining of the media ecosystem; trusted commentators and podcasters are the omakase chefs of the news world, and close community connections can feel like the bartender who listens to your woes.

But sometimes all we want or need are empty calories, and McDonald’s fills that gap admirably, as does a weekly delivery of frozen meals you can reheat. And we should admit that a fair amount of what we produce in the news world is not gourmet cooking, but somewhere between an energy bar at our desk and a forgettable snack when no one is looking — the content that’s most easily replaced by AI. And despite what Starbucks is doing with hand-written names on cups, no one is going to McDonald’s for the experience or human connection, any more than a flyby visitor to Apple News feels like they have a deeply personal relationship with the app.

Which is to say, what kind of restaurant are we? And how do we — assuming Imas is right — find a way to maximize human connection or provenance in our products? We don’t all have to be the same kind, of course, but we each have to pick a lane or two; offering both fine dining and take-out pizza doesn’t just risk confusing your customers — it confuses your staff, your mission, and eventually, yourself.

The high-end work maps neatly on to Tony Haile’s “archipelagos of trust,” and the omakase chefs are presumably the Fareed Zakarias and Ian Bremmers of this world, whose worldviews we’re interested in seeing. Are wire services the fast food providers? And if so, what’s a sustainable strategy for that business if the value in the ecosystem is found in the human touch? Are newsrooms that lean heavily into events building the deep relationships that will bring customers back even if information becomes more of a commodity?

And sure, the analogy, if it holds up, does point to some ways to emphasize the human touch, even if the actual production of news is increasingly mechanized: reporters could more regularly respond directly to emails and comments, newsrooms could publish more how-we-did-it explainers that let readers feel like they’re privy to the newsgathering process, and so on. We could emphasize our humanity more than our output.

But there are also some uncomfortable questions that the analysis surfaces.

If human connection is valuable because it’s scarce, then does its value fall if everyone can access it? Which is another way of saying, scale is at odds with scarcity. If everyone can afford a Birkin bag, is having one special? Can news with a personal touch command a premium if everyone can have it?

There’s an argument that it can: that AI will unlock a flood of as-yet unmet demand for news. Ezra cites another paper, by ASU professor Eldar Maksymov, about the “Jevons Paradox,” the phenomenon that when the cost of something falls, we don’t cut back our use of it — instead, demand explodes because we find new uses for the now-cheaper product. When the first spreadsheets came out, they didn’t put accountants out of work; the collapsing price of financial analysis unlocked a huge market that couldn’t previously be met because accountants were expensive. Is there a huge market for more personalized, more community focused news that isn’t currently met because the existing business model for news favors economies of scale? That’s certainly the argument that Shuwei Fang makes in her smart piece about the coming age of machine audiences.

Still, regardless of whether demand grows or not, the unsentimental economic analysis here is that people will value human connection because it’s rare; and if that’s true, then it implies that growing past the ability to meaningfully offer human connection to the communities we serve will start to diminish our value.

It’s not an argument not to do it; it’s an argument that the ecosystem is best served by multiple medium- and small-sized organizations, not giant national and international behemoths.

And there’s another point: In an economy driven by human connection, the quality of the product is secondary. Starbucks serves a coffee that is — honestly — indistinguishable from one you could get around the corner. What it’s selling, if it gets the new strategy right, is the feeling of connection. The friendly bartender who knows you like Islay scotch isn’t selling you a special brand of scotch you can’t buy from the corner liquor store; he’s selling you the service of pretending to be interested in your life story.

So what does that mean for newsrooms, who are focused on the quality of our output — our Pulitzer Prize-winning stories, our brand of scotch — when perhaps we might want to emulate the bartender more? Are we providing a product, or a service? And if it’s a service, is it personalization, engagement, or actually working on behalf of our readers (as I’ve suggested before)?

And here’s an even more uncomfortable thought: what if machines can fake the human touch? If there’s one thing LLMs do very well, it’s mimic how people talk and interact, except tirelessly and cheerfully. It’s certainly managed to convince Richard Dawkins that AI is conscious, and if it can win over a famously skeptical scientist, can the rest of us be far behind?

Which is to say, maybe even the human touch won’t be the moat we think it is. Or that the real value won’t be in the appearance of humanity, or the perception of community, but the proof that a real human was involved in the making of that piece of information: A sort of Fair Trade stamp that asserts that no humans were harmed in the creation of that story.

There’s a lot to chew on here; I have real questions about whether human connection is the killer app, even as I believe that increasing personalization and better understanding of users’ news needs and intentions will lead to better (and also worse) outcomes for our information commons.

Perhaps the analogy is a little stretched, or perhaps this AI revolution will be like nothing that has come before it. But if nothing else, all this provides — wait for it — food for thought.