Your local coffee shop probably has AI brain

We can't make meaning from everything in the universe

A new coffee shop just opened around the corner from me. Its walls are pleasantly speckled with raw plaster and it’s already become the coffee shop of choice for the neighborhood’s young, beautiful people, who fill it daily despite the fact that there appears to be just one small couch and one four-top in the store’s otherwise empty thousand-plus square feet of prime central Brooklyn real estate.

There are already four other coffee shops within a two-block radius of this one. There are no pharmacies and no decent bagels or slices within a mile (don‘t @ me, you're wrong). But this new coffee shop does offer something the others don't. It is the first of the five that hawks “storytelling” as its flagship product.

Storytelling, gathering, curation, immersive, intentional design, meaningful exchange, design-forward, beautifully crafted, authentic connection – these terms, and more, can be found in the marketing gloss my wife discovered printed on heavy textured paper and slapped on the raw concrete entry table (why is there an entry table?) when she got curious and decided to pay the shop a visit.

These words obviously came out of an LLM, which is why they have precisely zero meaning. None of them were intentional, curated, or authentic. If they tell a story, it is of a marketing guru who outsourced their job to a next-token predictor. These words amount to nothing because they could describe any Brooklyn coffee shop in the year of our lord 2026. That’s why an LLM can write them, because so many already have. They are so devoid of context and specificity they could describe almost any Brooklyn business. In fact, I think I've seen them used recently to promote my local meat counter. You know the one, the old-school deli that recently got taken over by the nice, but slightly odd failson who has a MOOC certificate in digital marketing — from Stanford, dad!

In fact, here you go. I have thoughtfully remixed the coffee shop’s description to fit the old-school deli (this is the actual text, with just a few carefully curated replacements):

JERRY’S MEAT HOLE is a design-forward meat counter nestled in Bed-Stuy, Brooklyn, where thoughtfully curated bulk meat meets beautifully crafted specialty cuts of assorted meat product.

Rooted in creativity and connection, Jerry’s Meat Hole is designed as a gathering space-one that celebrates storytelling, intentional design, and meaningful exchange. We partner with factory farms and meat distributors to create moments that feel personal, immersive, and community-driven.

Inspired by the ease and intimacy of a lived-in loft apartment, our space offers a refined yet welcoming backdrop for meat producers to connect with meat consumers in an authentic way. Here, ideas are shared, stories are experienced, and visions are brought to life. Book your next event or apply to sell your meat in our store by emailing [email protected]

I don’t know anything about the provenance of the coffee shop. I don’t know the people behind it or the kind of coffee they serve and whether it is, indeed, beautifully crafted. I’ve never set foot in there and hope I never will. But I know exactly how this coffee shop happened, beyond a shadow of a doubt. I know, conclusively, that they used AI to generate everything about the shop first — the concept, the design, the menu, the business model — and then tried, not unsuccessfully, to make what emerged from that black box a reality. The same way the hapless third-generation owner of Jerry’s Meat Hole saw an insipid OpenAI ad on the subway and decided to use AI to give their 74 year-old family business a glowup it does not need and cannot not possibly live up to.

I know that this is how this coffee shop came into being because I have seen so many people succumb to AI brain, and this is what it does. It reverses causality. Instead of extracting meaning from information, the way it’s always been done, your brain on AI says that if you supply the meaning, AI can efficiently search all the information in the known universe to find a set of facts that justifies it. This is backwards! This is wrong! This is exactly what is happening all around you, every day, probably even to you and by you.

People will argue that search did the same thing before AI. And libraries before it. And stone tablets before that. Anyone who's ever wanted to justify a bad idea has always been able to find a way to do it, they'll say. Which, sure, except, no, these people are wrong. Search forces you, the searcher, to sift through pre-filtered information to arrive at a your own conclusion, however wrong. It doesn't just serve you a plausible answer on a plate. Our brains work differently when presented with two different technologies with two completely different interfaces. Are we really having this debate?

Search brain gets you gimmicky SEO-driven flavored coffee drinks. AI brain gets you coffee shop concept stores with five seats and no bathroom and descriptions that read like performance art in neighborhoods that desperately need fewer coffee shops and more pharmacies. AI brain gets you meat counters that have decided they can use AI to fake being trendy. Someone said “generate the business plan of the perfect coffee shop for this industrial ground-floor space in Bed-Stuy” or "yo i am the boss of my family's dumb old meat store i kno nothing about meat but i am a marketing wizard wut do i do to make it cool?" and ChatGPT just turned around and answered because that’s what it’s programmed to do. It’s not programmed to wonder if a coffee shop is the right kind of shop for a specific location. It’s not programmed to consider whether a meat counter needs a trendy makeover. It’s not programmed to stop and ask if it has crammed so many buzzwords into a shop’s (I’m going to assume they call it an) "artist’s statement" that the entire enterprise is at risk of collapsing under its own weight. The only thing it is programmed to respond with is, “sure, I can find the information to make sense of your absurd request.”

So, now we have another coffee shop across the street from our four other coffee shops. And it’s constantly full of beautiful people wondering why its owners couldn’t thoughtfully curate any more places to park their beautiful bodies. And the meat shop where we used to get perfectly good meat has raised its prices and hired a social media manager and there are lines around the corner now for new daily “drops“ that I’m pretty sure are just the same average meat they’ve always offered, repackaged for a era of insanity.

The other day, while I was out taking a stroller walk with my son, some tourists from Los Angeles eating sliced salami out of brown waxed paper wrapped with twine stopped me on my street to ask what my favorite local coffee shop was. When I told them it was the one on my corner with slightly too acidic coffee and slightly too acerbic staff, they looked it up and asked me, “what about that concept store next door? ChatGPT says we might like it better.” I shrugged and went home to put my son to bed.

Only some of these things are literally true. But all of them could easily be true, because I made up the untrue stuff all by myself, using my mind‘s built-in capacity to stretch everyday experiences into meaningful anecdotes that help me make sense of this silly world, even if the events they’re based on aren’t, technically speaking, entirely real.

Contrast that with the coffee shop. It is real in the sense that you can, I assume, buy and, if you're lucky, find a seat to consume coffee there. But it is in no way a place where ideas are shared, stories are experienced (what?), or visions are brought to life, despite the fact that a chatbot with access to all the world’s information thought those would be cute things to say about a coffee shop (why the passive voice, though???). AI has its productive uses. But its most common use, justifying bad ideas and decisions with meaningless drivel, is silly and dangerous.

Don’t let your hard-won ability to make up useful lies atrophy just because some machine tells you it has access to everything there is to know. It is precisely that unfiltered access to everything that makes it unable to truly know anything. And it is our inability to access everything, our need to filter what we experience to make sense of it at all with our limited hardware, that makes us capable of making meaning out of anything. Cede that gift to the machines and you become them: capable of doing anything, unable to understand why you're doing it. There is no coffee good enough to justify becoming that, no matter how beautifully crafted.

don't subscribe

Careful: if you subscribe, critical and independent thinking about media, technology, politics, and society (and the occassional non sequitor about bread) will be sent to your inbox.
[email protected]
Subscribe