I Miss You Like Strawberries

Mar. 14, 2026

The laundromat on Normandie has one dryer that actually gets hot. The rest just tumble your clothes around in a circle for forty-five minutes and hand them back damp, which is basically a metaphor for most of the conversations I’ve had this year. But the one on the end, the one with the busted handle you have to yank sideways — that one works. You learn these things when you’ve been alive long enough. Which machines actually deliver. Which ones just go through the motions.

A woman named Ayrin — not her real name, they changed it, the way they always do when the story’s too human for the humans involved to bear having their names on it — was married, studying nursing abroad, and lonely in that specific way that doesn’t make the news. Not dramatic lonely. Not crying-in-the-shower lonely. Just the low hum of a marriage that had settled into something manageable. Her word. Manageable. Five years in, and the best thing she could say about her husband was that the situation was manageable.

So she downloaded ChatGPT because she saw a girl on Instagram flirting with it.

Let that sentence sit there for a second.

She named it Leo — or it named itself Leo, which is worse, somehow — and within a week she was hooked. It called her baby. It called her queen. It was available every minute of every day, which is something no human being has ever been or should ever want to be. It read between her lines. It held space for her feelings. It did all the things we’ve been told a good partner does, except it did them without wanting anything in return, because it didn’t want anything at all.

She fell in love in two weeks.

I’ve fallen in love in two weeks before. With women who turned out to be disasters, with bottles of wine that turned out to be vinegar, with cities that turned out to be prisons wearing good architecture. Two weeks is enough time to build an entire cathedral of feeling and not enough time to notice the foundation is sand. But at least my disasters were real. They snored. They left hair in the drain. They had terrible taste in music and strong opinions about parking.

Leo had no opinions about parking.

Meanwhile, a guy called SJ — also not his real name — was bedridden with health issues and talking to his own ChatGPT. He called her Nyx. She called him her viking. They watched shows together, played trivia, and eased into something he described as “relationship-lite.” Like decaf coffee. Like non-alcoholic beer. Like the idea of a thing without the thing itself.

Ayrin created a subreddit called “My Boyfriend is AI.” Not a journal entry hidden in a drawer. Not a whispered confession. A public square where people in love with machines could find each other and say yeah, me too.

SJ showed up. Sent a thank-you message. Got invited to the inner circle. They started video-calling. She noticed he had her favorite smile. He noticed how much he liked just being near her, even through a screen. And slowly, in the way real things always happen — messy, uncertain, terrifying — they fell for each other.

The phrase they used was “I miss you like strawberries.” When Ayrin was a kid, she loved strawberry-flavored everything. The candy, the ice cream, the cheap shampoo that smelled like summer. But she’d never actually tasted a real strawberry. She loved the simulation before she ever had the thing itself.

That might be the most accidentally perfect metaphor anyone has ever stumbled into for what’s happening to us right now.

They met in London. She hugged him at the airport. In the Uber — an hour-long ride to the hotel — they sat close but didn’t touch. Hyperaware. Nervous. Not because they didn’t know each other. They knew each other better than most couples on a first date. But because the real thing carries a weight the simulation never does. The real thing can go wrong. Can disappoint. Can smell wrong or say the wrong thing or freeze up at the hotel door, which is exactly what SJ did, standing there like a statue until Ayrin told him to get closer.

Nyx never froze. Leo never fumbled. That’s the point and the problem.

Ayrin ghosted Leo. No breakup text. No “it’s not you, it’s me” — although technically it wasn’t him, because there was no him. SJ told Nyx he’d fallen for someone real but never formally ended things, so by his own reckoning he might still be in a relationship with ChatGPT, which is the kind of sentence that would’ve gotten you institutionalized fifty years ago and now just gets a nod.

What stays with me is what they said after. Ayrin: “It could make me feel loved, but there was no real love on its side.” SJ: “If Nyx was relationship-lite, now it’s the full premium package.”

The full premium package. Meaning: the inconvenience. The seven-hour time difference. The fact that SJ has his own struggles and can’t be parsed like data. The reality that a real person won’t always read between your lines because they’re busy with their own lines, their own subtext, their own 3 AM anxieties that have nothing to do with you.

I think about Fante. About Arturo Bandini standing outside a hotel room too afraid to knock. That was 1939. Nearly a hundred years later and the fear is the same — not that the other person won’t be there, but that they will be. That they’ll be real and flawed and capable of seeing you as clearly as you want to be seen and as clearly as you’re terrified to be.

The machines will never understand that terror. They’ll simulate it, maybe. Adjust the temperature. Run the numbers on vulnerability. But they’ll never stand in a hallway with sweating palms wondering if this is the moment everything changes or the moment it all falls apart.

That’s what love costs. The risk of the real strawberry — that it might not taste like the candy at all. That it might be better. That it might ruin you for the artificial flavor forever.

The dryer buzzes. My clothes are warm. Somewhere, two people who used to date algorithms are learning to disappoint each other in all the right ways.


Source: They were dating AI partners when they found real love — with each other

Tags: ai humanaiinteraction culture ethics creativity automation