If you haven’t already heard of Friend, the company that makes a $129 wearable AI companion—a plastic disk, containing a microphone, on a necklace—you probably also have not seen Friend’s recent ad campaign. Late this past summer, Friend paid $1 million to plaster more than 10,000 white posters throughout the New York City subway system with messages such as I’ll binge the entire series with you.
People hate these billboards. Revile them, even. Across the city, the ads are covered in graffiti criticizing the pendant (it doesn’t have eyes, bruh; CRINGE) as well as the idea of AI altogether (AI wouldn’t care if you lived or died); some vandals invite you to befriend a senior citizen instead of a chatbot, or volunteer with a community garden—you will meet cool people! Many of the ads have been ripped and torn. The backlash has grabbed far more attention than the product itself, so I wondered: How does Avi Schiffmann, the 22-year-old founder and CEO of Friend, feel about being the most despised tech founder in America’s largest city?
To my surprise, he was visiting New York from San Francisco when I reached out to ask about this. He told me that he was in fact in the city to see his vandalized billboards—and he was game to meet me last Wednesday in the West 4th Street station, where he’d purchased a prominent array of Friend ads in two long entry corridors. That morning, every single Friend.com ad I’d seen in the station had been scribbled over, but only a few hours later, they had all been replaced with new posters. Still, a few were freshly vandalized; when we approached one that said Fuck AI!, Schiffmann, with a Friend device dangling over his black T-shirt, said, “I love it.”
As Schiffmann tells it, the backlash was all part of the plan. The ads were meant to work as a canvas and provocation, he told me, because traditional marketing is passé: “Nothing is sacred anymore, and everything is ironic.” (He’s made the same point on X and in an interview with Politico.) To get attention, you need to be “a little on the nose,” he told me, and the images of vandalized Friend ads circulating the web are the best PR that Friend could ask for. “The picture of the billboard is the billboard,” Schiffmann said (also recently posted to X). Some of the ads implying that an AI is superior to a human friend—I’ll never bail on dinner plans, I’ll never leave dirty dishes in the sink—are clearly meant to goad. In fact, many of the posters, my colleagues and I have noticed, seem to be marked with verbatim messages in similar handwriting; had Schiffmann not only courted the vandalism but also instigated it? He denied any meddling: “Then I wouldn’t enjoy it that much.”
Friend is Schiffmann’s first foray into the AI industry, although he has experience building viral software. When the coronavirus pandemic began, and Schiffmann was still in high school, he rose to fame after creating one of the world’s most popular websites for tracking COVID-19 cases; the project was lauded by Anthony Fauci. When Russia invaded Ukraine, just months after Schiffmann had dropped out of Harvard, he created a website to match Ukrainian refugees with hosts. In 2023, his attention turned from crisis response to start-up mode (or perhaps the loneliness epidemic), and he began developing the Friend, then known as “Tab,” which he described at the time as a “wearable mom.”
Friend debuted in July 2024 with a promotional video that features brief clips of young adults navigating the world with a prototype pendant around their neck. In the final scene, two teenagers sit on a rooftop, apparently on a date. “I just kind of like to come up here to be myself. I’ve never brought anybody else—I mean, besides her,” the girl says, gesturing to her pendant. “I guess I must be doing something right, then,” the boy responds. In a time when the world seems to have agreed that Facebook, Instagram, and the social-media era have inflicted anxiety and loneliness on generations of adolescents and young adults, it’s hard to see the video as anything other than satire or tone-deaf.
Perhaps it is both. Schiffmann told me that he doesn’t think the company’s vision is dystopian or that AI companionship will degrade human friendships. “I don’t think this kind of ‘friend’ replaces any relationship in your life,” he said; rather, it provides a new category altogether. Schiffmann likened his AI pendant to a therapist, a best friend, and a living journal all at once. Seated on a bench in Washington Square Park, near the West 4th station—we had fled to avoid some overly loud busking—he paused, contemplating whether to continue. “This is what I said a while ago, and I don’t think a lot of people liked it,” he began, “but I would say that the closest relationship this is equivalent to is talking to a god.”
I was taken aback, though not terribly surprised; Schiffmann had indeed made the same analogy when Friend launched last year. There are so many clearly well-documented problems with AI companions—they confidently present false information as true, may push people toward mental-health crises or even suicide, flirt with children. “For an AI relationship to be real,” Schiffmann told me when I objected, “I think it has to have the possibility to lead you astray.” He likened the situation to replacing human drivers with self-driving cars, which still get into accidents but less frequently than people do. (This was confusing: Schiffmann had just told me that AI pendants will not replace human relationships.) There’s “a lot of responsibility,” he continued, but he was confident that it would work out, in part because the AI pendant, by virtue of being trained on all of the internet, has “read every book on how to be a good friend.”
Friend extends the generative-AI paradigm that ChatGPT sparked nearly three years ago: Algorithms whose ability to talk lucidly about anything, anytime, makes it easy to assign them magical and terrifying properties. As with ChatGPT at its launch, Friend has some serious flaws—reviewers have called it “an incredibly antisocial device” and “unattractive, and clunky to use”—and like OpenAI, the company has spent a lot of money without any immediate hope of making it back. Schiffmann has raised a few million dollars—$1.8 million of which was used to buy the URL “Friend.com”—but only about 1,000 Friend pendants have been activated. By Schiffmann’s own admission, the pendant has “plenty of issues,” and he does not yet know how to make the business profitable; running the AI model constantly is expensive, but he has no intention of adding a subscription fee. He did say that he’d like to have Friend pendants in Walmart next year.
For now, he’s prioritizing what he calls “mindshare”: to have as many people as possible thinking about, hating on, and discussing his product. As he tells it, all of this will jam into the zeitgeist the controversial notion that AI can be a “friend,” just as ChatGPT cultivated and became synonymous with the allure of chatbots. Friend also has ads all over Los Angeles, and Schiffmann said that Chicago is next. He also said that the company is working on a “feature film” about Friend, although he gave no other details. I could see why he was so game to meet with me and stand in front of one of his posters, on which someone had crossed out almost every word and declared, in red Sharpie, that a friend is A PERSON. As he leaned back to pose, someone passed by and offered a fist bump. “I have no idea who that was,” Schiffmann chuckled.
As I listened to his ideas, I kept coming back to Schiffmann’s observation that “everything is ironic.” Throughout the AI boom, picking apart sincere statements from hyperbolic PR, or just plain trolling, has become harder and harder. When Sam Altman, the CEO of OpenAI, says he wants to build a gigawatt of AI infrastructure every week—a data center that uses as much electricity as a major American city—it is both ridiculous and completely serious. He’s capturing mindshare and receiving funding for these efforts, in spite of a lack of clarity about how generative AI will make money or truly serve society. When Anthropic CEO Dario Amodei warns that AI models could replace half of white-collar jobs in a few years, even as his own company keeps marketing those very AI models, he sounds at once grave, naive, and absurd. To outright market an AI “friend,” rather than the more measured “companion” or “assistant” or chatbot, is to play with that confusion head-on.
A microphone in a plastic disk on a necklace connected to a chatbot is not a god, but Altman and Amodei both have declared that they are racing to usher in a sort of superintelligence. In a way, Schiffmann has simply said aloud the truth of many AI leaders’ grand vision. Meanwhile, the people defacing Friend’s advertisements are expressing a much larger, inchoate rage at the broader AI industry, not just these plastic pendants that practically nobody owns. Schiffmann has created spaces throughout the city for millions of New Yorkers to provide their own “social commentary on the topic,” as he put it, and for that commentary to then circulate on the World Wide Web.
Schiffmann told me that he was inspired by The Gates, an art installation of more than 7,000 orange steel gates along paths in Central Park that attracted tourists from around the world. Friend’s ads can provide a place to “see what the world thinks about AI,” he said, which apparently is “fuck this slop.” Indeed, Schiffmann was more prone to citing postmodernist aphorisms and artists than famous venture capitalists and tech founders. Of late, he told me, he has been pondering a quote attributed to Andy Warhol: “You have to be alone to develop all the idiosyncrasies that make a person interesting.” Warhol, of course, is known for at once satirizing and embodying mass production through his art and his studio, the Factory. Friend and its advertisements, at the moment, can be better understood as installation art than as a business, a performance instead of a product—an attempt to prod public attitudes toward AI, but perhaps not direct them.
The post A Walk With New York’s Most Hated Tech Founder appeared first on The Atlantic.
