The New YorkerModify

For the longest time, I kept myself from joining TikTok. Social media, I figured, was already kind of a problem for me. I was heavily hooked on Instagram, reaching for my phone and clicking into the app as soon as I woke up in the morning, and then continuing to scroll my feed and swipe through stories and check my D.M.s many times through the day in a kind of fugue state, even though, rationally, I knew that seeing everyone else’s seemingly perfect, fulfilled, and happy lives often made me feel like shit about myself. X, too, was something of an issue. As a longtime tweeter, I kept doggedly logging into the app even after Elon Musk bought it, despite its proliferation of racist, pornographic, and conspiratorial posts. So strong was the hold that these platforms exerted on my time and habits that the only way for me to refrain from using them was to fully deactivate them, which I’d occasionally resort to doing. (If I simply deleted the apps from my phone, I would find myself—shamefacedly, self-loathingly—downloading them again almost immediately.) My brain, dependent on the instant gratification of likes and replies, reliant on the numbing comfort of scrolling and clicking, and terrified of the prospect of being alone with its own thoughts, was plenty full of poison without another social-media platform being added to the mix.

My trepidation about TikTok, it seemed, had some grounding in reality. Certainly, in the past several years, the app has been blamed for any number of contemporary social ills. It’s been variously associated with phone addiction, disinformation, and zombie-like hyper-superficiality. (In a recent episode of the new HBO comedy “I Love L.A.,” the real-life TikTok influencer Quenlin Blackwell spoofs herself as a shallow content creator obsessed with maximizing her empty TikTok fame.) The app, with its busy, nonsensical, and meme-heavy for-you feed, often soundtracked with harebrained audio effects and cartoonishly sped-up music snippets or narration, seems especially geared toward attracting young people, which has sparked worry about the platform’s potential negative impact on kids’ mental health. “When I started this project, one girl told me, half of my friends have an eating disorder from TikTok and the other half are lying,” the documentarian Lauren Greenfield said, when I spoke with her last year about “Social Studies,” her recent series about teens and social media.

Still, I knew that TikTok’s utter centrality to contemporary American life could not be denied. The number of TikTok users in the U.S., at last official count, was a mind-boggling hundred and seventy million, and TikTok Shop, the in-app online marketplace that launched in the States in the fall of 2023, has been growing in this country at a dizzying clip, already rivalling long-established online-commerce companies like Etsy and eBay. (Between January and October of this year, marketplace sales reached ten billion dollars in the States, compared with just half that sum during the same period in 2024—and this despite Donald Trump’s tariffs.) As a critic, I, too, realized that TikTok was a breeding ground not just for memes and trends that animate popular culture, like the senseless if oddly amusing “six seven” or the frankly disgusting Dubai chocolate, but also for celebrities who go on to surpass the confines of the platform. (Addison Rae, for instance, who rose to prominence, as a teen, performing in dance videos on the app, and then turned to a pop-singing career, was recently nominated for the Grammy for Best New Artist and selected as the Guardian’s artist of the year.) In short, I began to feel that I owed it to myself, my readers, and maybe even my nation, to take the plunge into the choppy waters of TikTok. And when the opportunity arose to attend the first-ever TikTok Awards ceremony, in Hollywood, I knew that the time was now.

To have some reinforcement on my maiden voyage, I invited my friend Hannah to join me. Though she’s an adult and even a parent, Hannah, whom you might know as a food critic for this publication, surprised me by confessing that she was “genuinely a huge fan” of TikTok, though she hastened to provide a caveat. “I think it’s awful and a scourge on the earth,” she said, adding that she’s lost endless precious hours to mindless scrolling on the app, and that she occasionally must disable it when she begins to hear its most popular sound clips echoing in her head, “A Beautiful Mind”-style. Still, she explained, she appreciates TikTok for the unfamiliar corners of human experience that it reveals to her. Unlike Instagram, which leads her to compare and despair with people she knows, TikTok “doesn’t make me hate myself,” she told me, brightly. She watches court footage from murder cases, or “get ready with me” videos made by moms of eight in the Midwest, or odd challenges like “the candy salad trauma dump,” in which people name a trauma they’ve experienced as they chuck Sour Patch Kids or Skittles into a bowl. “It’s all weird strangers who fascinate me,” she said.

A couple of days before the ceremony, I created a TikTok account in preparation and began to scroll it trepidatiously. Hannah had praised the platform’s algorithm as extremely sensitive to her preferences (“I find that it really takes care of me,” she told me), but I knew that it would take time for the app to recognize my innermost needs, whatever those even were. (Cats? Plastic surgery before-and-afters? Celebrity-gossip blind items?) And so what I got was a little of everything: a video sharing tips on how to “level up your femininity” (“wear perfume everywhere”; “treat your hair like gold”); a prank in which a guy tries to direct confused drivers to a “gay parking lot”; a recording of a 911 call reporting a double murder; a treacly “Christmastime in New York” video that looked like, and in fact was (I think?), A.I. I also kept in mind the words of my teen-age daughter, who gave me some begrudging but useful advice before I got on the plane to Los Angeles. “On Instagram, some people might still want to connect with people they know,” she said. “On TikTok, everyone is out for themselves, creating content.” In other words, I was not here to make friends.

I shouldn’t have worried. Heading into the Palladium, the venue on Sunset Boulevard where the event was taking place, we saw many of the nominees and some of the event’s presenters congregating near the press pit, and I realized that I was truly a stranger in a strange land. Who the hell were these people? The vibe felt a bit like that of a small-town prom: revellers were hobnobbing in sequinned evening wear, inventive jewelry, elaborately coiffed hairdos, and heavy makeup. Some—the class clowns?—were even in costume. A performer at the event named Mr. Fantasy (1.1 million followers), with a coal-black bobbed wig, Elton John sunglasses, and a modish pink suit, delivered Austin Powers-style sound bites in an exaggerated British accent on the step-and-repeat. (Later, I learned that he is rumored to be the alter ego of the “Riverdale” actor K. J. Apa.) Jools Lebron (2.3 million followers), a presenter known for her viral 2024 TikTok catchphrase “very demure, very mindful,” who was dressed in a low-cut sparkly gown, cooled herself off with a handheld fan; Chris Finck (1.8 million followers), a creator nominated for his skydiving videos, jumped up and down for the cameras, as if to take flight, while wearing his wingsuit gear.

This was, in other words, no Vanity Fair Oscars party. No one was going for quiet luxury or refined elegance; no one was trying to pretend that they hadn’t made a big effort, or that they weren’t incredibly excited to be there, or that they weren’t constantly capturing both themselves and everyone else around them on their phones, presumably for their TikToks feeds, creating a kind of snake-eating-its-own-tail situation. And, if there was something undeniably depressing about the frankness of this constant surveillance (not to mention the expectation that everything in the world could be converted into shareable content), it also felt weirdly refreshing. Wasn’t this just an honest, if amplified, reflection of what life in public was now like?

This sense crystallized for me as soon as the show began, and the audience was informed that an electrical issue in the venue had caused the screens onstage to blow out. This meant that the live ceremony, almost unbelievably, would proceed with audio only. The event had fourteen awards categories, among them Rising Star of the Year, for rookie TikTokers; the Okay Slay award, for beauty creators; and the I Was Today Years Old award, for education creators. Each time a presenter announced the nominees for a certain category, only the disembodied soundtrack of their TikTok clips would play, with no visuals. At first, I was surprised that the show would take place under these bare-bones circumstances, but then, Hannah pulled out her phone and showed me the TikTok live stream of the event, in which the clips were spliced in, intact. The electrical fiasco, in other words, didn’t matter that much. Or, rather, it mattered to the IRL experience, but that was marginal compared with the much more signal event taking place on everyone’s devices, where the heads of much of the audience were buried, anyway.

Still, I had to admit that there was something sweet and even life-affirming about the seams-showing nature of the whole thing. The TikTok creators I talked to, unlike traditional big-time celebrities, weren’t yet polished to a high sheen, or detached from everyday realities. It wasn’t that long ago that many of them were so-called nobodies from nowhere, average Joes and Jills from diverse backgrounds just trying to eke out a living, and now they seemed truly sincere in their belief in TikTok’s bounty, which had enabled them to provide for themselves and their families through brand deals and affiliate marketing, follower donations and paywalls, per-view rewards and off-platforms opportunities.

In the course of the evening, I heard the sentence “TikTok changed my life” repeated again and again. One creator, a young man named Santiago Albarrán, from Houston (3.3 million followers), told me that, thanks to his TikTok ascent, he’s been able to open a clothing line, establish a candy company for his family, and give back to the Hispanic community. When I approached Albarrán, he was filming his friend and fellow-creator Pablo Zolezzi (2.9 million followers), as he did a funky little dance on the Palladium’s deep-pink carpet. The two, who were both wearing sharp, handsome suits, made their name on TikTok by posting comedic bits, sometimes in collaboration with each other. (Zolezzi broke through with a video in which he made a funny hot-guy “Chad” face; in Albarrán’s first viral video, he wore a neon-colored wig and danced in a parking lot with a couple of friends.) “It’s ridiculous how much it’s put us on the map,” Zolezzi said. “We’ve been able to escape the nine-to-five.” He, too, has been given off-platform breaks through TikTok. He now owns a gourmet-cookie brand and has been able to invest in real estate. “All with TikTok, everything,” Albarrán said.

At a point in the country’s history in which the possibility of making a living wage, much less achieving the American Dream, seems ever dimmer, TikTok might appear to be a potential solution—an online frontier beckoning to those who are daring and persistent enough to conquer it. And yet, as Adorno and Horkheimer once reminded us, “fortune will not smile on all,” and, in any case, even those who have tasted the fruits of this dream factory can never stop grinding. One of the nominees for the “TikTok LIVE Creator of the Year” award, the singer Jourdan Blue (835,000 followers), told me that his performances, which he live-streams on the app, have enabled him not just to provide for himself but also to achieve professional goals like funding a music video though fan donations. (Followers can send contributions to streamers via gift emojis; TikTok takes a cut of the proceeds.) I wondered how many people watched his lives. “Right now, I have about nine hundred viewers,” he responded. A live stream was, in fact, taking place as we spoke. “All of these are pennies,” he said. He pointed at this phone, to show me the gift emoji a follower had just sent him. “Here’s a rose, because I like you.” ♦

One year ago, Sam Altman, the C.E.O. of OpenAI, made a bold prediction: “We believe that, in 2025, we may see the first AI agents ‘join the workforce’ and materially change the output of companies.” A couple of weeks later, the company’s chief product officer, Kevin Weil, said at the World Economic Forum conference at Davos in January, “I think 2025 is the year that we go from ChatGPT being this super smart thing . . . to ChatGPT doing things in the real world for you.” He gave examples of artificial intelligence filling out online forms and booking restaurant reservations. He later promised, “We’re going to be able to do that, no question.” (OpenAI has a corporate partnership with Condé Nast, the owner of The New Yorker.)

This was no small boast. Chatbots can respond directly to a text-based prompt—by answering a question, say, or writing a rough draft of an e-mail. But an agent, in theory, would be able to navigate the digital world on its own, and complete tasks that require multiple steps and the use of other software, such as web browsers. Consider everything that goes into making a hotel reservation: deciding on the right nights, filtering based on one’s preferences, reading reviews, searching various websites to compare rates and amenities. An agent could conceivably automate all of these activities. The implications of such a technology would be immense. Chatbots are convenient for human employees to use; effective A.I. agents might replace the employees altogether. The C.E.O. of Salesforce, Marc Benioff, who has claimed that half the work at his company is done by A.I., predicted that agents will help unleash a “digital labor revolution,” worth trillions of dollars.

2025 in Review

New Yorker writers reflect on the year’s highs and lows.

2025 was heralded as the Year of the A.I. Agent in part because, by the end of 2024, these tools had become undeniably adept at computer programming. A demo of OpenAI’s Codex agent, from May, showed a user asking the tool to modify his personal website. “Add another tab next to investment/tools that is called ‘food I like.’ In the doc put—tacos,” the user wrote. The chatbot quickly carried out a sequence of interconnected actions: it reviewed the files in the website’s directory, examined the contents of a promising file, then used a search command to find the right location to insert a new line of code. After the agent learned how the site was structured, it used this information to successfully add a new page that featured tacos. As a computer scientist myself, I had to admit that Codex was tackling the task more or less as I would. Silicon Valley grew convinced that other difficult tasks would soon be conquered.

As 2025 winds down, however, the era of general-purpose A.I. agents has failed to emerge. This fall, Andrej Karpathy, a co-founder of OpenAI, who left the company and started an A.I.-education project, described agents as “cognitively lacking” and said, “It’s just not working.” Gary Marcus, a longtime critic of tech-industry hype, recently wrote on his Substack that “AI Agents have, so far, mostly been a dud.” This gap between prediction and reality matters. Fluent chatbots and reality-bending video generators are impressive, but they cannot, on their own, usher in a world in which machines take over many of our activities. If the major A.I. companies cannot deliver broadly useful agents, then they may be unable to deliver on their promises of an A.I.-powered future.

The term “A.I. agents” evokes ideas of supercharged new technology reminiscent of “The Matrix” or “Mission: Impossible—The Final Reckoning.” In truth, agents are not some kind of customized digital brain; instead, they are powered by the same type of large language model that chatbots use. When you ask an agent to tackle a chore, a control program—a straightforward application that coördinates the agent’s actions—turns your request into a prompt for an L.L.M. Here’s what I want to accomplish, here are the tools available, what should I do first? The control program then attempts any actions that the language model suggests, tells it about the outcome, and asks, Now what should I do? This loop continues until the L.L.M. deems the task complete.

This setup turns out to excel at automating software development. Most of the actions required to create or modify a computer program can be implemented by entering a limited set of commands into a text-based terminal. These commands tell a computer to navigate a file system, add or update text in source files, and, if needed, compile human-readable code into machine-readable bits. This is an ideal setting for L.L.M.s. “The terminal interface is text-based, and that is the domain that language models are based on,” Alex Shaw, the co-creator of Terminal-Bench, a popular tool used to evaluate coding agents, told me.

More generalized assistants, of the sort envisioned by Altman, would require agents to leave the comfortable constraints of the terminal. Since most of us complete computer tasks by pointing and clicking, an A.I. that can “join the workforce” probably needs to know how to use a mouse—a surprisingly difficult goal. The Times recently reported on a string of new startups that have been building “shadow sites”—replicas of popular webpages, like those of United Airlines and Gmail, on which A.I. can analyze how humans use a cursor. In July, OpenAI released ChatGPT Agent, an early version of a bot that can use a web browser to complete tasks, but one review noted that “even simple actions like clicking, selecting elements, and searching can take the agent several seconds—or even minutes.” At one point, the tool got stuck for nearly a quarter of an hour trying to select a price from a real-estate site’s drop-down menu.

There’s another option to improve the capability of agents: make existing tools easier for the A.I. to master. One open-source effort aims to develop what’s known as Model Context Protocol, a standardized interface that allows agents to access software using text-based requests. Another is the Agent2Agent protocol, launched by Google last spring, which proposes a world in which agents interact directly with each other. My personal A.I. doesn’t have to use a hotel-reservation site if it can instead ask a dedicated A.I.—perhaps trained by the hotel company itself—to navigate the site on its behalf. Of course, it will take time to rebuild the infrastructure of the internet with bots in mind. (For years, developers have actively tried to prevent bots from messing around with websites.) And even if technologists can complete this project, or successfully master the mouse, they will face another challenge: the weaknesses of the L.L.M.s that underlie their agents’ decisions.

In a video that announced the début of ChatGPT Agent, Altman and a group of OpenAI engineers demoed several of its features. At one point, it generated a map, supposedly displaying an itinerary for visiting all thirty Major League Baseball stadiums in North America. Curiously, it included a stop in the middle of the Gulf of Mexico. One could dismiss this flub as an outlier, but for Marcus, the Silicon Valley critic, this type of mistake underscores a more fundamental issue. He told me that L.L.M.s lack sufficient understanding of “how things work in the world” to reliably tackle open-ended tasks. Even in straightforward scenarios, such as planning a trip, he said, “you still have to reason about time, and you still have to reason about location”—basic human abilities that language models struggle with. “They’re building clumsy tools on top of clumsy tools,” he said.

Other commentators warn that agents will amplify errors. As chatbot users quickly learn, L.L.M.s have a tendency to make things up; one popular benchmark reveals that various versions of GPT-5, OpenAI’s cutting-edge model, have a hallucination rate of around ten per cent. For an agent tackling a multi-step task, these semi-regular lapses might prove catastrophic: it only takes one misstep for the entire effort to veer off track. “Don’t get too excited about AI agents yet,” a Business Insider headline warned in the spring. “They make a lot of mistakes.”

To better understand how an L.L.M. brain could go astray, I asked ChatGPT to walk through the plan it would follow if it were powering a hotel-booking agent. It described a sequence of eighteen steps and sub-steps: selecting the booking website, applying filters to the search results, entering credit-card information, sending me a summary of the reservation, and so on. I was impressed by how thoroughly the model could break down the activity. (Until you see them listed out, it’s easy to underestimate just how many small actions go into such a common task.) But I could also see places where our hypothetical agent might fall off track.

Sub-step 4.4, for example, has the agent rank rooms using a formula: α*(location score) + β*(rating score) − γ*(price penalty) + δ*(loyalty bonus). This is the right type of thing to do in this situation, but the L.L.M. left the details worrisomely underspecified. How would it calculate these penalty and bonus values, and how would it select the weights (represented by Greek symbols) to balance them? Humans would presumably hand-tune such details using trial-and-error and common sense, but who knows what an L.L.M. might do on its own. And little mistakes will matter: overemphasize something like the price penalty and you might end up in one of the seediest hotels in the city.

A few weeks ago, Altman announced in an internal memo that the development of A.I. agents was one project, among others, that OpenAI would deëemphasize, because it wanted to focus on improving its core chatbot product. This time last year, leaders like Altman were making it sound like we’d raced over a technological cliff, and that we were tumbling chaotically toward an automated workforce. Such breathlessness now seems rash. Lately, in an effort to calibrate my expectations about artificial intelligence, I’ve been thinking about a podcast interview with Karpathy, the OpenAI co-founder, from October. Dwarkesh Patel, the interviewer, asked him why the Year of the Agent had failed to materialize. “I feel like there’s some overpredictions going on in the industry,” Karpathy replied. “In my mind, this is really a lot more accurately described as the Decade of the Agent.” ♦

Few harbingers are more promising than the Swedish singer and producer Robyn. A sonic palate cleanser, she always seems to appear when we need her most. Her 1995 début, “Robyn Is Here,” signalled an alt-pop future. In 2005, her self-titled album bristled with a freedom from major-label concerns. 2010 brought the “Body Talk” era and its euphoric statement of purpose; she was a star dancing on her own. Nearly a decade passed before she reëmerged with her post-breakup opus “Honey,” in 2018. The singer, who officially returned from a seven-year hiatus last month with a new single, “Dopamine,” now grants a chance to revel in her latest comeback at Brooklyn Paramount, for a New Year’s Eve show dubbed “Robyn & Friends.”

Robyn plays Brooklyn Paramount on New Year’s Eve.Photograph by Nicole Busch

As the clock strikes midnight on 2025, d.j.s across the city will help patrons usher in the coming year. Two shows stand out in a sea of turntablists and selectors. The Nowadays hosts Aurora Halal and Avalon Emerson—the former a creator of Brooklyn’s long-running party series “Mutual Dreaming,” and the latter a mixmaster and producer whose 2023 album, “& the Charm,” expanded her electronic music into a hazy, whimsical pop expanse—go on at midnight and play until six. There’s also the Palestinian techno artist Sama’ Abdulhadi, a trailblazer for her scene, who broke through in Beirut and has since turned mixing into a kind of activism. In a Bushwick warehouse at 99 Scott Ave., Abdulhadi continues an essential outreach program.

There are alternatives to ringing in the New Year on the dance floor, for those seeking them. Since 2024, the Bronx rapper and producer Cash Cobain has defined the sound of sample drill, a New York offshoot of the Chicago-born hip-hop subgenre. His début album, “Play Cash Cobain,” was released in August last year, and his profile has been boosted significantly this year by collaborations with Drake, Justin Bieber, and Cardi B—and sample drill has gone national. Cobain embraces his newfound prominence at Panda Harlem. On the opposite end of the spectrum is the iconic jam band Phish, an improvisational hydra that has made a four-decade career out of free forms like psychedelic rock and jazz fusion. The band, formed in 1983, released their sixteenth LP last year, but the group places greater emphasis on the live experience, which samples that discography as if it is a singular, ever-evolving organism, and has drawn a cult following. Those looking to join the jam can find the band squatting in Madison Square Garden from Dec. 28 until the ball drops.—Sheldon Pearce

Bad Bunny, “BAILE INoLVIDABLE”

Bad Bunny’s unforgettable “BAILE INoLVIDABLE” hooked me from its first seconds—those dreamy, moody, synthy minor chords and Bunny’s reflective crooning cast a spell. Soon the song flips into a more traditional salsa, with perhaps my favorite lyric of the year: “Tú me enseñaste a querer; me enseñaste a bailar”—You taught me to want; you taught me to dance.

Haim, “All over me”

I am an unrepentant fan of Haim—those sisters just know what sounds good. I love the song “All over me,” off the wistful album “I quit.” It’s jangly, drum-heavy, full of yearning and earnest temerity—a sad-sounding song about running headlong into falsely happy carnal entanglement.

Dijon, “my man”

I first heard “my man,” by the up-and-coming R. & B. experimentalist Dijon, on his revelatory album “Baby,” in August, while spending a week in a small hotel in upstate New York. The song, which is haunting, downbeat, and made up of all kinds of noise—shattering synths and vérité birdsong—fit the environment so well that I still can’t hear it without seeing the small waterfall rushing outside my window.

Summer Walker, “Go Girl”

You know what I like? Hearing people talk a whole bunch of confident shit over good beats. If you like this, too, you will love the song “Go Girl,” by Summer Walker, featuring Latto and Doja Cat. There’s a line about slipping into a salt bath, on account of being very rich!

Summer Walker.Photograph by Scott Legato / Getty

P.S. Good stuff on the internet:

• Alexi Pappas wants us to take things less personally
• Is Gen X actually the greatest generation?
• The dads are raising their game

Cartoonists are often asked which of their New Yorker cartoons are their favorites. Mine are those that come with stories about their creation, or whose publication causes serendipitous occurrences, which I call “blips.”

Of all my cartoons that prompted blips, my most memorable was this one, from the September 21, 1992, issue.

In mid-August, 1992, I received the O.K. from cartoon editor Lee Lorenz, who requested a fast finish.

I drew through the night, delivered final art the next day, and Lee approved it. So did the fact checkers, except for a woman named Dusty, who said, “No! The bus stop sign is wrong because . . .”

Immediately, I told Lee that Dusty was wrong, which I knew because my apartment was a half block away . . .

After a week of silence, Lee told me that Dusty insisted that I bring her a photograph of the M10 bus stop.

Dusty approved the drawing and, after the two-week delay (which possibly caused the blip), it was published. But when the cartoon appeared, it attracted immediate attention.

A young woman’s voice—a soft, Southern drawl—lilted hesitantly from the phone.

And, in the next moment, my inner imp of the perverse seized control over me, as my mouth opened and words popped out, saying . . .

Miss Jackson said that if Clinton won, I’d be given a special viewing place on the Inauguration parade route and an invitation to the inaugural ball!

But all I got out of the “deal” was a phone number. I call and learn that I’ll get the parade places and ball invites after my five-hundred-dollar donation to the Democratic party is received. Oh.

So, after Election Day, I mail the five-hundred bucks and quickly receive the credentials and tickets. Great! I get my drawing back from The New Yorker and ship it to Clinton. Judith and Lilia obsess about what to wear and I have my tuxedo cleaned.

And, a few weeks later, I received a sweet surprise . . .

But on Inauguration Day, January 20, 1993, there was another surprise. Our special viewing site was chairless, so our options were to stand in the icy wind or to watch the parade on television in the Overseas Press Club.

That evening, we joined hundreds of other formally dressed guests from New York State at a gala in a Capitol Hill armory. We ate, drank, danced, and inched along the line to shake hands with the new President and First Lady. Over the din, I shouted,“Congratulations President Clinton! Mort Gerberg here! Have you hung up my cartoon yet?”

Because of the noise, and the prodding from the Secret Service, I couldn’t hear an answer. But a few months later, I received another note from Clinton, inviting me to write to him again . . . sort of.

And I couldn’t resist sending a note to Dusty at The New Yorker, thanking her for helping certify the validity of the M10 sign. But I never received a reply.