The new essay collection “Searches: Selfhood in the Digital Age,” by Vauhini Vara, opens with a transcript. “If I paste some writing here, can we talk about it?” Vara asks. Her interlocutor, the large language model ChatGPT, responds, “Of course!” The chatbot asks what specific themes it should focus on. “Nothing in particular,” Vara replies. “I’d love to just hear your reaction, if that’s OK?” This is, of course, O.K. with the chatbot. “I’m nervous,” she admits.
Vara, a novelist and tech journalist, began experimenting with A.I. products in early 2021. The previous year, OpenAI released GPT-3, a precursor to the company’s commercial chatbot, ChatGPT. Through an interface called Playground, GPT-3 could be given any text-based input—a sentence, a paragraph, a block of code—and, in turn, generate additional text to “complete” the prompt. Vara had been drawn to the technology after reading texts it had generated: some outputs were passable as human writing, and others were charmingly stilted, as with a Times “Modern Love” column that became trapped in a recursive loop: “We went out for dinner. We went out for drinks. We went out for dinner again. We went out for drinks again. We went out for dinner and drinks again.” (“I had never read such an apt Modern Love in my life,” she writes.)
At the time, Playground was invite-only; Vara, in her capacity as a reporter for the California Sunday magazine, had profiled Sam Altman, the C.E.O. of OpenAI, and she messaged him to ask if she might play around with it. (Vara, who got her start in media as a tech journalist for the Wall Street Journal, also previously worked as an editor at The New Yorker.) She was curious about how GPT-3 would respond to her fiction. At first, her experiments felt “illicit”: a dalliance with a technology that seemed to threaten her line of work. Vara was uninspired by GPT-3’s responses to her writing, but over time came to appreciate the algorithm’s ability to offer language where there was none. She enlisted it to help her write about the death of her older sister, Krishna, when the two were both in college. The story, a foundational one, had never come easily to her; perhaps GPT-3 could help.
The resulting essay, “Ghosts,” was published in August 2021, by The Believer. It is structured in nine movements—what Vara calls “stories.” For each, Vara supplied the opening sentences, or paragraphs; GPT-3 filled in the rest. The algorithm had moments of clarity, humor, and sadness. It was also prone to tropes and clichés. When Vara offered, simply, “My sister was diagnosed with Ewing sarcoma when I was in my freshman year of high school and she was in her junior year,” the response from GPT-3 read a bit like a college-application essay about overcoming hardship. “Eventually, she went into remission and got the all-clear and was able to play lacrosse with me for a season,” the program suggested, before concluding, heartbreakingly: “She’s doing great now.”
The essay went viral. It was a time of rising curiosity and anxiety about artificial intelligence; text and image generators were not yet in widespread circulation, and the essay seemed to capture the technology’s ambitions and shortcomings. With each iteration, Vara’s prompts grew longer, more detailed, and more personal. Later, she would describe this as an impulse to counteract the “falsehoods” generated by the L.L.M.—to assert herself, and the truth, against the statistically derived fabrications of the machine. Ultimately, GPT-3—which she had believed might “deliver words to any writer who has found herself at a loss for them”—seems to have been most useful to her in its capacity to elicit text, rather than supply it. In the ninth and final story, by far the most personal account of Vara’s grief, only the last two lines are GPT-3-generated. The regular chatbot dynamic had been inverted: GPT-3 was prompting Vara.
“Searches” belongs to a subgenre of memoir that I jokingly—and, as a contributor, self-deprecatingly—call “Me, Online.” The books and essays in this category tend to blend memoir, reportage, and criticism; they question the author’s own complicity in technological capitalism, while leaning on structural analysis to explain, and exonerate, both writer and reader. They are usually written by millennials, who witnessed the rise of the consumer internet when they were children; came of age in a world still only partially defined by digital technologies; and were educated—and scrolled Tumblr and Twitter—in an era of widely applied critical theory. (Ours is a zoom-in, close-read culture.) Memories of life before the internet are inextricable from memories of youth, so nostalgia comes easily. There is often a scene of the author waiting patiently, against the dial tone, to sign in to AOL. “The computer burst into a long, staticky screech, punctuated by a series of sharp beeps, as if the machine were hyperventilating,” Vara writes, of a childhood friend’s modem.
Vara is an appealing narrator—smart, funny, honest, and a little neurotic. She harbors ambivalence about her reliance on technology products, but doesn’t beat herself up about it. She is thoughtful but not too heady; principled but not preachy. The book’s theoretical touchstones include Shoshana Zuboff’s “The Age of Surveillance Capitalism” and Safiya Umoja Noble’s “Algorithms of Oppression.” These are invoked with a light touch. The questions Vara asks of technology and artificial intelligence are not terribly complex, but they are meaningful: What biases, and structural hierarchies, do L.L.M.s and text-to-image generators reproduce? How do digital technologies shape, expand, or pervert communication?
Where “Searches” departs is in its formal experimentation. The collection braids essays—some memoiristic and journalistic, others more playful—with transcripts of exchanges between Vara and ChatGPT. Many of the essays use as their structure, or raw material, digital forms and data sets that will be familiar to internet users in the twenty-first century. One chapter, “Elon Musk, Empire,” is an alphabetized list of “Interests,” intended for advertisers, associated with her X account: “Babies,” “Bolivia’s Government issue,” “Cuisines,” “Financial Services,” “India travel,” “Lover,” “Meta,” “Nursing & nurses,” and so on. Another chapter is composed of Vara’s Amazon reviews from 2021 to 2024—writeups of products like Lactaid, fig bars, and a guide to Disney World, all accompanied by slightly guilty disclaimers about why the items had to be purchased from Amazon. (“I checked out this book—the previous edition—from the library,” she writes, of the Disney guide. “The problem was that it came with a section at the end that you could get Disney characters to autograph, which we couldn’t do with the library copy.”)
About halfway through the book, Vara begins to riff on the tropes of the tech-founder investor pitch; she imagines pitching her own startup, which will provide a time-travelling vessel to a parallel universe inhabited by “all the girls and women who died prematurely in our universe,” including Krishna, and Vara’s aunt, who also died young. She does so in an essay titled “Resurrections,” which is structured like a slide deck and illustrated with images generated by OpenAI’s DALL-E 3 and GPT-4o, and Bing’s Image Creator. The catalyst for the essay was Vara’s realization that there weren’t many photographs of Krishna on the internet; and, because Krishna died in 2001, before digital photography became ubiquitous, her own digital footprint is nearly nonexistent. The gap between text and technology is one of yawning inadequacy, and the juxtaposition has a dark humor.
Still, the format quickly takes on the feeling of a joke that’s gone on too long. If the strength of “Ghosts” was that the computer-generated text provided structural support for charged material, the central flaw of “Searches” is that the collection inverts this balance: the material is too often in service of the structure. One essay, “I am Hungry to Talk”—a meditation on communication, translation, and language gaps—was originally written in Spanish, and appears in a two-column format, alongside an English version generated by Google Translate. It’s a nice concept, but the insights of the essay never quite rise to the occasion of the conceit. Another, “What Is It Like to Be Alive?” aggregates responses to a Google Forms survey that Vara created in 2023, asking women about their lives. (“What do you know about the lives of the people—maybe parents, maybe not—who raised you?”) The responses, submitted by a fairly arbitrary group of women, range from profound to half-hearted. In a book concerned with the collection, monetization, and homogenization of personal experience, the move toward something like a collective voice makes sense on a conceptual level. But in its literalism, the essay struggles to transcend the form.
The collection’s eponymous essay is a selection of Vara’s Google Search queries, organized by category (who, what, when, where, why, how). In an introductory essay preceding the piece, Vara writes that she found her Google Search history “unexpectedly moving”: “a comprehensive record of the previous decade of my life; it taught me about the person I’d been during each day of my existence.” The litany—with queries like “who is cardi b,” “what makes someone charismatic,” “what happens to syrian refugees who return,” “when to switch from car seat to booster seat,” “where is cantonese spoken,” “why are americans unhappy,” “how to warm up pork chop,” and “how to orgasm”—is, at turns, funny, charming, and banal. It feels both personal and not. In this respect, the chapter is representative of the book: technology is used to suggest intimacy, but instead sometimes operates like a scrim. The effect is intriguing; it’s not always clear if Vara offers her digital detritus in the spirit of disclosure or concealment. What emerges is less a personality than a sensibility—a portrait of an author searching for connection, and for herself.
Why is it so hard to transpose digital forms into literary ones? Done right, the results can be magnificent, magisterial: the blog-post comments in Dana Spiotta’s “Innocents and Others,” which move from gossipy to insightful to spammy; the sputtering, asynchronous Gchat conversation in Ben Lerner’s “Leaving the Atocha Station,” shaped by a shaky internet connection; the hypertext-like footnotes in David Foster Wallace’s work; Tony Tulathimutte’s “Our Dope Future,” a short story written in the form of a Reddit post. There are subtler approaches to depicting a world, or consciousness, infiltrated by digital networks: when a narrator in Joshua Cohen’s “Book of Numbers” is unable to access the internet, his narration—unaugmented by information gleaned from search engines and databases—loses some of its verve.
The most famous recent example might come from “A Visit from the Goon Squad,” Jennifer Egan’s prismatic, multi-narrative novel from 2010. Toward the end of the book, the pages go horizontal, to accommodate a chapter presented in the form of a seventy-five-page PowerPoint. The author of the presentation is a twelve-year-old girl, Alison. The content is sentimental—one slide is titled “Signs That Dad Isn’t Happy”—but it plays with, rather than against, the sterility of the software, and its inherent values of speed, simplicity, and linearity. Egan had been inspired by an article she read about the 2008 Obama campaign, in which a PowerPoint presentation was credited with turning the campaign’s fortunes around. She recalled thinking, “If PowerPoint has become that basic a form of communication, then I have to write some fiction in it.”
Egan’s use of PowerPoint operated on the levels of character and context. It was the medium of twenty-first-century office culture; futzing around in PowerPoint is exactly the sort of thing a computer-savvy tween might do to get her overworked mother’s attention. Just as the e-mails in Elif Batuman’s “The Idiot,” set in the mid-nineties, read and function differently from the e-mails in Sally Rooney’s more contemporary “Beautiful World, Where Are You,” the use of digital technologies in literature can, at its best, capture not only character but the broader cultural context.
Previous communications technologies—PowerPoint, Gchat, e-mail, blog comments—were designed to organize, and transmit, human-produced material. But how should a writer think about a technology that is itself a text-generator? In 2023, the writer Stephen Marche, under the pseudonym Aidan Marchine, published “Death of an Author,” a whodunit novella; the majority of the text was generated by L.L.M.s. The book is fine. (“Well, somebody was going to do it,” Dwight Garner wrote, in the Times. “If you squint, you can convince yourself you’re reading a real novel.”) That same year, Sheila Heti published a short story, “According to Alice,” in this magazine; it was written collaboratively with a customized chatbot. The story succeeds, in part, because it is not trying to pass: it leans into the strangeness of algorithmic text. The cadence is bizarre, freakish, funny. The language had been produced by algorithms, giving the final story something in common with found-text poetry, or Dadaist découpé. But the human author was a sort of conductor: prompting, arranging, intervening. Left to its own devices, A.I. remains a generator, not a writer.
Inside the OpenAI-DeepSeek Distillation Saga & Alibaba’s Most Powerful AI Model Qwen2.5-Max
De ChatGPT a mil millones de agentes
