The Best Book I've Read This Year
Is that incentive enough
That’s one way to intro a book: it’s the best thing I’ve read this year. Also the most exhilarating, whimsical, and insightful book I’ve read this year. I could say “just trust me,” as I often do, and hope you’ll see part of why in the conversation below. But if you want to read something that will reorganize the way you think about the way AI, the performance of self, and just the internet broadly in this moment, something propulsive and unexpected, — Vauhini Vara’s Searches is that book.
Some of the thinking in Searches stems from Vara’s viral essay, “Ghosts,” published in The Believer in 2023. After her sister’s death, she couldn’t figure out how to write about her. So she asked Chat-GPT — then called GPT-3 — to write it for her. The resultant hybrid is remarkable: unsettling and beautiful. But what is it? Whose is it? Who is it for? That’s the sort of experiment (and resultant questions) that guide Searches. If you want something to rattle around in your brain for weeks, months, years, again — this is that book.
You open the book with a description of your first encounter with an AOL chat room after moving to a suburb of Oklahoma City in the early ‘90s — and your unspoken desire to understand what it took to be, well, hot. As an elder millennial I’ve read so many descriptions of similar experiences amongst our micro-generation (my partner, hilariously now a tech journalist at The Atlantic, was convinced he was messaging with Adam Sandler for months) but yours was such a wryly hilarious set-up for this insight:
“Chat rooms had prepared us for what the internet was becoming — a place where, anonymized and disembodied, we could seek answers to our truest questions.”
I’m hoping we can use that as a set-up for the premise of the book, too: the internet as a place of searching. Can you elaborate a bit more on how you arrived at that framework? How does it allow us to approach the internet’s expanse differently? What does it allow us to see?
Oh my god, I loved the the line about your partner, which excavated a long-lost similar memory of my own, having to do with a friend getting the email address of Nick Hexum, the lead singer of the band 311 (I don’t need friends that act like foes, ‘cause I’m Nick Hexum, the one who knows about things! — remember that?), and sharing it with me but asking me not to abuse it, at which point I promptly abused it, losing that friend’s friendship forever.
I loved how the concept of searching set up for me the central question the book wanted to wrangle with: How is our desire of knowledge — of the self, of others, of the universe — both fulfilled and exploited by big technology companies? The human desire for knowledge is so fundamental: It’s what has made us so capable of exploration and so capable of exploitation. For me, that’s the subject of this book.
One of the many scratchy delights of this book is the way you lay out your own arguments about why we shouldn’t buy from Amazon — how you arrive at them, how Amazon affects its workers, how it’s reshaped the social and physical character of Seattle — alongside your purchases (with their real, listed, often ridiculous names), others’ reviews of them, and your own internal/external monologue about why you needed to buy from Amazon even though you try not to buy from Amazon.
This reminds me of a point in one of my books where I tried to list all the emotions I actually felt while scrolling Instagram — mortifying, but also clarifying. (I get a similar feeling whenever I’m asked to look at Google Search History, or someone else is asked to do something similar on a podcast — something you also do in the book).
How did you think about the purpose of this section, and why did it seem important to put the clash between “trying to be the person I want to be in the world” and our hunger for ease and immediacy on the page? What’s still mortifying?
I wanted to make the case for examining our own agency in big technology companies’ exploitation of us and of the world. In this chapter made up of my Amazon reviews, I start out wanting, in good faith, to pull away from my use of it — and I end up using it more than ever. The point here isn’t to show how great Amazon turns out to be; it’s actually to position myself as a villain — which felt really uncomfortable, but also really important, because I thought if I did it right, it would allow others to see their villainy and all of us to see our collective villainy. [AHP note: I suggest reading this section before you start arguing about Amazon in the comments]
This felt connected to me to the centrality of searching to the human experience. I wanted to move toward asking how else we might deploy our curiosity, beyond how we’ve deployed it to date. If human desire — and, therefore, curiosity — is a given, and if we have agency in how we deploy our curiosity, what are some of the better, more generative, places it could take us?
Every few chapters, you feed the text into Chat GPT and ask for its thoughts, which return to you in cheery summary format. I always feel deeply annoyed but also deeply owned by these summaries — I feel like I’m reading a 9th grade book report telling me how dull and predictable I am. I feel transparent, but maybe the better word is hollowed out. What feelings did they evoke in you — and how do you conceive of their narrative purpose in the text as a whole?
I didn’t expect ChatGPT to give me insights into my own writing, and it didn’t. I did expect it to give me insights into how it functions — specifically, into its own use of rhetoric, as a product conceived by an investor-funded technology company — and it absolutely did that. I’d argue that ChatGPT’s text production is inherently meaningless as a form of communication — that is, as literature — which means that any meaning that its presence has in the book has to be created by my interactions with it and by readers’ engagement with it.
I knew readers would either either notice or not notice all its rhetorical tricks (summarizing my writing in a way that describes my writing as more positive about big technology companies than it actually is; using what appears to be bland or generic language to in fact reinscribe hegemonic values; and so on), and I wanted those segments of the book to function narratively whether readers noticed those tricks or not.
If they did notice those tricks, I was hoping to derive tension from the reader’s curiosity — or annoyance! — about the apparent failure of the narrator (that is, me!) to call ChatGPT out in the context of the conversation itself, especially early on. If they didn’t notice those tricks, I was hoping to derive tension from the reader’s curiosity about why the narrator is including all this junky ChatGPT-produced material in the first place: What’s her — my — end game here?
I don’t want to give spoilers, but hopefully by the time the reader gets to the end of this exchange, their understanding of it is different from how they understood it at the beginning.
I am obsessed with the chapter you wrote in Spanish (with Google Translation in English alongside) about the process of trying to become fluent in Spanish during your husband’s sabbatical in Spain, compelled, at least in part, by your shame at how simple, passive, rudimentary and altogether at odds with your self-image you came across as a novice Spanish speaker.
I feel like I’m trying to describe how a joke works, but the essay works (for me, as an English reader) in part because you write it in Spanish, and the Google Translation back into English creates all sorts of small flubs — the sort of thing that makes the reader think, ah, she’s just not quite there yet (or, secretly, not that smart). I felt this acutely when I studied in France: that tension between how you’re accustomed to being able to express yourself and the level at which you actually can. My French mother never knew I was a great writer, just one whose little notes (about when I’d be back for dinner) she’d gently and repeatedly correct.
At the end of the essay, you write: “Soon, I will translate this essay into English using Google Translate. I will look at the two versions – the Spanish version and the English version — together. I will immediately notice how odd the English version looks, how awkward to read, but I won’t have enough Spanish to figure out why. Only much later, after more Spanish study, will I realize that the situation is more complicated than I thought. The blame is shared. There will be sentences that I have written perfectly in Spanish, which Google Translate will have translated poorly. But there will be sentences that I have written horribly—changing “this” for “that,” as is my habit; inventing words that don’t exist; putting accents where they don’t belong—that Google translate will have translated into something much more comprehensible than it should be. In the end, I will understand that what I thought was a shortcut was not; to communicate, in Spanish, I had to learn Spanish. But is that all there is? If I want to be able to communicate to everyone, do I have no choice but to learn all the languages in the world?”
I’d love to hear you riff a bit on where you arrived at the end of this experiment — sometimes I find myself thinking back on that period in France and realizing there was tremendous value, being humbled that way, and that it prepared me to learn how to “speak” in so many more languages (on the internet, in newsletters like this one, in podcasts). And sometimes I’m just exhausted by the impossibility of communication, full stop.
Oh, I’m glad you’re obsessed with it — so am I! I wanted to be making a case in this book for human self-expression and communication as a way to connect with one another and to the universe around us, and it was important to me to define this self-expression and communication capaciously. What I mean is that I care as much about the kind of art that sits on museum walls or on bookstore shelves as I do about the kind of art that doesn’t — the language and images we all use all the time (and have been using for the past 100,000 years) as a path to communion.
This is maybe going to sound unrelated to the topic at hand — this specific chapter that I wrote in Spanish and translated into English using Google Translate — but part of what I wanted to do was to handicap my own ability to express myself by forcing myself to do it in a language in which I’m not entirely proficient, right? I wanted to make the case for the value of my effort to express myself even when I lack what would be described as the correct language to do it.
That’s the goal of the Spanish part on its own. Readers with Spanish fluency will recognize my clumsy grasping for language; they’ll recognize the inelegance and even incoherence of what I’m trying to get across. And yet I hope they’ll also recognize the beauty in the effort itself.
For those who also — or only — read in English, there’s this additional layer, having to do with what happens when Google Translate is asked to convert my Spanish into English. On the one hand, it smooths out a lot of my errors, making me sound more elegant and coherent than I was in Spanish; on the other hand, it botches some of my correct Spanish, turning it into inelegant and incoherent English. I have to admit it allows for cross-language understanding on one level — and yet, its ability to do is inextricable from its corruption of my self-expression, both when it’s making me sound good and when it’s making me sound bad.
In your chapter on the aftermath of “Ghosts,” you cite the headline of one of the hot takes it sparked: “Rather than Fear AI, Writers Should Learn to Collaborate With It.”
As a way of closing — I keep thinking of that word, collaborator, and its historical precedence, and the narrative justification in the moment, not dissimilar to what you lay out in the Amazon snippets. What does this sort of collaboration do to us, and our compasses of creativity?
So, I’m always interested in thinking about the rhetorical purposes of language, and that word, collaboration, is a really rich one to explore. What is a “collaborator,” and what is the rhetorical purpose of describing artificial intelligence as one? Well, Merriam-Webster defines a collaborator as “a person who collaborates with another” — the operative word here being “person.” To “collaborate” is “to work jointly with others or together especially in an intellectual endeavor.”
I write in the book about getting an email after publishing “Ghosts” from an OpenAI employee named Jay Dixit — someone whose job it is to talk to authors about using AI in their work — in which he used a variant on that term: “I think your story could be an inspiring example for other writers on how to use ChatGPT as a creative collaborator that works in service of the writer’s vision, using it not to generate copy but as a catalyst for their own creativity,” he wrote.
With big technology companies working hard to get us to use — and trust — their products, they have an interest in rhetoric that makes their products come across as not just human-like but as human-like in ways that support our goals; they’re leaning into this particularly hard with AI models that generate language. To call big technology companies’ AI products potential “collaborators” implies not only that they’re human-like but that they share our goals — they’re working jointly with us — and are capable of intellectual engagement. But in fact these companies’ products are built to serve the companies’ goals and are not capable of intellectual engagement.
What I’m interested in is interrogating the notion that collaboration with AI products owned by big technology companies is even possible in the creative sphere. To me, it seems definitionally impossible. ●
You can buy Searches here and wow do I recommend doing so.
If you liked that and want to keep stuff like this outside the paywall — consider becoming a paid subscriber.
Subscribers get access to the weekly discussion threads, which are so weirdly addictive, moving, and soothing. It’s also how you’ll get the Weekly Subscriber-Only Things I’ve Read and Loved Round-Up, including the Just Trust Me. Plus it’s a very simple way to show that you see the work that goes into creating this newsletter every week. If you’re already a paid subscriber: thank you so much for making this work sustainable.
