September 9, 2024
A link roundup of articles and thoughts on practices of reading, why ChatGPT is still extraordinarily limited at writing, and the “missing questions” in our conversations about the future.
As always, I generate summaries using AI and edit those summaries for accuracy and usefulness. Then, I offer some thoughts of my own.
Oscar Schwartz’s article “Against Rereading” reflects on the tension between the profound, transformative experience of reading a book for the first time and the widely celebrated practice of rereading. While literary greats like Nabokov and Barthes advocate for rereading as a way to deepen understanding and even produce new meanings, Schwartz remains skeptical. He argues that the magic of first reading is rooted in its immediacy, intensity, and emotional resonance, which subsequent readings might erode. Ultimately, Schwartz suggests that rereading can be a conservative act, tied to tradition, and at odds with the fleeting, yet enriching, experience of encountering a text for the first time.
I am a chronic repeater. In large part, this is a symptom of being, for lack of a better term, a control freak. I remember once reading that Immanuel Kant lived each of his days exactly the same, down to his walking patterns in his house, and that this disciplined existence is associated with his extremely influential philosophical output. This was in college, and I decided that was how I wanted to live (what the fuck) and made myself a rigid schedule where every moment of my time was accounted for. My friend looked at the schedule and asked when I would have time to go to the bathroom.
I still fetishize the idea of that sort of life but have come to accept that it is impossible. The paradox, however, is that I have always been convinced that rereading, or re-consuming (TV shows, films, etc.), the same texts is generally inferior to reading as diversely and widely as one possibly can. This, again, is a reaction to my tendency to read and watch the same things over and over again when left to my own devices.
With that in mind, this article really resonated with me. I am persuaded both by Schwartz’s argument against the unquestioned supremacy of rereading. I am also very impressed by the depth with which he explores the argument in favor of rereading, as well. Ultimately, the following passage reflects my own attitude:
In the end, maybe the crucial difference between those who read once and those who reread is an attitude toward time, or more precisely, death. The most obvious argument against rereading is, of course, that there just isn’t enough time. It makes no sense to luxuriate in Flaubert’s physiognomic details over and over again, unless you think you’re going to live forever. For those who do not reread, a book is like a little life. When it ends, it dies—or it lives on, imperfectly and embellished, in your memories. There is a sense of loss in this death, but also pleasure. Or as the French might put it, la petite mort.
Again, I’m a control freak.
The article discusses an experiment by a writer hired to produce a 200-word article about a new coffee shop. After completing the task, the writer wondered if AI tools like ChatGPT could deliver a similar product. To test this, the writer entered the same press release into ChatGPT and compared the AI-generated copy to their own work. ChatGPT’s output was slick but filled with clichés and lacked meaningful content or human appeal. The key critique was that while ChatGPT can generate grammatically correct and polished text, it struggles to understand what engages readers on a human level​.
A nice, short, and shareable article on why ChatGPT cannot replace writing. I like its emphasis on how context and meaning can be so easily lost in generated text. I maintain that generative AI can be useful as a tool (see these link roundups, of course) but not as a replacement for writing. Writing, after all, is definitionally a form of human thinking.
Benedict Evans’s article discusses how technological predictions often ask the wrong questions. He illustrates this with examples of forecasts from the mid-20th century that missed key developments, like the internet, focusing instead on less relevant issues. Evans emphasizes that predicting future tech trends involves more than guessing advancements—it requires understanding underlying platforms and frameworks. He also reflects on the importance of questioning assumptions about what comes next in technology and the potential for unforeseen breakthroughs.
When I was a youthful undergraduate scholar I was afforded the opportunity to attend an academic conference for honors students across universities. I only remember a few moments from it, but one of them comes back to me with surprising frequency. I would guess that I think about it a few times a month.
Another honors student was giving a presentation on science fiction and its social commentary. It was, even for an undergraduate honors student discussing science fiction at a conference, incredibly techno-optimistic. His essential point was that we can look to histories of science fiction to historicize our understandings of technology in the present. I was too shy during the Q&A, but afterward I approached him and asked (some version of): “Isn’t there a possibility that science fiction could normalize understandings of technology? And couldn’t those normalizations be harmful, wrong, or a combination of both?”
It’s hard to capture this and difficult to not sound like I am overstating, but the presenter’s response seemed extraordinarily similar to what I imagine it would look like if I had asked a question in a foreign language. For him, science fiction simply charted how technological progress “actually is.” And technological progress, of course, is good. Keep in mind that this was probably 2011-2012, a time period of enormous techno-optimism. These years and the ones immediately following were sort of the inverse of the “shot, chaser” adage. Look at all of these wonderful new ways to participate in society. (Chaser.) Look at what people, largely, are choosing to do with that… (The most disgusting shot imaginable.)
I’m rambling. My loose point is that this story resonates with what Evans is talking about here. I may be picking on the student, but truthfull we all extrapolate current visions of technology into the future in one way or another. I’m persuaded by Evans’s call to consider the unasked questions when we do so.