Fixing the ticker-tape problem, or the disconnect between how we write and how we read
Between the tedious wash steps of the experiment I’m running, I’ve been tinkering with Python. The result is aiRead.
aiRead integrates the ideas about active reading I’ve accumulated over the last four years. Although its ChatGPT integration is its most powerful feature, this comment is about an insight I’ve gleaned by using its ticker-tape display feature.
Mostly, I sit down at my desk to read articles on my computer screen. I click a link, and what appears is a column of heavy-looking blocks of text. This column descends down into the depths, below the bottom edge of my browser.
With aiRead, a delicate line of characters traces its way across the screen. As they tiptoe forward, my brain switches on. I have time to imagine. To critique. To notice my confusion.
Having spooned up a bite-size morsel of text, aiRead waits for me to decide on what to do. Do I move on to the next paragraph? Do I rewind, or read it again? The software nudges me to make deliberate choices about what to read next, based on my reaction to what I just read. As a result, I feel more interested, and come away with a better understanding of other people’s ideas.
Most people’s writing benefits. We write in ticker-tape fashion, with a single line of words unspooling from our minds onto the page. Only later do we reshape the material into sections and move around our language by the paragraph. Despite this, we typically read in blocks. It takes effort to tune into the melody-like flow of note-like micro-ideas. aiRead fixes the disconnect between how writers write and readers read.
Yet not all writing suffers from this problem. When I read Scott Alexander, it’s easy for me to follow the continuous flow of his ideas and language.
After reflecting on the opening paragraphs of his classic essay, I Can Tolerate Anything Except The Outgroup, I notice three main features that allow his writing to solve this “ticker-tape problem” I have described, making it much easier to read his material carefully.
First of all, he goes beyond perfunctory examples, using elaborate stories to illustrate his ideas. Where another author would eschew an example, he uses one. Where they would write a single sentence, he gives us a paragraph. Where they would introduce their essay with a quote from Howl, he gives us Meditations on Moloch.
This brings us to the second key difference that makes his writing unusually easy to follow. Because he uses such elaborate stories, he can compose his sentences as a series of images. Each image is vivid, and my mind enjoys holding onto it for a moment. Then, I let the image of the beloved nobleman collapse, and raise the image of him murdering his good-for-nothing brother in a duel.
Finally, Scott chooses sonorous language in which to express these images. He doesn’t describe “a noble who everybody in town loved kills is brother in a duel decades ago.” He says “a beloved nobleman who murdered his good-for-nothing brother in a duel thirty years ago.” I relish his language, and naturally take the time to savor it.
Great writing, like Scott’s doesn’t need a software application like aiRead to ensure careful and attentive reading. Other writing benefits greatly from a ticker-tape display. I am surprised, and pleased, to find that a long and careful investigation into the details of reading has also produced valuable insights about the act of writing.
Fixing the ticker-tape problem, or the disconnect between how we write and how we read
Between the tedious wash steps of the experiment I’m running, I’ve been tinkering with Python. The result is aiRead.
aiRead integrates the ideas about active reading I’ve accumulated over the last four years. Although its ChatGPT integration is its most powerful feature, this comment is about an insight I’ve gleaned by using its ticker-tape display feature.
Mostly, I sit down at my desk to read articles on my computer screen. I click a link, and what appears is a column of heavy-looking blocks of text. This column descends down into the depths, below the bottom edge of my browser.
With aiRead, a delicate line of characters traces its way across the screen. As they tiptoe forward, my brain switches on. I have time to imagine. To critique. To notice my confusion.
Having spooned up a bite-size morsel of text, aiRead waits for me to decide on what to do. Do I move on to the next paragraph? Do I rewind, or read it again? The software nudges me to make deliberate choices about what to read next, based on my reaction to what I just read. As a result, I feel more interested, and come away with a better understanding of other people’s ideas.
Most people’s writing benefits. We write in ticker-tape fashion, with a single line of words unspooling from our minds onto the page. Only later do we reshape the material into sections and move around our language by the paragraph. Despite this, we typically read in blocks. It takes effort to tune into the melody-like flow of note-like micro-ideas. aiRead fixes the disconnect between how writers write and readers read.
Yet not all writing suffers from this problem. When I read Scott Alexander, it’s easy for me to follow the continuous flow of his ideas and language.
After reflecting on the opening paragraphs of his classic essay, I Can Tolerate Anything Except The Outgroup, I notice three main features that allow his writing to solve this “ticker-tape problem” I have described, making it much easier to read his material carefully.
First of all, he goes beyond perfunctory examples, using elaborate stories to illustrate his ideas. Where another author would eschew an example, he uses one. Where they would write a single sentence, he gives us a paragraph. Where they would introduce their essay with a quote from Howl, he gives us Meditations on Moloch.
This brings us to the second key difference that makes his writing unusually easy to follow. Because he uses such elaborate stories, he can compose his sentences as a series of images. Each image is vivid, and my mind enjoys holding onto it for a moment. Then, I let the image of the beloved nobleman collapse, and raise the image of him murdering his good-for-nothing brother in a duel.
Finally, Scott chooses sonorous language in which to express these images. He doesn’t describe “a noble who everybody in town loved kills is brother in a duel decades ago.” He says “a beloved nobleman who murdered his good-for-nothing brother in a duel thirty years ago.” I relish his language, and naturally take the time to savor it.
Great writing, like Scott’s doesn’t need a software application like aiRead to ensure careful and attentive reading. Other writing benefits greatly from a ticker-tape display. I am surprised, and pleased, to find that a long and careful investigation into the details of reading has also produced valuable insights about the act of writing.