Routine Writing Is About to Be Free
ChatGPT portends huge productivity increases and major disruptions in any enterprise in which writing matters. Also, animation..
Image of a late-night dorm room and a student with writer’s block, generated by Dall-E-2 using a prompt generated by ChatGPT.
I know two kinds of people: those who have been obsessively playing with and discussing ChatGPT and those who have at best a hazy notion that it exists. I’m in the obsessive group, as you already know if you read the Tennyson experiment I posted earlier.
For those in the hazy group, ChatGPT is a system that uses massive amounts of text to create a predictive model that enables it to mimic human writing. The shorthand is that it’s an AI chatbot, or autofill on steroids. You type in a request and it spits out an answer. This CNET column provides a solid backgrounder:
For example, you can ask it encyclopedia questions like, “Explaining Newton’s laws of motion.” You can tell it, “Write me a poem,” and when it does, say, “Now make it more exciting.” You ask it to write a computer program that'll show you all the different ways you can arrange the letters of a word.
Here’s the catch: ChatGPT doesn’t exactly know anything. It’s an AI that’s trained to recognize patterns in vast swaths of text harvested from the internet, then further trained with human assistance to deliver more useful, better dialog. The answers you get may sound plausible and even authoritative, but they might well be entirely wrong, as OpenAI warns.
Even in its current, relatively primitive form ChatGPT portends both huge productivity increases and major disruptions in any enterprise in which writing matters. Instead of writing boilerplate corporate memos, managers will soon assign them to bots. The run-of-the-mill college grads who get paid to flood my mailbox with press releases and promotional emails should start thinking about careers as nail techs or phlebotomists—something in the physical world. Insight and beauty are still rare, but serviceable prose isn’t.
With the right prompts, ChatGPT can already mimic routine political journalism and quotidian high school and college essays. “What GPT can produce right now is better than the large majority of writing seen by your average teacher or professor,” writes Daniel Herman, a humanities teacher at Maybeck High School a small independent school in Berkeley, in The Atlantic.
I asked the program to write me a playful, sophisticated, emotional 600-word college-admissions essay about how my experience volunteering at my local SPCA had prepared me for the academic rigor of Stanford. Here’s an excerpt from its response:
In addition to cleaning, I also had the opportunity to interact with the animals. I was amazed at the transformation I saw in some of the pets who had been neglected or abused. With patience and care, they blossomed into playful and affectionate companions who were eager to give and receive love. I was also able to witness firsthand the process of selecting the right pet for the right family. Although it was bittersweet to see some animals leave the shelter, I knew that they were going to a loving home, and that was the best thing for them.
The application essay is vapid but convincing. The variety of word choice (“blossomed,” “bittersweet”) and sentence structure marks it above average. “Had the opportunity to” is a stylistic tell: Here’s a privileged person who’s been taught to sound grateful rather than to write succinctly. “I was also able to…” is the same. I’m sure admissions officers see thousands of such essays every year. If their value goes to zero thanks to automation, this reader, writer, and teacher won’t object.
While crashing the value of mediocrity, ChatGPT could increase the returns to excellence. (“Average is over,” as Tyler Cowen put it.) Think about what happened to graphic design. Many people used to make a living doing routine tasks, from laying out pages to selecting typefaces, that are now easily handled by software. Thanks to the graphic intelligence embedded in everyday tools, the standards for routine graphics, from websites and PowerPoint presentations to restaurant menus and wedding invitations, have increased.
But that doesn’t mean there’s no work for graphic designers with the conceptual chops to take on complicated tasks. Powerful tools make iteration and brainstorming easier, but cleverness is still a valued skill. When my friend Shikha Dalmia launched The Unpopulist on Substack, she asked me to look at some logos she’d come up with using easily available tools. They weren’t terrible, but neither were they distinctive. “Hire a professional,” I advised, and she got a real logo.1
As I write, there are 28 student papers awaiting my grading attention. I doubt any used ChatGPT, partly because mentioning it in class produced mostly blank stares. (The most tuned-in student, however, said he’s started using it in place of Google.) Already, we’re getting confirmed reports of cheating on exams given on Canvas, the web-based system used by many colleges for assignments and grading. By next term, every class will have to take account of ChatGPT, either explicitly incorporating it as a starting point or going back to handwritten tests and essays.
The kind of mediocre writing that earns grade-inflated Bs is now replaceable by a bot. Maybe if those B-essay students started with AI-generated prose it would be easier to teach them to do better: to refine the ideas, dig down more on the facts, improve the writing style. Can ChatGPT be a time-saving tool, like a calculator or text search, rather than a threat?
Don Allen Stevenson speaking at the Atlantic’s Progress Summit.
One of the most interesting talks at The Atlantic’s recent Progress Summit—kudos for having it in L.A.—was Don Allen Stevenson III, a Dreamworks veteran now working as a self-employed “metaverse creator.” His talk started with a slide listing the many departments responsible for pieces of an animated film:
Storyboards
Character/Environment Art
Expression Sheets
Modeling reference
Surfacing textures
Rough Scene Layout
Keyframe animation poses
Crowds
Lighting/color
FX
Now AI can substitute for much of this background work. “All departments are kind of now text prompts,” he said. “And that’s the biggest shift that I've ever seen.” In the photo above, Stevenson shows how Dall-E 2 prompts can produce storyboards. (OpenAI’s Dall-E 2 is to art what its ChatGPT is to text.) The art is in response to the following prompts:
“An Anime mech robot sits in bed under a soft light and discovers the meaning of life, digital art.”
“A purple dinosaur wearing a pair of overalls sitting at the edge of a lake skipping a rock on the river emolga art, storyboarding”
“Two friends having an argument, anime style, manga story boarding”
If you don’t like these, you can ask for more. I used the same prompts to get different answers, including these.
AI-driven “sketching” allows a one-person shop to take on more ambitious projects. It lets creators go quickly from idea to concept art. It makes it cheap and easy to try out ideas that would be too costly and time-consuming to experiment with using human labor. At this point, the results are too crude to appear on the screen but they give human artists a place to start working.
Stevenson also demonstrated how the two forms of AI can work together, by asking ChatGPT for a prompt to use to generate Dalle-E 2 images. He asked for “a short and detailed prompt that is perfect for generating images of interior environments for Dalle 2. Ideally, have it be a futuristic, solar punk room that has a lot of nature inspiration. Looking for a short prompt.” ChatGPT spat out the following response:
Imagine a futuristic, solar punk living space that incorporates natural elements in its design. The room has an open layout, with a curved wall made of transparent glass that overlooks a lush, urban garden. The furniture is sleek and modern, with geometric shapes and metallic accents. The color palette is vibrant, with shades of green, blue, and purple that contrast with the warm, golden glow of the artificial sunlight streaming in from the skylights. The air is fresh and clean, thanks to the plants and the advanced air filtration system. This is a space where technology and nature exist in harmony.
Stevenson then put this response into Dall-E 2 as a prompt and got the images you see above. During his talk, he demonstrated the process on the fly, using a park as the image and showing how he could use both tools to iterate and improve the resulting image, getting it closer to the ideal in his head. ChatGPT might do the same for writing. In both cases, the true shortage is thinking.
I haven’t taken my own advice, but this is a free, individually authored newsletter that doesn’t even have a real title, while The Unpopulist is a regular publication that deserves appropriate production values.
History should comfort us to some degree here. If a task can be passed off to a “lesser” critter or machine such as a horse, watermill, or computer, we’ve always done so. At scale we save human bodies and brains for stuff only humans can do. In that trade, however, there’s often a period where we’re unsure to what the human capacity will shift once liberated. So far we’ve largely figured it out; we’ll figure it out again. That ability to superintend our own contribution is something uniquely human (for now), and we’re pretty good at it.
As a career newspaper copy-editor, editor, reporter, feature writer and op-ed columnist, I spent thousands of hours 'cleaning up' and punching up the better-than-average writing of pro writers and the semi-abysmal scribbling of letters-to-the-editor writers. I think ChatGPT could be a very good teaching tool -- I'd create an article or column, give it to each student and have he/she/they etc. work to improve it. Seeing how you can make a mediocre bit of writing much better -- clearer, sharper, funnier, meaner, edgier -- would teach newbie writers a lot of good writing tricks. Comparing and contrasting your 'fixes' with your peers -- and with Professor Virginia -- will add to the lesson.