Seeing Creativity Like a Language Model

I’m writing a book about the worldview I’ve developed by writing, coding, and living with AI. Last week I published the third piece from it, about how AI will impact business. Here’s the fourth, about how AI will impact creative work. It’s a synthesis of my previous writing on the topic, along with some updated thinking.—Dan Shipper

Was this newsletter forwarded to you? Sign up to get it in your inbox.

In 2023, I was at a low point in my journey with Every. I'd been running the business for several years but our growth was stagnant. My writing felt stagnant, and my co-founder had just left to run another company. I dreaded walking into a party because I knew I was going to be asked about what I do. And when I was asked about it, it would stick in my throat, like a first-time actor frozen backstage after hearing their cue.

I needed to figure out what to do with myself—with my life, with my work, with my company. I'd always struggled with the tension between being an entrepreneur and being a writer. For the previous several years, I had leaned more towards the entrepreneur side of things. It felt more legible, lucrative, and prestigious. My last company had been a startup, too, so I was used to being able to walk into a room and talk about myself as a founder.

I realized that if I was really honest with myself, I actually wanted to be a writer. But I also wanted to have a great business—or maybe a few great businesses. And the two seemed mutually exclusive. I had no examples to turn to. Most of Silicon Valley lore says that a CEO doesn’t have time for personal writing or any other form of creative work. My own experience confirmed as much. Being a CEO was both time-consuming and emotionally taxing. It wasn’t conducive to the space required to be really creative.

This was in the very early days of ChatGPT, when the model was still GPT-3. Hoping to resolve my existential dilemma, I decided to ask for a list of writers who also had great businesses—in other words, a list of people to look up to.

It immediately spat out a few great answers. One was Sam Harris, a writer, philosopher, and neuroscientist who has published many books, has a really popular podcast, and also has an extremely popular meditation app called Waking Up. I'd known about Harris for a long time, but I'd never really connected the dots—that he'd built the kind of career and business that I aspired to.

Another example that came up was Bill Simmons, the famous sportswriter and podcaster who started Grantland and later founded The Ringer, which he sold to Spotify for around $200 million. Again, I'd heard of Bill Simmons, but I'd never thought about him in this way.

ChatGPT came back with a bunch of other names: Ezra Klein, and Nate Silver, and John Green. All writers who I knew about and loved, but who, suddenly, I was seeing in a different light.

In the Google era, I would only have been able to Google for “writers who are also great business people” and hope that someone else had thought to create an article or blog post naming those people. In my case, no one had really done that. My path in life was not yet well-defined.

Build a project in a single day

AI can feel abstract until you build something real—and tools “made for programmers” can stop you before you start. Spend one day with Every’s Dan Shipper to get acquainted with the best tools on the market, assign your AI agents real tasks, and ship something live. Claude Code for Beginners runs Nov. 19, live on Zoom. You’ll leave with a project, a reusable workflow, and a head start on building the agentic apps that will populate the future.

Save your seat now

Want to sponsor Every? Click here.

How technology reshapes creative work

Most discussions about how AI will affect creative work revolve around fears that it might completely replace it. We imagine a doomsday world where AI agents autonomously creating tasty slop push real artists to the side. I don't think that's going to happen. But AI will change art, creativity, and creative work.

Technology doesn't tend to eliminate artists or art forms, but it does tend to change art forms and the skills required to be an artist. TV, radio, and the Internet didn't eliminate books. But it did turn books from a mass medium, the primary way in which educated people learned about the world, to something more niche and aspirational. The same thing happened with musicals. Once the dominant form of popular entertainment in America, musicals still exist, and the aspirational nature of being on Broadway is still a part of popular consciousness. But musicals are more expensive, more special, compared to, say, TikTok videos where young people sing and dance for the camera.

Concerns around AI and art largely boil down to the fear that AI can reproduce exactly any work that an artist does. Taking any paragraph of text from a writer, an AI could probably write that. It’s the same thing with a given painting or film clip. But if you look closely, this is the same old rationalist concern. It identifies art and its value as just the text on the page or the pixels in the image—completely divorced from context and history. But as we've seen in science and business, context is king. The context within which an AI-generated work is created is necessarily different from that of a human creative, and while in certain circumstances those works will be interchangeable, in many other ones they never will be.

Moreover, I think that AI will be a boon to creativity and creative work for those who embrace it. Most groundbreaking art is taboo at first. And the cultural taboo around using AI for art signals that it is fertile ground for the brave creatives who embrace it to produce novel works.

Let's talk about what that looks like.

The allocation economy

The way I think about how AI will affect the economy at large is to move us from a knowledge economy to an allocation economy. In a knowledge economy, you are compensated based on what you know. In an allocation economy, you are based on how well you can portion out the resources of intelligence.

You need vision, a taste for ideas and for language, and an ability to effectively communicate what you want in words. You need to know how to plan and how to estimate timelines, how to break up projects and distribute them among different people who are skilled at different tasks. You need to know when to step in and what to micromanage to ensure the details are right.

All of these skills already exist in the knowledge economy: They are the skills of human managers. Not just managers in business, but anyone who manages humans: Showrunners, directors, and conductors, all do the same kinds of things.

But human managers make up only about 10 percent of the workforce. And in order for everyone to effectively work with AIs on a day-to-day basis, these are the exact skills that we all need to develop.

In broad strokes, this is how AI might affect creative work. For many creative tasks, we’ll move up one level of abstraction from doing everything ourselves to directing what work needs to be done and how.

From sculpting to gardening

Another way to think about this is to consider the difference between a sculptor and a gardener. A sculptor takes a block of marble and turns it into a work of art. Every line and curve in a sculpture is there because the sculptor decided to put it there with her hands.

A garden is different. A gardener can’t directly cause a rose to bloom. Instead, the gardener’s job is to create the conditions under which a garden can flourish. It is to make sure there is an appropriate amount of sunlight and water and nutrients, to prune and shape plants, and to keep pests at bay. The gardener guides and nurtures but doesn’t directly create.

Working with AI is more like gardening than sculpting. We’ll need to create the right conditions, provide the right prompts and direction, and shape the output—but the actual generation of content will be handled by the AI. This shift requires different skills but opens up new creative possibilities.

Of course, one difference between gardening and working with AI is you’ll always be able to intervene as a sculptor in situations where it’s necessary or desirable to do so. And just like many people practice and cherish the craft of, say, hand-making tables or clothes, even though mass manufacturing can make them with machines, many people still choose creative work entirely by hand.

But what does it look like to do creative work in an allocation economy as a gardener instead of a sculptor? Let's discuss.

Pattern collectors

For almost five years now, I’ve kept a note on my phone called my ineffable list. It’s a trick I learned from the writer Robin Sloan. Every time I come across something in the world that makes me pause—that has a little bit of that flavor of whatever it is that I like—I put it in my ineffable list.

For example, the top note of my ineffable list is a quote from my nephew. He asked his mom (my sister) recently, “Where does the wind go?” Something about that just caught my attention. I don’t know what exactly that something is or what, if anything, I’m going to use the quote for, but I love it.

Right underneath that is a quote from an Anton Chekhov story called “About Love,” where he describes a man and a woman who are in love but who can never be together because she’s already married. They’ve gone to the opera together, they’re sitting next to each other; the quote reads: “I would silently take the opera glasses from her hands, and at that point, I would feel that she was close to me, that she was mine, that we couldn’t live without one another.” I love this because it conveys some universal experience of longing and forbidden love, and does it with so much vividness and economy.

If I scroll down a little bit more, I have a note about my dad. The family business is cemeteries and funeral homes, and my dad has a process that he calls “The System” for how to increase cemetery sales. There’s something there that resonantes with my writing such a long work about the limits of rationality in systems. I just have a little note about it to remind me. Maybe I’ll use it for this very piece.

Collecting little moments like this is at the heart of creative work—even before AI. Writers have writers’ notebooks, artists have sketchbooks, and musicians often record snippets of melodies. These collections become the seeds of future work. If we approach creativity like a language model, we see that these collections are also the context for our future art. They’re the training data. Despite concerns about AI “flattening” art and culture, I think it has the potential to make our creative work even more unique, personal, and valuable, for three reasons:

1. AI can help you find more of what you like—even if no one has collected something like it before.
2. AI can help you articulate what you like.
3. AI can help you to create more, quality work with your taste.

Finding more of what you like with AI

When, at a low point in Every’s journey, I asked GPT-3 for a list of writers who are also business people it immediately connected me to a set of people that I already knew about, but collected and presented them to me in a newly inspiring way.

Once I had these role models, I completely reoriented the business. Instead of spending most of my time in meetings, I made it a priority to write for the first half of every day. I previously would never have done this. It would have felt too embarrassing to tell my team that I would be unreachable. But having examples made me confident that my writing could be the engine driving the rest of the business. I also saw how many of these writers hired talented executives who understood their vision and who could help operate the day-to-day of the businesses, giving them space to do their creative work. So I did the same. And the business subsequently took off in a way that it never had before.

The examples AI help you find go beyond individual role models too. After I sold my last startup, I spent a year or two trying to write a novel. I'd come around to a particular scene—maybe a character would be walking down a street at night, or a character would be having a fight with their girlfriend—and I’d want to paint the scene or write the dialogue. And I’d want examples. But you can’t really Google for examples of something like this. You needed to be an expert. You needed to have read tons and tons of novels with a particular eye for dialogue and scene-setting. Or, you needed to work with an expert who had—which is difficult, expensive, and out of reach for most people.

AI does this. It can help you gather the examples that inspire you to create work that you've always dreamed of. And to do it far more quickly and easily than you ever could have before.

Once you have these examples, AI is also quite good at helping you articulate what about them moves you.

Articulating what you like with AI

Being able to describe what you like is one of the most important skills for creatives and, interestingly, for managers. Of course you can never fully articulate what you like—fully eff the ineffable—in part because what you like changes depending on context. But having some way to talk about it helps guide your own creative work and helps you collaborate more effectively with others.

For example, I know that I love writing that explores deep philosophical themes about human nature, makes complex concepts accessible, beautiful, lyrical, and emotionally resonant, and that deals with nature, science, technology, psychology, and philosophy. I've always loved this, but I've only really been able to say it recently because of AI—around the same time that ChatGPT helped me identify heroes, it also helped me articulate the kind of writing that I love.

I threw a list of my favorite writers into ChatGPT and asked it to describe what was common about them. Here's what I got:

Based on your admiration for these authors, it seems you enjoy writing that:

1. Explores deep philosophical themes about human nature, existence, and consciousness.
2. Makes complex concepts accessible, whether in psychology, philosophy, or science.
3. Integrates narratives or metaphors that help to illustrate or illuminate these concepts.
4. Expresses a deep sense of empathy, introspection, and curiosity.
5. Utilizes beautiful, lyrical language to express these ideas.

It appears you appreciate writing that is both intellectually stimulating and emotionally resonant, combining clear, accessible explanation with deep, thoughtful introspection.

I also took my full ineffable list and did the same thing, but this time I put it into Claude:

Based on the information provided, here is a summary of your preferences and values:

Themes and Subject Matters:

1. You seem to enjoy thought-provoking and philosophical ideas, especially related to psychology, spirituality, and existence. There are references to Jung, Frankl, and Anthony de Mello indicating an interest in depth psychology, meaning-making, and mindfulness.
2. You have an appreciation for nature, science, and discovery. References to Tinker Creek, the 'invisible form,' and interconnection suggest a sense of wonder about the world. Mentions of terraforming planets and space point to an interest in science fiction.
3. Relationships and human connections are meaningful to you. Quotes about love, intimacy, friendship, and longing for places and people point to the importance of relationships.

Style and Tone:

1. The overall tone seems contemplative, introspective, and poetic. The language used is often metaphorical, lyrical, or turns of phrase. References to 'slits in the skin of the present' and 'the careful emptiness inside a cello' suggest a poetic sensibility.
2. There is an occasional wry, sardonic wit and humor. Mentions of 'revolutionary and courteous' manners, and illogical political logic point to a capacity for dry, ironic humor.
3. A sense of nostalgia and wistfulness pervades some entries, indicating a tendency to feel sentimental about the past. Mentions of missing Kyoto, childhood treats, and formative memories of family point to nostalgic sentiments.

In summary, you appear to value philosophy, relationships, aesthetics, nostalgia, music, food, wit, quirkiness, and childlike wonder about the world. Your tastes seem to stem from a poetic, introspective, and meaning-oriented sensibility that finds depth and beauty in life. A richness of references across subjects points to intellectual curiosity and openness to a diversity of ideas, experiences, and forms of meaning-making.

Reading this output almost made me cry. I felt like someone had handed me a CT scan of my soul, allowing me to see myself in a way that I'd never seen before. It gave me a creative direction. I could start to evaluate whether my work truly embodied the characteristics that I was trying to embody.

Once you know what you like, and you can articulate it, AI can also help you make more of it.

AI as a creative tool

People often fundamentally misunderstand AI as a creative tool: They assume it’s meant to do everything, replacing human creativity entirely. They’re afraid that the resulting work is soulless, inhuman.

But using AI well is like using any other tool well. You employ it for tasks that are best suited to its abilities. You act as an editor for everything it produces. And it becomes a creative partner, helping you to not only work faster but also to discover and bring forth a kind of work that wouldn’t have been possible without it. It also reveals things about creative work that were previously invisible.

First, it shows just how much hidden drudgery exists in the creative process—work that could be automated without compromising artistic vision. For example, as a writer I normally spend hours and hours a week composing tweets about articles I’ve written or summarizing complex ideas that I already know into one or two sentences for a piece I’m writing.

Second, it highlights how our creative potential has been limited by lack of access to information, teachers, and specialist skills. AI can democratize access to expertise that was previously available only to a select few.

Third, and perhaps most surprisingly, AI illuminates the subtle differences in how our brains work—differences that may have invisibly limited the kinds of creative work we could do. These aren’t necessarily what we’d traditionally call neurodivergence, but they’re real constraints that AI can help us work around. It can:

1. Understand complex material: Dense texts in philosophy, machine learning, or literature are often too intimidating to tackle outside of a graduate program. I keep my phone next to me with ChatGPT open. I’ll snap photos of the text or turn on advanced voice mode—it translates abstract language into plain English, helps me understand the historical and cultural context around books like Leo Tolstoy's War and Peace, and even have conversations with the characters.
2. Get ideas out faster: When I sit down to write, there’s a dense network of ideas, concepts, feelings, and other totally unnameable things happening in my brain. They exist as a sort of direction or a potential energy, and writing is the process of turning it into linear sentences. That’s why writing is so hard. But it’s often easier to talk about what I’m writing than to actually write it. I can hold a button, watch the sound waves of my voice appear on screen, and see words materialize in my document.
3. Summarize quickly: One of the biggest sources of drudgery for a writer is writing summaries. Even when I know much of the material, getting out a three-sentence summary of something like the history of machine learning normally takes half a day of googling and rereading old books. AI generates those summaries in seconds. I can start 80 percent done and spend the rest of my time making it great rather than reconstructing everything from scratch.
4. Edit with multiple perspectives: As a professional writer, I’m lucky that I have talented editors to work with. But my first editor is always AI. It’s not a replacement for humans, but it’s always on, quick to respond, never in a bad mood, and deeply knowledgeable about every subject. It can simulate a New Yorker editor, a poet, or even thousands of different readers at once using tools like Rally, giving me aggregate feedback that would normally be out of reach unless I had millions of dollars or an established audience.
5. Adapt your work for different audiences: Another source of creative drudgery is making sure people see it. If you’re a writer, that means putting your work on X or LinkedIn, or querying agents with a summary of your book. If you’re a YouTuber, it means creating clips, thumbnails, titles, and headlines. If I’ve created work I want people to engage with, I need to create smaller versions to entice them to go deeper. It’s repetitive, it’s boring, and it can feel a little hacky, but it’s massively important. Creatives can't afford—often literally—to leave their marketing up to someone else. AI handles this transformation work beautifully.
6. Build things without huge upfront investment: A lot of creative work is off-limits because it requires you to get people to believe in you before you can make the thing. If you want to spend three months writing a book, you have to get a grant. If you want to make a sculpture, you have to get money for materials and a space. If you want to make software, you have to get investors interested. AI changes this calculus. I was able to, in a couple of days, spin up a prototype of Spiral, which at the time automated the above-mentioned creative drudge work—and as of this week is a fully fledged agentic ghostwriter.

Generalists own the future

In ancient Athens, when Socrates was around, citizens were expected to be generalists. A man was a farmer, a lawyer, a defendant, a juror, a congressman, and a warrior all in one. It was a golden age of culture, and it was enabled by the generalist nature of that society. The British classicist H.D.F. Kitto wrote that the Athenian ethos implied, “a respect for the wholeness or the oneness of life, and a consequent dislike of specialization.”

As Athens became an empire, specialization started to win the day. A man couldn't be expected to know how to run a tiny colony in Sicily while also managing trade routes to Egypt. The complexity of the empire required specialists. This pattern has repeated throughout history—as societies grow more complex, they tend toward specialization. Adam Smith wrote about it in The Wealth of Nations when he described how in a pin factory, one worker making pins alone could perhaps make 20 pins per day. But when the work was divided among specialists—one drawing out the wire, another straightening it, a third cutting it, a fourth pointing it, and so on—the same workers could make 48,000 pins in a day.

This dramatic increase in productivity through specialization became a foundation of modern capitalism and industrial society. It also brought its own problems. Karl Marx famously pointed out the alienation that workers felt when they were reduced to performing repetitive, specialized tasks. They quit Henry Ford’s assembly lines in droves, forcing him to shorten the workday and double their pay. The assembly line worker, disconnected from the final product, lost the satisfaction of creating something whole.

AI may provide a solution to this problem—and a return to Athenian generalism and the ideal of the Renaissance man. With AI as our collaborator, we can bridge the gap between specialized knowledge and generalist capabilities. Having what is effectively a thousand PhD-level experts in your pocket allows you to work across diverse domains without sacrificing depth or quality. Like the Athenian citizen who seamlessly moved between roles, AI can enable us to expand our creative and professional repertoire, breaking free from the constraints of narrow specialization.

In doing so, it can help a much broader part of the world to live a fuller and more creative life.

Next week, we’ll wrap this series with a final essay on how tools shape the world that we see.

Read the first three pieces in this series, about the new worldview enabled by AI, and how AI will impact science and business.

Dan Shipper is the cofounder and CEO of Every, where he writes the Chain of Thought column and hosts the podcast AI & I. You can follow him on X at @danshipper and on LinkedIn, and Every on X at @every and on LinkedIn.

We build AI tools for readers like you. Write brilliantly with Spiral. Organize files automatically with Sparkle. Deliver yourself from email with Cora. Dictate effortlessly with Monologue.

We also do AI training, adoption, and innovation for companies. Work with us to bring AI into your organization.

Get paid for sharing Every with your friends. Join our referral program.