Essay Writing as Personal Sovereignty

Announcing a runner-up in the Cosmos Institute essay contest

Over 115 readers entered our inaugural Cosmos Institute Essay Contest, where we asked for reflections on the theme of human autonomy in the age of AI. We received essays from across the globe, with submissions from the UK, Italy, France, Germany, and India. The response was remarkable, and we were impressed by the range of ideas.

It took several weeks of careful deliberation—and more than a few spirited debates among our judges—to narrow down the field.

Today, we are proud to announce one of our two runners-up: Michael Dean, who is charting new territory with his essay, “Essay Writing as Personal Sovereignty: Shaping AI to preserve the slow craft of critical thought.”

An architect-turned-writer and O'Shaughnessy Fellow, Michael challenges us to imagine a future where creativity and technology enhance, rather than diminish, each other. He does this quite fittingly through a heartfelt description of the role of the essay.

---

Essay Writing as Personal Sovereignty

Shaping AI to preserve the slow craft of critical thought

This 2,800-word essay took me 45 hours to write. I wrote it from scratch five times, and only 10% of my words made the final cut. That comes out to an average pace of one word per minute; imagine typing a single word, taking a brief walk, and then coming back to type the next. Writing happens at an unbearably slow pace for a culture that’s glued to vertical feeds with split-second reward loops, but thinking takes time. Good thinking takes a lot of time and even more toil. Essay writing is a process I’ve grown to love, a process I believe is deeply human, and yet it’s a process that’s becoming endangered.

Someone in the near future, equipped with a superintelligent AI, might consider the manual labor of essay writing to be a tremendous waste of time. They might be able to generate an essay similar to this one in under 45 seconds. But we cannot blame AI for the death of essays, for the essay has been in a slow decline for decades—long before household computers. Essays are slow to write, hard to make good, and terribly unpopular as a form of recreation. They’ve become the symbol of rote assignments and rigid rules. In popular literature, they’re basically irrelevant; the essay section at the biggest Barnes & Noble in the world fits within your arm span. It’s not surprising that essays are being automated by people forced to write them.

Despite the genre’s poisoned brand and questionable future, the mainstreaming of the essay is important to fight for, considering it was a latent but unrealized potential of the Enlightenment.¹ It’s a vehicle for personal sovereignty, and I believe the vibrancy of a society’s essay culture is a proxy for how much its citizens value critical thinking. Its disappearance would signal trouble, especially in an age where sensemaking is only getting harder.

If the essay is the best tool we have to foster independent thought at scale, then it might be the most important written medium we have. We must consider why it failed, why its revival matters, and why—thanks to technology—its flourishing is more possible than ever.

Which principles will help us design and govern AI so that the tradition of the essay can experience a golden age in the 21st century?

---

I: Writing as a personal quest for truth

---

If I had ChatGPT in high school, I might have cheated. I didn’t understand why I was assigned a 5-paragraph essay on Of Mice and Men with required citations. In this paradigm, writing is transactional; we hand it in for a grade, to pass a class, to get a degree, and finally, a job where we hopefully never have to write again. AI is exposing this broken incentive system, and it’s resulted in the “Homework Apocalypse.”

But looking back to the essay’s origin, you see that we somehow completely inverted its nature; it’s not about advancing through educational gates, but excavating our own minds. The word “essay” itself suggests independence; in French it translates to “to attempt” or “to try,” and it’s no coincidence that it emerged at the dawn of The Scientific Revolution. An essay is a personal experiment to discover truth through writing. 

In 1580, Montaigne’s first volume called Essais set a course for future intellectuals, not only in pioneering the genre of the essay, but in spreading the practice of public-facing free-thinking. The printing press was invented in 1439, and its first century was dominated by religious works, classical texts, and practical documents. Original, secular works started to emerge and circulate in the mid 16th-century; Montaigne was perhaps the first to use this new distribution medium to spread an incipient mode of skepticism (he challenged convention and derived his own conclusions through essay writing). He started not with dogma, but with questions. In 1863, essayist Alexander Smith called the essay writer a “chartered libertine, and a law unto himself.”

While writing is the tool to externalize what you think, editing is the tool to transform what you think. There’s even a distinct word related to editing: “assay,” which means “to analyze something and determine its contents.” Half a century before Montaigne, the alchemist Paracelsus used this word—is alchemy not the perfect metaphor for this process? Writing freezes your amorphous, gaseous thoughts onto the page; editing is the act of dissolving prose to see and interrogate your ideas. It is a slow, often daunting process, but it is the cost of intellectual sovereignty. As you edit, you don’t just edit the words on the page; you edit the nature of your identity, your beliefs, and your morals. As author S. Kelly Harrell said, “a good editor doesn’t rewire words, she rewires synapses.”

What’s remarkable about this transformative process is that it’s accessible to anyone, simply by arranging words in their native tongue. Of course, composition is hard, but unlike other forms of expression, essays don’t require foreign syntax (like in coding), dexterity (musicianship), resources (architecture), or time (painting). In a single day, at low cost, anyone can compose, refine, and publish their thoughts to the world. The essay is the most democratic tool of expression we have; if we can scale the practice of composition, we can scale critical thinking. 

But even though the essay helped shape the Enlightenment and its vision for education, something important was lost in translation.

---

II: The challenge of scaling critical thinking

---

Montaigne became known as the founder of modern skepticism; he influenced Enlightenment thinkers like Bacon, Pascal, Rousseau, Voltaire, Locke, and Hume, several of whom cited him as inspiration. The essay was not only the medium these figures used, it also mirrored their vision for education: intellectual liberty, freedom from authority, and the development of reason and self-reflection. While some early institutions embraced these Enlightenment ideals, the rise of mass education in later centuries ended up departing from them.

The 19th-century project of American mass schooling was a radical success, but it came with a compromise. Literacy rates rose from 50% to 95% over a century, but we had to prioritize standardization over autonomy. While the Enlightenment may have inspired mass education, it didn’t inform its implementation. Instead of being decentralized, non-dogmatic, reflection-oriented, and fitted to each individual, Western education is state-controlled, Protestant, memorization-oriented, and uniform. The goal of our system was similar to that of its Prussian inspiration: to build a workforce to meet the needs of industrialization. Technology skewed the humanities. As a result, our writing programs were highly utilitarian; instead of teaching the essay in the tradition of Montaigne, programs focused on grammar, bureaucratic documents, and structured legal defenses.

However, even if we tried to build an education system of the purest Enlightenment ideals, we would’ve faced insurmountable issues of scale. In 1861, Emerson published an essay on education to criticize the emerging system. He called for one-on-one attention from mentors that could nurture the innate genius of each child down tailored, lifelong learning paths that usher in intellectual, moral, and spiritual development. While correct in spirit, it didn’t address the challenge of finding and training such a volume of qualified teachers. In 1857, the NYC superintendent of education—Henry Kiddle—said that teachers were “placed in the school-room with very little but the knowledge of what is to be done, the how to do it is, as it were, a distant haven on the other side of a pathless ocean which is to reach without pilot, compass or chart….”

Schools failed to inspire essay writing, and for a range of reasons it never took off as a popular genre of literature to read, either. A single essay is too slim to be bound and published, so they’re often gathered in disjointed collections. Unlike the novel, which became a primary form of entertainment in the 19th century, the essay was more abstract and less accessible. There certainly is a canon of timeless, well-rounded essays, but they’re only appreciated in narrow literary or intellectual circles. In 1977, E.B. White decried the essayist as a “second-class citizen,” destined to ramble through an undisciplined, unappreciated existence. The essay could have been the people’s medium, but the citizenry never found the intrinsic drive to read or write them.

As we enter the age of artificial intelligence, it’s worth being honest about the shortcomings of the Enlightenment, education, and the essay; our existential problem isn’t about chatbots replacing our greatest authors, but that the median citizen will have no meaningful reason to write. All the occasions where we have to write are instrumental. Whether we’re writing an email, posting on social media, or drafting a contract, the words matter less than the results; writing is a social tool, a means to an end. As long as we weave together an acceptable, legible, and coherent sequence of characters and spaces, we might get what we want. Once we automate the menial act of “correspondence,” what reason is there to suffer over sentences? We can all just exchange machine-generated strings of text and leave capital-W Writing to the real Writers: the poets, memoirists, and novelists.

It’s plausible that most people won’t need to write in the near future; we might even celebrate the triumph. But writing and arranging words is the process of careful thinking. By offloading it, will we not atrophy the muscles of thought? We’ll each gain the illusion of competence, without even being aware of our cognitive amputation. We might lose our personal sovereignty in an age where we need it the most.

The human operating system is experiencing a scale, frequency, potency, and specificity of media that is psychologically overbearing. We are in, what David Foster Wallace called in 2007, an environment of “Total Noise.” When sensemaking gets stressful, it becomes tempting to outsource our thinking to “trusted” experts. When we do this, our worldview risks getting hijacked and programmed by algorithms, influencers, and power structures. If we submit to convenience, we fall into groupthink, but if we take the time to think and write through our circumstance, we earn back our agency.

---

III: Technical principles for the golden age of the essay

---

Our challenge is to preserve a classical mode of thinking in a very modern context. No matter how well we convey the importance of Montaigne, the merits of skepticism, or the glory of essay writing, we are working against a deeply human attribute: impatience. 

It is becoming increasingly hard to inspire slow and hard activities with delayed rewards. Will anyone actually prioritize their future self when 100x shortcuts are rampant? Every environment that could foster essay writing is incentivized for speed and ease. Teachers are locked in bureaucracies, and students have low autonomy. Software is in a cult of efficiency, where users pay to eliminate friction. Social media networks devolve into contests for attention, causing writers to focus on growth hacking instead of thought crafting. These contexts all frame writing as a bothersome chore to be cheated.

In order to inspire the slow craft of essay writing, there needs to be (1) a compelling vision to make and share essays in the age of YouTube and TikTok, and (2) technology that makes the process magical without reducing it to a prompt and click.

At the end of Paul Graham’s 2004, “The Age of the Essay,” he notes that, “the Web may well make this the golden age of the essay.” Two decades later, I’d say this is very much in progress; while Internet long-form is relatively eclipsed by tweets, video, and content marketing, there are independent essayists who are bypassing gatekeepers and unlocking a hidden continent of opportunity. 

Online writing inverts the paradigm of English class; instead of writing formulaic essays on forced topics for a teacher who might not care, you write creative essays on emerging interests for an audience who might pay you. Why don’t more people do this? It’s that perennial problem: it’s prohibitively difficult. Even the writers who already see the potential of the medium get blocked by the challenges of drafting, structuring, editing, and articulating. Can technology fix this?

At the core of our solution is a paradox: how can AI help writers without doing any of the writing? Is it possible for generative text to help us construct essays without automating the process that refines our thinking? Instead of offering specific solutions—on how to reform schools, software, or social media—I’ll propose five principles on how we can implement AI to preserve the tradition of essay writing:

1. Ensure writers craft their own sentences. This is the cardinal rule. If crafting prose is the main function that refines thought, then environments that care about writers should make it logistically impossible for AI to craft sentences. In an open market, there will always be an array of chatbot options that will offer a single-prompt essay, but this doesn’t prevent alternatives that (a) keep the writer in charge by preventing mindless copying and pasting, and (b) engage their thinking in such a profound way that it makes the friction worthwhile.

2. Build a protocol for  “margin muses.” Most AI writing happens within one of two interfaces: either (a) through a chatbot or (b) through a document editor, where the writer has precision over where text generation occurs—but we need a third option. Imagine a relatively analog editor, but one where each sentence is processed in real-time by a “margin muse,”² offering different types of insights on the side. In this interface, only the human can edit the document, but AI-powered agents provide the Socratic service of asking great, contextual questions. This interface could serve as a framework for developers to build margin-based applications: research assistants connected to personal knowledge bases, philosophers that find holes in your argument, or tutors that shape micro-prompts to drive the writer forward.

3. Unite around a shared composition philosophy. We can build an AI-powered editor that’s trained on history’s best writing, but it requires us to answer a hard question: what is good writing? While certain qualities may always be subjective—like moral worth, cultural relatability, or historical relevance—there may be an objective composition language³ that underlies all great essays. We need to analyze the classics,⁴ find and quantify the timeless patterns, and turn this philosophy of craft into code. The existence of a history-backed quality standard would let writers upload drafts and get powerful feedback on their idea, form, and voice.  If we map the secret architecture of great essays and build it into AI, it will make a hard craft substantially more approachable.

4. Incentivize quality publishing. Social media incentivizes writers to publish essays that are likely to go viral, but we can build much better methods of gamification. Once we’re able to use AI to identify great writing, we could build curation algorithms that reward quality. Considering the Internet is about to be flooded by bots to hack attention algorithms, it’s important that we build better ones that reward humans for publishing good work consistently.

5. Reframe the essay as an intellectual matchmaker. Feeds are currently shaped based on consumption patterns, but what if they were based on the ideas you wrote? Every time you publish an essay, it could semantically analyze your paragraphs—using embeddings—and construct a feed of other essays or excerpts. Not only would this connect writers who cover similar topics, it could find deep conceptual alignments between thinkers who are using different vocabularies and frameworks. In this model, the essay is a key that unlocks a social reality.

AI’s goal should never be to make writing “easy”; it should aim to make it approachable, intuitive, and rewarding. The Internet has unlocked a new reason to write essays, and AI—if we build it right—can enable anyone with patience to embark on the kind of lifelong learning process that Emerson dreamed of.

While the ratio of novel readers to novel writers might be something like 1,000:1, the ratio of essay readers to essay writers could theoretically be closer to 1:1.

If we can step outside of the productivity vortex and build writing tools that fuse AI with humanistic values, we might realize the potential of the essay and elevate our culture to a new plane of thought—finally fulfilling the original vision of Enlightenment education. 

1. Enlightenment thinkers wrote prolifically through essays and used the medium to dream of education reform, yet they never realized that the essay itself—if taught properly—could have been the method to realize their vision at scale.

2. Here’s a rough prototype I built called Margin Muse. You’ll see that as I complete sentences (triggered by a punctuation mark) it uses AI embeddings to bring related notes into the margin.

3. I recently published an essay called A Pattern Language. Inspired by architect Christopher Alexander, I’ve organized timeless writing concepts into a hierarchical and interlocking system.

4. I’ve been reading, scoring, and reviewing essays in public on Dean’s List.