I’m giving $10,000 to the best essay of 2025

Unlike literary prizes w/ vague standards, I've turned an unreasonably detailed rubric into software + am giving you unlimited access. Runs 9/15-11/9. Goal: publish an anthology w/ 10+ essays >4.5/5.

The Internet needs a quality algorithm. Without one, it will only get harder to discover and celebrate great writing. It used to be the role of literary institutions to sift through silos of submissions, to find the signal in the noise, to elevate great work from unsuspecting people to inspire the larger culture.

Consider the 21-year-old college student who won the Prix de Paris essay contest in 1956: she won $1,000, a job at Vogue, and went on to become Joan Didion. Or consider the unknown, 38-year-old sheet music copyist who beat establishment intellectuals at the Academie de Dijon essay contest in 1750 and went on to become Jean-Jacques Rousseau. Now consider how similar things happened to James Baldwin, David Foster Wallace, Brandon Sanderson, and whole generations of writers, known and unknown.

But when a culture loses its ability to properly fish for talent, you get something like today’s Internet: infinite hooks, but everything is seaweed. Writing hits your glowing screens at numbing speeds and scales based on its popularity and extremity. Engagement algorithms are hackable. Consider the state of the Substack leaderboards: #2 was recently hung for plagiarism, and multiple accounts flutter into the top 10 with AI slop and cogsucking armies of commentbots. Trite little Notes that read like greeting cards are getting more attention than entire editions of The Iowa Review. Meanwhile, the platform has an underbelly of growling, undiscovered talent. This is merely a minor annoyance, for feedmakers care less about merit than total user milliseconds.¹

The journals, anthologies, magazines, and MFAs—the purported guardians of writing education, economics, and quality—haven’t adapted well to the 21st century (this could be its own essay, but for now I’ll make a generalization and say that old giants were gutted and turned into media companies²). Whether it’s The New Yorker or Best American Essays, they seem to be more driven by prestige than virtue. There’s little incentive to modernize, either; essay competitions are run the same way in 2025 as they were run in 1750. Some still use mail. They charge you a $24.50 entry fee, disappear with your work for 3 months, only to send back a template rejection letter without any feedback.

As I build the software for Essay Architecture—which generates feedback by running thousands of 1-5 evaluations—I realize I’m building more than an educational tool. There’s a social component to this: it’s an engine for essay discovery, the beginnings of a quality algorithm.

So I’m putting up $10,000 to test a thesis: can I create an internet-native technology-forward essay prize that is more accessible, more transparent, and more synergetic, resulting in an anthology that is commercially viable and culturally catalytic?

In two weeks I’ll share all the details, but I’ve used this essay to explore and articulate (first to myself, and now to you), how the Essay Architecture Prize is different from traditional essay contests. I analyzed 65 of them. I entered 2 of them myself³ to feel the dynamics (in 2024 I lost the Astral Codex Ten book review contest, but won the inaugural Cosmos Essay Prize). There are dozens of design decisions⁴ to consider, and with the right mission, the right technology, and a big-enough prize, I think it’s possible to create a new discovery engine that rewards the right things, otherwise timeless essays will continue to hide in the slopstorm, dwelling unread in the gutters of Substack.

---

---

I, MORE ACCESSIBLE: At $10,000, this is the world’s largest “open” essay prize. It doesn’t use the essay as a tool to promote an agenda; its agenda is to promote the essay.

---

I noticed that all the competitions with significant prizes have restrictions around location, age, or topic. The CBC Nonfiction Prize ($6,000) is only for Canadians (it is funded by their government). The John Locke Institute Essay Competition ($10,000) is only for college applicants. The Willie Morris Awards ($12,000) is for “under-recognized writers” writing about the US South. The Ayn Rand Institute Essay Contest ($25,0000) is only for high school students willing to write a book report on one of her three books. The Berggruen Prize ($50,000) is only for someone capable of exploring the frontiers of unsolved problems in philosophy and consciousness. The Nine Dots Prize ($100,000) asks you to submit a 3,000-word essay, but the winner has to agree to spend a year expanding it into a book for Cambridge University Press.

I think you get it. The point is, these contests use the essay as a tool to advance some institutional agenda. There is nothing inherently wrong with that; if you’re going to give away free money, it might as well be strategic. However, this means that the big prizes are usually off-limits to the average writer.

Now, there’s a whole other breed of competitions which I call “open” essay prizes. They celebrate a specific medium of literature (essays, fiction, poetry) and try to find emerging voices. They give little restrictions—sometimes just a word count—and the prompt⁵ is open-ended enough for anyone to enter. The problem is, the largest open essay prize is not very big. The William Hazlitt Prize (~$23,000), hosted by Notting Hill Editions, quietly disappeared five years ago, and so now the largest one is $5,000.

At $10,000, I’ll be hosting, “the world’s largest open essay prize.” It’s neat to say that. As a sum, it’s a lot for me to casually give away, and for the recipient, it’s big enough to work hard towards. But still, it’s not necessarily life-changing money, and it’s definitely not society-changing money.

Think about how the X-Prize (and others) give away $1,000,000s per year for scientific innovations; why is literature/education/philosophy valued 100x less? As writing gets threatened by AI in the decades ahead, it will become urgent to incentivize and champion the essay, our best folk medium. Unlike fiction, memoir, or poetry, it’s the one literary form that, theoretically, could be taken up by everyone. Everyone (everyone) has space in their life to write one essay per year. If AGI/ASI automates all our cognitive labor, then essay writing could be the recreational hobby that sharpens our thinking, the gym for the mind.

So let’s imagine a utopian future where an open essay prize offers society-changing money: $10 million. People are freaking out, borderline obsessed, kind of like in Ready Player One. Joyce Carol Oates submits, but loses to an unexpected teenager who lives in a trailer park. If the 28% prize:entry ratio⁶ holds, that would be an awful lot of reading and judging to do. 2,800,000 entries at 2,500 words comes out to 7 billion words, and if the judges read slowly and carefully at 30 WPM, it would require 3.8 million reading hours, or 486,000 days of full-time reading, which could be pulled off in a year with a team of 1,300 readers, reading full-time, 8 hours a day, without weekends or vacation. Even if you paid everyone minimum wage, it would cost something like $62 million to pay your judges.

The point is, a society-wide open essay prize can’t scale with human judges. Paradoxically, if you want to build an institution to preserve human writing, it might only be possible with AI evaluation.

---

II, MORE TRANSPARENT: The judging standards are transparent, detailed, and built into software that you get for ~$3-9/draft. This lets you get instant, unlimited feedback, making it possible for anyone to iterate towards a winning essay.

---

I looked through the top 10 open essay prizes from 2024 to see how they communicated their quality standard, and couldn’t believe the lack of specificity. They ask for “writing with a command of craft.” They say “writers are encouraged to explore the form,” and “the Selection Committee judges applications on the quality of the writing samples.” Sometimes it gets poetic, asking for “daring formal and aesthetic innovations that upset and reorder literary conventions, historical narratives and readers’ imaginations.” If they’re not vague, they’re often completely missing. The Iowa Reviews Awards ($1,500), associated with the renowned Iowa MFA program, has nothing to say about what they want. They’ll know it when they see it.

Of the contests that I analyzed, the average submission fee was $24.21, and most don’t offer feedback (though I found two that did, for $90-120 per entry). So basically, you pay money to submit into a black box.

With Essay Architecture, all essays are scored on 27 patterns, each with a public definition,⁷ and there is a 20,000-word wiki that explains what each pattern is, why it matters, how to implement it, and it gives examples and rubrics. If that’s not enough, you also get access to a tool that scores your essay and gives you instant⁸ feedback on how to improve it. At $3-9/per draft,⁹ it’s 10-40x less than market, and once you’ve uploaded, submission is free.

What makes this radical is that you have access to the same tool that I’ll use to shape the longlist. This means, if you’re motivated, you can get into an iterative loop and get as close to a platonically perfect essay as you can. Even if you lose (and of course, almost all entrants do in fact lose), you leave with hopefully the best essay you’ve ever written, and new insights on the revision process.

I know what you’re thinking: isn’t this hackable? If everyone has the judging tool, won’t everyone get a 5/5? Won’t trolls discover AI hacks and get perfect scores by uploading gibberish?

Well, the main insight is that AI generation is significantly harder than AI analysis. I fed Claude my textbook (which has all the answers) and asked it to turn some notes into a perfect essay; it only got a 3.6. For reference, a 3.0 is good, a 4.0 is great, and a 5.0 is world-class. From my beta testing, writers found they could increase their scores by 0.5-1.0 with editing. At a certain point though, your form hits a limit, and you can only advance by significantly restructuring your essay. After five rewrites, I got the essay you’re reading from a 2.61 up to a low-4.¹⁰ And even if you do find a way to technically hack the software to score a perfect 5 with a slop submission, it would be pointless because the winners are ultimately determined by human readers.

The judging process has three pillars of equal weight: (1) Essay Architecture scores on objective composition patterns (also, I’ll be manually scoring the longlist to make sure it’s accurate); (2) a panel of readers will be scoring for non-compositional, subjective values (do they feel the essay is culturally potent, timeless, singular?); and (3) I’ll invite a guest judge to serve as a wildcard, stack ranking the short-list based on their own taste. The top essays from this competition will be well-crafted, well-liked, and recommended by a trusted judge. The whole triad is covered: logos, pathos, and ethos. All three branches are present: judicial, legislative, and executive. This is as American as it gets. Can a process like this make a substantial difference in the quality and success of an essay anthology?

---

III: MORE SYNERGISTIC: This is built for independent online writers. You’re allowed to publish essays that are already on your Substack, and if you’re a finalist, you’ll be featured in and earn royalties from a Metalabel anthology.

---

I’ve noticed that essay competitions aren’t too popular among online writers, and I think it’s because of friction around publishing rights. These institutions typically want “first North American serial rights.” This means, while you still keep the copyright, they want to be the first outlet to publish your work. You’ll see things like “Entries must not have been previously published, either in print or online, or been accepted for publication elsewhere.” Some will consider simultaneous submissions, but only if you “notify them immediately” if the work has been accepted, generously giving you “24 hours to decide if you’d like to withdraw.” This is the language of a gatekeeper.

The Essay Architecture Prize is built for independent writers who already self-publish, which means you’re welcome to submit anything: unpublished essays, or ones that already exist on your Substack. If you are a finalist, you’ll be asked to grant me “non-exclusive publishing rights,” which gives me permission to publish your work in an anthology of the winners. Unlike traditional anthologies though, I want writers to have equity in the book.

Here is a theoretical breakdown of how an essayist gets paid on their path to getting featured in a prestigious anthology. The writer is (sometimes) compensated by the journal they get published in: maybe it’s $0.20/word from The New Yorker, but probably under $500 for a Tier 2 journal, and often $0 for a smaller literary outlet. Anthologies often select essays from journals and prize winners. 70% run on prestige and offer nothing other than free copies of the book, 20% offer a flat fee (maybe $200), and 10% offer royalties. I would guess that a large anthology like Best American Essays doesn’t offer royalties. If we generously assume it sells 50,000 copies per edition at $20, and then factor in costs for distribution and printing and such, then Harper Collins might profit something like $250,000. At 22 contributors, that’s like $11,000 in value generated per essayist, but they earn something closer to $400.

This doesn’t feel like synergy, this feels like trading your work and time for a “featured in X” badge. What if writers owned 100% of the anthology?

Imagine if 10 finalists were assembled into a digital zine on Metalabel, a platform that automatically splits sales between the owners of a collection. If it sold 2,500 copies at $10, that would be $2,500 per owner (which makes the royalties function as a meaningful prize for being a finalist). Now let’s run the crazy simulation. In a fantasy where this becomes a commercial breakthrough—because unlike BAE, its obsession with quality leads to a collection that is well-rounded, trance-inducing, and recommendable—selling 500k copies of a $19.99 softcover, that’s something like $2.5 million in revenue, which, split 20ish ways, comes out to $125,000 per finalist.

Of course, I’m excited just to sell the first 100 copies, but the fantasy proves a different point: a business that pays writers well is, generally, a bad business. The idea of giving 100% of royalties to the original essayists makes absolutely no sense to Harper Collins, or to anyone aware of the complexities of book-selling. But what makes this fundamentally different from a publishing company is that the whole thing is anchored in software. It’s a different business model, one that happens to produce high-quality essay books. The hope is for the software to be successful enough so that I can make decisions from the place of what helps writers? and what helps essay culture? Synergy means the more I help the community (via tools, prizes, distribution), the more it helps me (bigger prizes + better books = word of mouth + more software revenue), which then fuels better tools, bigger prizes, farther reach, and on and on and on.

---

The Essay Architecture Prize

So this inaugural Essay Architecture Prize is an experiment to see if any of this is actually feasible. Are there enough entrants, submissions, and subscriptions to sustain software development and host a bigger prize next year? How accurate is the scoring? How does Substack respond to the concept and process? Where is the bottleneck in infrastructure? (Does the whole thing crash and burn the day before the November deadline?) Is the grassroots anthology we produce better than Best American Essays and how many copies can we sell if everyone promotes it? Did I enjoy running this and would I want to run it every year?

Out of all the businesses I could make from this tool, I’m most excited about the one where software and literature work together in a flywheel. Mr. Beast started by giving away $10,000 to his audience, and now he routinely does 7-figure giveaways. But, unlike Mr. Beast, this won’t use jaw-dropping prizes and fabricated situations to hack your attention; I see it like a Trojan Horse, an unreasonable sum of money that tricks a generation into manifesting the golden age of the Essay, where good ideas get amplified and inspire a wave of reading, writing, editing, thinking, and self-publishing.

---

Let me know if you have questions/ideas and I’ll make sure to address them in next post.

---

Footnotes:

1. By no means am I implying that you will achieve any sort of worldly fame by winning my inaugural, experimental essay prize—nor do I think fame is the point of winning. Jim Harrison (paraphrasing Honoré de Balzac) said “fame is the sunshine of the dead.” My interpretation: if you cling to fame instead of the creative process, you become symbolically, artistically dead, even though your reputation spreads like rays around the world. Recognition doesn’t have to be corrupting; don’t aim to win to be seen, aim to win to push the boundaries of what you can be.

2. I’ve become skeptical of the phrase “media company.” Even when one starts with a virtue-driven mission, it becomes very easy to become ideologically captured by the constraints of the business/market, where media gets selected/crafted less by standards, and more so by algorithmic demands, promotion opportunities, status boosters, upsells, metrics, etc. It sneakily turns into a propaganda machine while maintaining a facade of mission, so the creators don’t even notice. And so it’s helpful to clarify: don’t make media for the sake of maximizing reach, but make the media that won’t exist without your vision and figure out how to make it sustainable.

3. Before I wrote essays, I was exposed to competition culture through architecture school. I won 3 in-school competitions: for a recycling center (that got built), for an aviation museum, and for a modular construction technology for disaster-relief efforts. I lost 2 open competitions, one for a boating house in Central Park, and another for an experimental K-12 school in Cleveland. In every case though, it was an occasion to do my best and push my limits within a tight timeframe.

4. If you’ve never heard of Chill Subs before, they aggregate literary prizes and journals, and they just released a nifty “contest calculator” to evaluate how friendly a contest is for a writer. Check it out to get a sense of all the different design constraints that go into this.

5. My next post will cover the prompt of The Essay Architecture Prize in more detail, but I figured I’d leave an easter egg in the footnotes. It is something like “capture the essence of 2025.” I want you to write about a personal experience that is microcosmic of a larger phenomenon within this calendar year. It is open-ended enough for you to have multiple options, and yet also, it unifies all the submissions into a single collection that captures our zeitgeist.

6. Of the top 10 open essay prizes I analyzed, the average grand prize was $2,931 and brought in 834 entrants. This gives a baseline: estimated entries = 28% of the grand prize. It’s TBD how this changes at different scales. My first assumption would be that the rate gets lower with scale (ie: a limited number of writers will care to submit). Alternatively, this rate could increase since there is iterative feedback and unlimited submissions.

7. Public definitions coming soon! (The book will be hosted on my website on 9/15). Paid subscribers and software users will have access to the full wiki/book.

8. It is “instant” relative to the traditional contest turnaround of weeks/months. It still takes 15-30 minutes to get your final score (but you get pattern scores and feedback every 45 seconds). So maybe it’s 50x slower than a chatbot, but it still has a 50x quicker turnaround than a 24-hour return from a professional editor.

9. I surveyed a few hundred writers on what they thought this tool might cost per draft. A dozen writers said $0 (since chatbots are free), and another dozen said >$100 (anchoring this to the price of a human editor). Someone even said $2,500 (troll? believer?). The mean/median was $36/$10 per draft. I’m aiming to launch with a single-digit cost per draft ($9), and lower prices via subscriptions (ie: 4-10 uploads per month).

10. Note from a friend who read all versions of this essay: “Kind of blows my mind how much better the last two essays have gotten after the last revision.” I believe quality comes from destructive rewrites. A draft is not meant to be edited, but to be a springboard for a different, better draft, endlessly, until you get bored or run out of time.