Perell reads Popova one of the Sontag quotes she loves. Our task is not to find the maximum amount of content in a work of art. Our task is to cut back content so that we can see the thing at all.1
Popova lights up. I love that because she wrote that long before we used that horrific word "content" to describe cultural material today. The Sontag essay is from the 1960s. The word she's using has a specific old meaning — the intellectual content of a work, the what-the-work-is-about. Sontag's argument was that critics keep trying to find more content, and the over-finding obscures the actual work. Strip back the content-finding. See the thing.
Forty years later the word content has been captured by an entirely different economy. Content now means the material that fills the container. The container is advertising-supported web infrastructure. Content is what gets produced in exchange for the audience's attention, which is sold to advertisers. The word has been turned inside out.
Popova hates it. I'm allergic to the word content.1 She's never had advertising on The Marginalian in twenty years. Not a single banner. Not a single sponsored post. The reason isn't anti-commercial purity in the abstract. It's that the moment advertising arrives, the creative work becomes the content that fills the container. The container is the ad. The reader is the attention being sold. The writing is the bait.
The structural inversion has consequences. Once writing is content-in-the-container-of-advertising, the economics select for whatever maximally extracts attention. Clickbait. Listicles. Outrage. The lowest common denominator of cognitive engagement. The web has bent in that direction over twenty years not because writers got worse but because the economic form of the web rewards content-as-attention-bait and starves the work that doesn't fit.
Sontag's content was an honorific term used in a critical conversation. The content of this novel is the contradictions of modern self-formation. It meant the substance the work was carrying. The argument was about how to receive that substance rightly. Sontag thought too much content-finding got in the way of the work's actual force; the work's form was being neglected because the critic was over-focused on what the form was trying to say.
Sontag's content assumed the work was the destination. The critic was trying to understand the work better. The content was a property of the work that the critic was trying to read.
Content in the contemporary economy is the opposite. The work is not the destination. The work is the means to the destination, which is the audience's attention, which is then sold. The work is therefore fungible — interchangeable with any other work that produces the same attention. Quality matters only to the extent that quality produces more attention than non-quality. Once quality and attention diverge — as they do in many contemporary attention-markets — quality loses.
Popova's allergy is to the inversion. The word is the same. The position the word names has flipped.
She doesn't end on a moral note. I don't want to end on a cynical note as the old lady who's been online for 20 years. But she does describe what she's been watching. Content is rewarded in certain ways. Clickbait, which works because it sells pages, and they want to sell pages because they want to sell ads. There's been more and more and more kind of a shallowing of cultural material as the internet has moved more and more toward clickbait and listicles and 10 ways to.1
Twenty years ago, Brain Pickings (as it was called) operated in a web that hadn't yet fully bent. There was room for the long meditation on a 19th-century letter. The reader had time and attention to spend without that attention being commercially extracted. Twenty years later, the same essay has to compete with infinite attention-bait and the reader's attention has been trained on the bait. Brain Pickings didn't change. The web did.
She's moving away from the web as a primary form not on moral grounds, she says, but because as a human being who doesn't find it compelling. The form has eroded enough that the experience of being in it isn't worth the time anymore.
This page pairs with AI Cannot Suffer, Cannot Make Art (Popova) — same writer, same essay, related critique. The AI critique is about what the writer is. The content critique is about what the writer is producing inside of. Both are about the structural conditions of art-making being eroded by adjacent economic forms.
It connects to Annual Publishing Discipline (Wang) — Wang's twelve-month-essay cadence is his version of the same content-allergy. The Substack weekly schedule, the Twitter daily presence, the Instagram constant feed — all are content-economies in Popova's sense. Wang's annual cadence refuses to participate. Different operational form, same underlying critique.
It sits in productive tension with the creator economy tradition that has emerged in the last decade. The creator-economy frame says: writers can sustain themselves through direct audience relationships, no advertising required. This is partially true. Substack monetizes through subscriptions, not ads. The model is in the same family Popova is reaching for. But the cadence Substack expects (weekly newsletters) imports a content-orientation through the back door — the writer must produce content for the subscribers' attention at regular intervals. Popova's discipline of write when there's something to write doesn't fit Substack's economics. The vault should preserve the tension. The creator economy is a partial solution to the problem Popova is describing, not a full one.
Popova's site has run for twenty years without advertising. This is operationally unusual. Almost every other site of comparable longevity runs ads or has at some point. Many of the comparable sites have changed substantially over twenty years as their advertising contracts changed — the kinds of pieces they publish, the headlines they use, the lengths they tolerate.
The Marginalian hasn't changed in those ways. The pieces still tend to be long meditations on writers and thinkers, often dead, often less famous than the famous ones. The headlines are direct rather than baity. The reading experience is slow rather than scroll-optimized. The site hasn't optimized for engagement metrics because the engagement metrics aren't what's keeping the site funded.
What's keeping the site funded is direct reader support — donations, sales of books and prints, certain affiliate links that don't reshape the editorial. The model is unusual. It works because Popova built it over twenty years before the alternative models became standard. A writer starting now would struggle to replicate it; the audience habits that supported Brain Pickings in 2008 don't exist in the same form today.
The point isn't that Popova's model is generalizable. The point is that the model is possible — that writing outside the content-attention-advertising stack is operationally feasible, and the work that gets produced outside the stack is structurally different from the work produced inside it. The work outside the stack carries the weight of not having been bent for attention. The reader can feel the difference.
You write online. You've been doing it for a while. You're producing pieces on some cadence. Audit the cadence.
Ask: am I producing pieces because I have something to say, or because the cadence requires a piece this week? If the cadence is determining the production, you're inside a content-economy regardless of your subjective intent. The form has captured you.
Ask: am I writing headlines for what the piece is about, or for what will perform on the platform's distribution algorithm? If it's the latter, you're optimizing for the container (the algorithm's reach) rather than the work (the piece itself). The form has captured you.
Ask: do I edit pieces shorter for the algorithm or shorter for the work? Some pieces benefit from being short. Some don't. If you can't tell the difference for your own work, the algorithm's preferences have become indistinguishable from your editorial judgment. The form has captured you.
The capture isn't always avoidable. Most writers need income, and the income comes through platforms that have their own logics. Popova's twenty-year independence isn't broadly available. What's available is partial resistance — refusing to optimize headlines for performance, refusing to compress pieces beyond what the work warrants, refusing to publish on a cadence that exceeds what you have to say. Each refusal is a small piece of editorial sovereignty reclaimed. The aggregate effect over a career is meaningful.
You've been captured when your writing has predictable rhythms tied to the platform's expectations — newsletters every Tuesday morning, threads every weekday afternoon, year-end lists in December. The rhythm matches the platform's incentives rather than your interior production. The pieces you produce inside this rhythm have a particular sameness — they're skilled, they perform reasonably, they don't surprise you.
You're outside the capture when your production looks erratic from outside. Three pieces in a week, then nothing for two months, then a long piece that doesn't fit any of the platform's expected forms. The pattern is uncomfortable for the platform's distribution architecture. It's also the signature of writing that's tracking something real rather than feeding the form.
The structural critique Popova is making has substantial overlap with media-studies literature on the attention economy (Tim Wu's The Attention Merchants, Jenny Odell's How to Do Nothing, the broader Stanford Internet Observatory work). Popova is in the methodologically-defensible part of the critique.
The tension: the alternative model Popova represents is partially unreplicable. Most writers can't sustain twenty-year ad-free operations through reader donations. The model depends on early-mover advantage, on Popova's specific public persona, on the existence of a generation of readers who learned to support her in the pre-attention-economy era. Writers starting in 2026 face structurally different conditions. The principle (resist content-capture) is portable. The operational form Popova achieved is not.
A second tension. Some content-economy producers do produce good work. Not all clickbait is empty. Not all algorithmic optimization is incompatible with depth. The line between capture and adaptation is blurry in practice. Popova's allergy is to the form's structural pressure, not to every individual writer working inside it. Some readers will say her position is too absolute; some will say it isn't absolute enough. The vault should hold the tension.
Two writers in adjacent traditions arrive at similar critiques from different angles.
Macfarlane warns against Grammarly. Grammarly is an algorithmic optimization tool. It smooths writing toward statistical norms. The norms are the writing the algorithm has been trained on, which is the writing the prior generation produced inside content-economies. Macfarlane resists Grammarly because the smoothing erases distinctiveness. The mechanism is structurally identical to Popova's resistance to content-economies — both are about the form imposing patterns the work didn't earn.
Wang's annual-essay discipline is a similar refusal. Wang produces one essay a year. The piece is whatever has been accumulating in his attention for twelve months. He doesn't optimize for engagement, doesn't follow algorithmic feedback, doesn't trim for a platform's expectations. The piece is what it is. Wang and Popova are running the same operational refusal at different scales — Wang's annual cadence and Popova's twenty-year independence are versions of the same discipline.
Three writers, three angles, one underlying critique. The form selects against work that doesn't fit it. The writers who do the work the form selects against must do it deliberately, against the gradient. Most writers can't sustain the resistance economically. The aggregate quality of the form's output is therefore degraded by the form's own incentives.
What Popova is naming shows up in two other places worth surfacing.
Eastern Spirituality: Practice Corrupted by Its Own Success — Contemplative traditions consistently warn that a spiritual practice becomes corrupted at scale by the institutional forms that grow around it. The monk's individual practice was the original thing. The monastery is built to support more monks. The monastery's economic needs start shaping the practice the monks can do. By century three, the institutional forms have substantially reshaped what was originally a personal liberation discipline into something that serves the institution's continuance. The mechanism is identical to what Popova is describing about the web. The original thing was the writing. The forms grew to support more writing. The forms' economic needs reshaped what writing could be. The traditions and the web are different content; the structural problem is the same. The cross-domain insight is that practice-corruption-by-institutional-form is a domain-general phenomenon, not specific to writing or to spirituality. Knowing this doesn't fix the problem but does help the practitioner see what they're up against.
Behavioral Mechanics: Medium as Message: Form Shapes Content — Marshall McLuhan named the mechanism in 1964. The medium shapes the content more than the explicit content does. Television produces certain kinds of stories because television's economic and technical form makes those stories most viable. The web produces certain kinds of writing for the same reason. Popova is running the McLuhan analysis on the web's specific economic form. The handshake is one McLuhan would have recognized. The medium has become the message. The writing has become the content. The reader has become the attention being sold. The form's incentives are the actual editorial.
The Sharpest Implication If Popova is right, the writers doing the deepest work are operating against the economic gradient of their own form. Each refusal is small. Each refusal accumulates. The aggregate result is that the writers who produce the work readers will return to in twenty years are the writers who refused to bend their work to the platform's incentives — and most writers cannot afford to refuse. The form is therefore structurally selecting against the work that would best serve readers in the long term. This is a structural failure of the writing-economy, not a personal failure of writers. It is also nearly impossible to fix from within. The writers who could change the form most are the ones who already operate outside it (Popova) and therefore have least leverage on its incentives. The writers with most leverage on its incentives are the ones who depend on its incentives most. The fix, if it comes, will probably come from outside the form — patronage models, institutional support, structural changes in how readers fund the work they want. The forms that would enable this exist in fragments. They haven't coalesced yet.
Generative Questions