Before the interview starts, Popova taps Perell on the knee. Less time on LinkedIn, more time in archives.1 The instruction is half-joking. It's also the actual operational position she lives by.
She starts unpacking it. The internet is the surface of the ocean. Most of what has been thought, felt, written, created, drawn — dwells in university basements and library back-rooms and never made it online. People assume that if something isn't searchable, it doesn't exist. She finds this insane. So much that has shaped the present we live in more profoundly than the forty-year depth that's the internet.
She gives a small example. The 92nd Street Y in Manhattan. One of the most revered literary venues in the country. They hosted Pablo Neruda's first U.S. appearance. They hosted Susan Sontag's first major lecture. Those recordings exist. They're on reels. The reels are in the building's basement. They have not been digitized. Nobody knows they're there unless you go and spend time with them.
Multiply that by however many such institutions exist in the world. Public libraries with manuscripts and rare books. University archives with the working papers of writers and scientists. Religious archives. Municipal records. The recorded oral histories that someone made in 1972 with veterans who are now all dead. None of it on the internet. All of it real events in the history of the world, in the history of the creative universe, erased from the simulacrum of memory that is the internet.
Popova has been doing this for seven years for one book. She wrote about Mary Shelley. To write about Mary Shelley honorably, she had to read all of Shelley's existing writings — public and private. Public Shelley is digitized. Private Shelley sits in the Bodleian Library at Oxford. Mary Shelley's journals and diaries are extraordinary.1 The journals tell you who Shelley was in ways the published novels cannot.
Walt Whitman's notebooks are at the Library of Congress. Popova went. She read them. The notebooks tell you about the writing of Leaves of Grass in ways no biography could reach because no biography had read them at this density.
The pattern repeats. She decides to write about a 19th-century figure. She finds out where the unprocessed material is. She goes to it. She sits with it for days or weeks. She comes back with the material that the digital sources didn't have. The book she writes is therefore different from every other book about the same figure, because she's the writer who actually went to the unprocessed sources.
It's a real responsibility to do it honorably, to do it as accurately as you can, she says about writing across a 200-year gap. If you only go to what's preserved digitally, you're going to really misrepresent a life. I mean, it's not a small thing.
The deeper claim Popova is making, even when she doesn't quite say it, is that the internet has trained a generation of writers to mistake "what comes up in search" for "what exists." If a writer Googles "Mary Shelley" they get a curated digital subset of what's been published about Shelley. They miss the journals at Oxford. They miss the unpublished letters at various university collections. They miss the manuscripts with marginalia that have never been transcribed.
A book based only on what comes up in search will read like every other book based only on what comes up in search. It will represent Shelley as the search-results represent Shelley. The places where the search-results are misleading or incomplete will become the places where the book is misleading or incomplete. The reader cannot tell the difference because the reader hasn't been to the archive either.
Popova's discipline is to go past the search-results to the actual stuff. The actual stuff is harder to access. It requires travel, appointments, archivist relationships, hours of sitting with handwriting you have to learn to read. It produces material no Google search can produce. It produces books no all-digital-research writer can produce.
Popova mentions one specific accessible entry-point. The New York Public Library used to run a program called After Hours — once a month, archivists would bring out items from the collection that aren't normally on public view. Lewis Carroll's diaries. One of the four existing copies of Copernicus's De Revolutionibus. The archivist would be in the room and tell you about the item. You could ask questions. You could see the actual physical object.
This is what archives can be when institutions decide to make them accessible. Most institutions haven't decided yet. Popova thinks more should. The afterthought of her more time in archives instruction is partly that the institutions also have responsibility — to bring the material out from the back rooms, into rooms where curious people can encounter it.
This page pairs with Map vs. Territory (Hari). Hari's territory is the living people who hold the material; Popova's territory is the archived material that holds what dead people thought. Different forms of territory. Same discipline of going to it instead of trusting the available maps.
It connects to Observation-to-Contemplation Diary Form (Popova) — Popova's own diary practice is the form she's reading in others. Her diary work and her archive work are parts of the same operation, the attending to specific direct material as the actual source of writing.
It sits in productive tension with the thinkpiece tradition that synthesizes existing digital coverage. Most current essays are built from other essays, which were built from other essays. The chain rarely terminates in primary material. Popova's discipline forces the chain to terminate in actual archived stuff. The work that results is different in kind from the synthesis work.
Popova wrote a book recently that took seven years. Most of the seven years was archive work. Not just Oxford for Mary Shelley's journals — she read everything Shelley wrote, public and private. She read everything other people wrote about Shelley while Shelley was alive and the decades after. She read the journals of people who knew Shelley. She read the correspondence around Shelley's mother (Mary Wollstonecraft), Shelley's father (William Godwin), Shelley's husband (Percy), and Shelley's complicated relationships with Byron and Polidori.
By the time she sat at the keyboard, she had spent seven years inside Shelley's actual material. The Shelley she could render was not the Shelley anyone else could render, because no one else had been with the actual material in that density. The book that resulted is a book only Popova could have written, not because Popova is a uniquely gifted writer (she is, but that's not the operative claim), but because the discipline of going to the unprocessed sources produced material that the discipline of digital research could not produce.
The seven years is operationally significant. Most book projects don't have seven years of budget for primary research. Most book projects can't produce work at the same depth. The depth-quality of Popova's book is downstream of the time-investment in the archives. The trade-off is direct.
You have a writing project that needs primary material. Three operational moves.
Identify what unprocessed material exists. This requires some research. Where did your subject live? What institutions hold their papers? Are there working papers, correspondence, marginalia, drafts in any of these institutions? Most of this information is itself sometimes hard to find, but a librarian at a major research institution can usually point you. Start with the institution's finding aids — the documents that describe what's in a given archive. Most archives have these online even when the underlying material isn't digitized.
Contact the archive. Most institutions require advance arrangement. Some require academic credentials; others are open to independent researchers. Be specific about what you want to see. Be respectful of the archivist's time — they're the gatekeepers, and they decide whether your request gets prioritized.
Go. Sit with the material. Take notes. Photograph what you're allowed to photograph. Build a relationship with the archivist; they often know what else is there that you didn't think to ask for. Plan to come back. Substantial archive work usually takes multiple visits — the first visit is mostly orientation; the actual material work happens on visits two through five.
Most working journalists and essayists do not do this. The barrier to entry is in the time and the un-glamorous patience required. The reward is producing work no one else could produce, because you went to material no one else has been to.
Your book is built only from what comes up in search when your bibliography contains only published secondary sources, no archival citations, no manuscript references, no working-paper citations. Reviewers will notice. Readers who've been to the archives will notice. Other writers in your field will notice. The work will be received as competent within the digital corpus and thinner than the work that includes primary material.
You're running Popova's discipline when your bibliography includes archive locations, unpublished material, and material you saw because you went and looked. The work will be received as primary research, which is operationally different from secondary research even when the writer can't always articulate the difference.
The convergence across writers who run this discipline is substantial. Caro's Lyndon Johnson research, Patrick Radden Keefe's investigative work, Jill Lepore's American history, MacCulloch's Tudor and Reformation work — all involve substantial archive presence. The pattern is recognizable as the practice of writers whose books reach.
The tension: archives are unevenly distributed. Most accessible archives are in major cities in wealthy countries. Writers without travel budgets, without institutional affiliations, without proximity to the relevant materials, cannot run the full discipline. The principle is portable but the operational form is not. Digital surrogates (when they exist) are partial substitutes. The substitution is real but incomplete.
A second tension: the digitization frontier is moving. Material that was un-digitized ten years ago is now searchable. The advantage of the discipline shrinks as digitization expands. But the depth-frontier moves with it — there's always more un-digitized material than digitized, and the un-digitized portion is usually the more interesting portion (rare manuscripts, working papers, unpublished correspondence). The advantage persists, just at a different layer.
You spend a year reading every existing book about a historical figure you want to write about. By month three, you've started noticing that the books are recycling each other. The same anecdotes appear. The same quotations are repeated. The same interpretive framings recur. Each book adds a small layer to the prior consensus and changes very little.
Now picture spending the same year at the archive that holds the figure's working papers. By month three, you've encountered material that no published book mentions. A letter that contradicts the consensus story. A draft that shows the figure working out a position you'd been told they held automatically. A marginal note that complicates everything. The archive is producing material the secondary literature doesn't have, and once you've seen it, you can't go back to the secondary version as the truth.
The difference is what Popova is pointing at. Wang has a related discipline — manuscript copying as voice training — which is a small instance of the larger principle. Going to the original material changes you. You can't substitute reading-about for being-with. Different kinds of writers reach this conclusion from different angles. The convergence is strong enough to mark as a load-bearing principle of serious nonfiction work.
The archives discipline shows up in two other places worth naming.
Eastern Spirituality: Direct Experience vs. Doctrinal Knowledge — Eastern contemplative traditions distinguish anubhava (direct experience) from shastra (textual knowledge). The teacher who has only shastra is regarded as incomplete; the teacher who has anubhava is regarded as authoritative. Popova's writer is the journalistic version of the same distinction. The writer who has only the published secondary corpus has the shastra. The writer who has been to the actual archived material has the anubhava. Both communities arrived at the same hierarchy of knowledge — direct contact outranks accumulated commentary. The traditions developed centuries of vocabulary for distinguishing the two. Popova is running the discipline without the vocabulary, but the architecture is the same.
Behavioral Mechanics: Observation as Strategic Superiority — Military intelligence and competitive-strategy traditions both privilege direct observation over intelligence reports. The reports drift; the observation doesn't. The investigator who goes to the archive has the same structural advantage as the commander who walks the terrain. The accumulated reports may say one thing; the direct contact with the territory may show something else. The commander who skips the walk loses the diagnostic that would have corrected the reports. The writer who skips the archive loses the same diagnostic at a different scale. The principle is domain-general across strategy, journalism, and historical research.
The Sharpest Implication The current writing-economy makes archive work nearly impossible for most working writers. Trade publishers want shorter lead-times. Substack rewards weekly cadence. Newsletter audiences expect regular content. None of these compatible with the seven-year archive-immersion that Popova's Shelley book required. The implication is structural: the writers doing the deepest work are operating outside the economic structures most writers live inside. Popova is one of the few writers who have figured out how to fund the discipline. Most can't. Public-facing writing as a field is therefore being structurally pulled toward digital-only research even where archive material would dramatically improve the work. The aggregate quality of public-facing nonfiction is degraded by the structural constraint. Few people in the field talk about this honestly. The cost is invisible until you've read a book that did the archive work and compared it to a book that didn't.
Generative Questions