How he reads

Every Greene book starts as a pile of other people's books. Before he writes a page, he reads for a year or more, marks up what he reads, and comes back later to turn the marks into cards. This chapter is about the first half of that: what he reads, how much, where he finds it, and how he reads it. The cards themselves get their own chapter.

The material comes from 22 interviews he gave between about 2013 and November 2026, plus a few written sources: his old blog, a 2012 Mixergy interview, a 2013 Jeff Goins interview, and Ryan Holiday's posts. Chapter 1 explains how the transcripts were pulled and checked. The quotes below come from YouTube auto-captions, lightly cleaned, so treat the exact wording as close but not perfect.

One thing to keep in mind throughout. Greene talks about his reading a lot, but almost always in passing, as a story told to a podcast host. Nobody has published his reading lists. His books have bibliographies, but I found no count of their entries anywhere, and I did not count them myself. So the numbers in this chapter are his own estimates, given from memory, years after the fact. Where they disagree, I say so.

How much he reads#

Here is every reading figure I found, by book.

A few things stand out.

The per-book range is 200 to 400. The low end comes from a blog post that paraphrases a podcast. The high end comes from Greene himself, twice, eight years apart. The middle, "well over 300," is the only figure tied to a specific book he wrote alone. I'd use 300 as the working number, with 200 and 400 as the edges.

The unread pile matters. For The Laws of Human Nature he had about a hundred more books he meant to read and didn't, because "I did all of that research by myself. I had nobody helping me." So the reading list ran to 400 or more. The cut was made by time, not by judgment.

The Napoleon figure moves. In 2018 he told Brad Carr "maybe 4,000 pages." A year or so later he told Olivier Roland "about five biographies of him, maybe 3,000 pages." The 4,000 includes David Chandler's The Campaigns of Napoleon, which he puts at "about thirteen, fourteen hundred pages" on its own. Both figures are estimates from memory, more than a decade after the war book came out. Either way, it's a lot of pages for the main figure of one book. If Chandler alone is 1,400 of them, five biographies averaging 600 pages gets you close to 4,000. The two numbers may describe the same reading counted differently.

"Three books per paragraph" and "12 books per paragraph" are both about the Sublime. He said the first to Perell and the second to Holiday, both in the launch month. They don't really conflict. Neither is a count. Both are ways of saying the book was dense with reading, and the "12" came in a passage about how much time he had, so it reads as the upper bound. The Sublime has 12 chapters and runs 496 pages in the Viking edition. If every paragraph needed three books of its own, a 496-page book would need thousands. In practice the same books serve many paragraphs, so read these as a measure of density, not as a ratio you could multiply to get a reading list.

The 800-page Aztec Philosophy is the clearest single data point. He describes it as "on the surface quite boring" because of the detail, and then as "absolutely fascinating" once he got into how the Aztecs "thought in terms of shapes and patterns." He doesn't give an author. The title and length fit James Maffie's Aztec Philosophy: Understanding a World in Motion, but Greene doesn't say so, and I haven't checked which book he meant.

How long the reading takes#

The 2014 Rogan interview is the only place where he describes the research as a separate phase with a length. He was a year into The Laws of Human Nature at the time, still reading.

I'm not usually writing because my books require so much research. So right now I'm in a research period where I'm reading voraciously books about human nature, psychology, etc. And then in about a year I'm going to start writing.

That's from Rogan, 2014. Rogan's follow-up, "research for a full year before you write," got a yes. At Google in 2018 Greene said he had decided to write the book "about seven years ago" and "started devouring literature on the subject: neuroscience, evolutionary biology, psychology, psychoanalysis" (Talks at Google). Put together, Human Nature had something like two years of reading before the drafting began, though the reading never fully stopped.

For Mastery, his blog lays out the stages: first neuroscience and the evolution of the brain, then "biographies on as many iconic Masters in all fields as I could reasonably accomplish," then interviews with nine living masters (Greene's blog). Reading comes in waves, by field, with the theory first and the lives second.

The Sublime was different. He was writing chapter by chapter and reading for each chapter as he went: Weinberg's The First Three Minutes for the cosmos chapter, then "evolution and the dinosaurs," then "pagan religions" (Holiday live, 2023). After the stroke, research and writing ran in parallel. Chapter 8 covers how that worked.

Where the books come from#

The library years#

The first book was researched in libraries. He says so with some nostalgia, to David Perell in 2024:

Back in the day when I did the 48 Laws of Power, I'd actually go to a library, if you can believe that, right? I'm that old. And the library, you know, was like a magical journey.

The next sentence is the one people remember. In the section on con artists, there were "like 200 books. I'm like in Disneyland" (Perell, 2024). That shelf shows up all over The 48 Laws. Several laws open with a con artist's story, and Greene describes the trick of those chapters as leading the reader along before revealing the people are con artists (Talks at Google).

He learned library research before the books. In Hollywood he worked as an assistant to a director and to the director's wife, a screenwriter, who wanted to write a screenplay about coyotes. "I'd go to UCLA library and… honed some of my researching chops there" (Perell, 2024). He was also a researcher on a television show whose job was to find stories that could be produced, and he says he found "like 90% of the stories" on a team of six or seven (Danica Patrick). Chapter 2 has the rest of that history. The point here is that by 1995 he had years of practice at one specific task: walking into a stack of books and coming out with a story that works on a screen or a page.

The library also gave him his subject choices. When he started The Art of Seduction, he says, "I looked at a library… You could count on two hands the number of books about seduction" (Roland). He reads the shelf twice: once for what is on it, once for what isn't.

The Amazon years#

He stopped going. To Olivier Roland:

I used to go to the library. I don't go to the library anymore. I do it all on the internet with Amazon and websites. But I'm looking for books that have exciting new ideas, but that have things in it that are dramatic as well.

That changes the economics of the reading in a way he doesn't talk about directly. A library lets you pull 20 books off a shelf, read the first chapter of each, and put 18 back. Amazon makes you buy before you can triage. He describes the result as a gamble: "I look as closely as I can to see if a book I think is going to be interesting, and I buy it. And sometimes it is, and sometimes it isn't worth it" (Roland). So the books pile up at home, which is where the "8,000 books" line comes from (more below).

After the stroke the home became the only library. "All I can do is sit in my little room in my house at my desk and write the book… go through the books that I'm reading" (Diary of a CEO, 2024). He says the book is "so much better because of that."

The out-of-print rule#

He has a stated preference for books nobody else is reading. To Roland:

I like to find books that no one has heard of that were published in 1930, that are completely out of print, but that are incredibly profound, because then I know nobody else has covered this before. I'm the first. I'm like bringing this back to life.

The logic is about originality of material, not of ideas. If a story comes from a bestselling biography, his readers may already know it. If it comes from a 1930s monograph, they almost certainly don't. That is one reason his books feel full of things you haven't heard before, even when the underlying idea is old.

The best example he gives is The Power of the Charlatan, which he calls "an absolutely fantastic book, way out of print" and "a gold mine" for The 48 Laws (Perell, 2024). He remembers it this way: written "in the 30s" by "a German who was seeing Nazism rise," who never mentions Nazism but writes about 17th-century charlatans in a way that makes the parallel plain. He says he can't remember whether the author was a man or a woman.

The book is real and the description holds up. Its author was Grete De Francesco, born Margarethe Weissenstein in Vienna in 1893. Die Macht des Charlatans came out in Basel in 1937, and Yale University Press published Miriam Beard's English translation, The Power of the Charlatan, in 1939 (Wikipedia). She was Austrian rather than German, and she died in early 1945, probably at Ravensbrück. A German reprint came out in 2021. Greene's guess that "somebody might have brought it back" was right.

That's a useful check on his memory. Thirty years after reading it, he had the decade, the subject, the political subtext and the out-of-print status right. He had the nationality slightly off and the author's name gone. That is about what you'd expect from a strong reader who doesn't keep a reading log.

Roland caught another example on Greene's table during the interview: a French edition of Le Voile d'Isis, Pierre Hadot's history of the idea of nature, which traces two attitudes toward nature back to ancient Greece. Greene said he'd found it through a reference in another book, and that he'd read it "even if it's not completely relevant to my new book" (Roland).

Following the footnotes#

How does he find a 1937 book on charlatans? Mostly by following references from one book to the next. To Roland:

I'm reading a book and it's interesting, and then it references another book, right? And I go, "Hmm, that sounds interesting." And then I'll go look at it on the internet… it references three or four other books, that leads me to three or four other books, you know? I have to be careful because it's going to explode, and I'm going to have 8,000 books in my house.

The "8,000 books" is a joke about runaway growth, not an inventory. But the arithmetic behind the joke is real. If each book points to three or four others, three levels of chaining from one seed gets you to somewhere between 27 and 64 books. Do that from ten seeds and you have hundreds. The 300-book reading list is what's left after he stops the explosion.

He also follows less formal pointers. For the animals chapter of the Sublime, he had an octopus story that "was good, but it wasn't great." Then he noticed a blurb by Werner Herzog on the back of one of his books, recommending another book. "I bought it and it was absolutely perfect. It was one of those golden books that fit the chapter perfectly" (Perell, 2024). He doesn't name the book. The method is the same: one book points to the next, and the pointer can be a footnote, a bibliography entry or a quote on the back cover.

Researchers have a name for this. In systematic literature reviews it's called snowballing: you start with a set of papers, follow their reference lists backward and the papers that cite them forward, and repeat. Claes Wohlin's 2014 guidelines for software engineering reviews treat it as a real alternative to keyword searches in databases, with the caveat that the result depends heavily on the starting set (Wohlin, EASE 2014). Greene only does the backward half: he follows what a book cites, not what cites it. And he has no rule for when to stop, other than the size of his house and the deadline.

Triage: the first few pages#

With hundreds of books coming in, most of the work is deciding what not to read. He says he can tell fast. To Roland:

I have a nose for books that are giving me the material that I want, and I can tell when I open the book, cuz I ordered it on Amazon, I'm not really sure if it's good or not. I can tell in the first few pages this isn't what I want, I don't even bother reading it.

He told Perell the same thing in 2024. The first step "is finding the book," and "there's a lot of sifting through garbage, things that… don't have what I need… or stories that don't quite click." Can he usually tell right away? "Usually you can. You have a nose for it. But sometimes you start reading. You read, 'Oh man, this isn't going anywhere. Okay, I'm going to skim the rest of it'" (Perell, 2024).

So there are three outcomes, not two:

1. Dropped in the first pages. The book doesn't have the drama or the ideas he needs. He doesn't read it.
2. Skimmed. The book seemed promising and then stopped delivering. He skims the rest, presumably for the odd story or fact.
3. Read slowly, all the way. The book earns a full read with a pen.

What is he looking for in those first pages? He names two things. The ideas have to be "exciting" and "new," and the telling has to be "dramatic" (Roland). The second matters more than it sounds. His chapters open with stories, so a biography that lists facts in order is less useful to him than one that already has scenes in it. "So, I have to read books that cover it in that way, that describe the character, that make the character come to life," he told Roland, using Stalin as the example. A dull book about an interesting person can still be useful, but it costs him more work, because he has to build the drama himself.

A contradiction about skimming#

In 2014 he told Rogan the opposite. "I read three, 400 books and I read them, I don't skim them" (Rogan). In 2018 he told Brad Carr that "bad books or mediocre books I can skim" (Brad Carr), and in 2024 he told Perell he skims the rest of a book that isn't going anywhere.

The simplest way to square these is that the 300 to 400 are the books that passed triage. He reads those, and the skimming happens to the ones that get dropped or demoted along the way. The Rogan line came in a passage about discipline ("I'm so disciplined it's probably bordering on OCD"), so it was a boast as much as a description. I'd trust the later, more detailed accounts: weak books get skimmed, strong ones get read whole.

Reading slowly, with a pen#

The strong books get read slowly, and the reason is the pen. To Brad Carr in 2018:

I always write in the margins. Some people hate that, they feel like writing in a book is sacrilegious… but I'm not like that. I write a lot, and if I don't like a writer, and that happens often, and I get angry, I start putting giant X's and I say [expletive] you.

And then the other half: "Otherwise usually I like it and I'm putting exclamation mark, great, yes" (Brad Carr). He told Roland the same thing about a year later: "I put little marks on the side to highlight what I think is interesting. Sometimes I'll even write in it." Asked about sticky notes, he said no, "I write all through it… And if I hate a book, I'm… saying [expletive] you in the parts" (Roland). In the 2012 Mixergy interview he put it more plainly: "I underline it and put notes on the side and then I go back and put them on notecards" (Mixergy).

So the marks come in at least four kinds:

• Side marks next to passages he finds interesting. These are what the cards get made from later.
• Underlining, which he mentions in 2012.
• Exclamation marks and "yes," for agreement.
• Giant X's and insults, for disagreement.

He's explicit about what the pen costs him. "The problem here is it slows you down… unless it's a bad book." He thinks that's the right trade: "it's better to read one juicy book that grabs you, that really engages you, that really makes you think, than to read ten trivial books that don't have it" (Brad Carr).

The marks last. Talking about Rilke, whom he read in college, he says "all the books I have at home are from my college years and they still have my little notes in them" (Perell, 2026). He's been reading with a pen since at least his late teens. The habit may go back further: at Berkeley, studying ancient Greek, "you would have a passage of Thucydides and you yourself would write on the margins" (Perell, 2024). He links that marginal practice to the layout of The 48 Laws, with its quotes and fables in the margins. Chapter 5 covers the page design.

Why mark at all, rather than take notes while reading? Because the marks are only the first pass. The notes come later, on cards, after a wait of weeks or months, and the marks are what let him find his way back through a 600-page book. "Nothing is in my head," he told Roland, "because I can't possibly keep that all in my head" (Roland). The marginalia are an index he builds as he reads. Chapter 4 picks up from there.

Ryan Holiday, who learned the system as his research assistant, describes the same first step in nearly the same terms. "If you're not reading with a pen, you might as well not be reading," he says, and he folds pages and writes in Sharpie (Holiday's notecard video). Billy Oppenheimer, who now researches for Holiday, has a more elaborate code: underlining with a question in the margin, an all-caps phrase for anecdotes, boxed names and titles, and different folds for left and right pages (Oppenheimer). Greene's own code, as far as he's described it, is simpler: side marks, exclamations, X's.

Reading to bring the author back#

Greene has a theory of what reading is for, and he states it most fully in the Brad Carr interview. A written idea is a dead version of something that was alive in the writer's head. The reader's job is to bring it back. Brad Carr, 2018:

The material that you're reading, you tend to make it dead, and what you want to do is to make it all come to life. So reading a book is like bringing something dead to life. It's like taking Frankenstein and imbuing it with life.

He goes on: "when they had those thoughts that they're writing about it was something very exciting to them at the moment, very alive… whenever you write an idea down something is lost in the translation." So "your process is to get inside the writer and feel that excitement… it's partly empathy. In Mastery I call it thinking from the inside out. You go inside the book, inside the spirit, you try and recapture what they were feeling, not just what they were writing" (Brad Carr).

Then he gives Napoleon as the worked example. He read the 4,000 pages because he wanted "to feel like he's in the room with me." And he says the books didn't give it to him: "I found that the books I read were missing his genius… they were not quite getting at it." What he arrived at, after all that reading, was an interpretation none of his sources stated: that Napoleon's edge was "a mind for organizing information that was vastly superior to anybody else before the invention of computers" (Brad Carr). He told Jeff Goins the same thing in 2013: "Napoleon Bonaparte's genius was organizational" (Goins).

This is the most important thing in the chapter for anyone thinking about automating the process. The reading is not mainly about collecting facts. It's about reading enough of them, from enough angles, that a picture forms which none of the sources contains on its own. Five biographies of Napoleon each give you battles. The reader who has read all five and asks "what made him good at this?" might come out with "he was an information organizer." That claim might be right or wrong, but it is the author's, and it only exists because one person read all of it.

He applies the same idea to history in general. At the 2023 live event with Holiday, he said that over "the past 28 years… I've probably interacted more with the dead than the living," and that "98% of them bore the crap out of me." His complaint about dull history books is that "they just list George Washington on November 4th, he woke up… We make the dead people even more dead than they are" (Holiday live, 2023). His fix, when he writes, is to put the reader inside a decision as it was being made. He gives Caesar at the Rubicon as the example: "he's afraid, he's hesitant… it was a choice that he made."

He also reads for what's different about the past, not only what's the same. He mentions "a book called The Past Is a Foreign Country" and says that while the people of the past "have the same emotions as us… their cultural differences are vast." His own obsession was the court of Louis XIV, "the formality and the rituals," which "is not at all how we think nowadays" (Holiday live, 2023). The double reading (same emotions, different world) is what lets him move a story from 17th-century France to a modern office.

Reading people he dislikes#

He makes himself read writers who annoy him, and read them generously. To Andrew Huberman in 2023:

Steven Pinker, I don't like Steven Pinker, I find him really annoying, but I force myself to try and find a way to be seduced by him, to let him into my brain, to see where he's coming from, to open myself to the possibility that he could be correct.

He frames this as a form of intelligence: "vulnerability, letting people into your mental space." It's the same move as "thinking from the inside out," aimed at someone he disagrees with.

He also says he reads for challenge in general. "I read books that I think are going to confirm what I already know, and I like books that challenge me," he told Roland, adding that he's "learned to put myself back in square one each time I start a project" (Roland). That fits his ritual of starting every book by telling himself it will be his worst, which chapter 6 covers.

There's a tension with the marginalia here. The same reader who forces himself into Pinker's head also writes "giant X's" and insults in the margins of writers who make him angry. These aren't really in conflict. The X's are the first reaction, in the moment. The generous reading is a discipline he applies on purpose, to writers he has decided matter. Nothing in the interviews says which writers get which treatment, or how he decides.

The strangest example of reading across the aisle is one he says he'd never mentioned before. At the end of the 2022 Peterson interview he told Jordan Peterson that Maps of Meaning "was a very important book for me. I read it actually to help me with my war book… your notion of conflict and integrating internal conflict and external conflict" (Peterson). A book on myth and psychology, used for a book on military strategy. That's the cross-field reading the out-of-print rule and the footnote chains produce.

Reading for fun#

He says plainly that some of the reading is just to keep himself going. To Roland:

The writer has to be excited by the subject. And if you're not excited about it, it shows up in the book that the writer lost interest. So, I have to play a game with myself. I have to constantly make myself excited. So, sometimes I'll read books just because they're fun for me, and later on I'll figure out how to put them into my new book.

This is reading as maintenance. A five-year project means five years of reading on one subject, and he's said elsewhere that "if I had to write one book on narcissism for five years, I would kill myself" (Holiday live, 2023). Off-topic books are how he keeps the energy that, by his own theory, shows up on the page. "If what you're writing isn't alive in you," he told Perell, "it won't be alive in what you write" (Perell, 2026).

Note the second half of the Roland quote: "later on I'll figure out how to put them into my new book." The fun reading isn't wasted. He reads first and finds the use later. The Harriet Tubman story in the Sublime worked this way. He read about her "early on in the research process," didn't know where she'd go, and only found her place when he reached the chapter on the daemon (Perell, 2024). Reading ahead of need is part of the method, not a distraction from it.

What he reads#

Biographies, above all#

If one genre carries his books, it's biography. He says so to Barry Kibrick: "in the biography of anybody who's successful, which is something that I'm obsessed with, and this is the basis of so much of my research, there are always these moments in their adolescence, in their early twenties, where they're not powerful" (Kibrick, 2021).

Biographies turn up in nearly every account of a book:

• For The 33 Strategies of War, five or so on Napoleon, plus a biography of Gandhi that made him realise "Gandhi was actually a brilliant strategist" (Diary of a CEO, 2023).
• For Mastery, "biographies of all of the greatest masters that I consider worth reading: Da Vinci, Charles Darwin, Einstein, Henry Ford, you name it" (Goins).
• For The Laws of Human Nature, a 500-to-600-page biography of Mary Shelley for the envy chapter (Perell, 2024), a biography of Martin Luther King, and, while the book was still a working title, a biography of Abelard (NextShark).
• For the Sublime, a biography of William James (Kibrick) and several of Harriet Tubman.

One biography is rarely enough for a story he plans to lead with. On Tubman: "I realize there are pieces that are missing and I don't want to just depend on one book. So at the last minute I'll read a second, a third, even a fourth biography, or I'll read books by other slaves who escaped to get more color and detail" (Perell, 2024). That's the closest he comes to describing a cross-check. It's framed as a search for detail, not as fact-checking, but it does mean a lead story usually rests on more than one source.

The biographies are the raw material for the opening stories, which is the part of the work he says "has aged me a lot." Chapter 5 covers how a 600-page biography becomes a few pages of story.

Dense theory#

The second genre is theory, usually from psychology or philosophy. The analysis in each chapter (his "keys" sections) leans on it. His main example is Heinz Kohut, whose work on narcissism "grounds my whole theory of narcissism" in The Laws of Human Nature. Kohut's books are "very dense," and the job is different from reading a biography: "It's not a story. And then I have to go sift through this very abstract book and pick out the things that I can make real and put into layman terms" (Perell, 2024). In 2023 he was direct about where the ideas came from: "It's not my own theory. It comes from some great psychologists like Kohut" (Diary of a CEO, 2023). He also took at least one story from Kohut's clinical writing (Talks at Google).

He has a mixed relationship with academic writing. He calls most academic books on the sublime "so unsublime… The writer's not feeling it. They're going through all these theories" (Perell, 2026). The academics he names with approval are few: Harvey Mansfield on Machiavelli ("one of the few academics that I actually really like") and Johan Huizinga, whose Homo Ludens he calls "one of the greatest books ever written" (Williamson, 2024). On strategy he says "military writing is better than ever" and names van Creveld, Keegan and Luttwak, and points readers to the bibliography of the war book (Brad Carr).

What he likes best is philosophy written as story. His favourite example is Roberto Calasso, "one of my favorite writers that nobody will have ever heard of," who "writes books about the ancient world and he mixes it in with stories and anecdotes." The caption renders the title as "The Ruins of Kos," which is most likely The Marriage of Cadmus and Harmony. "That to me is philosophy because it stirs my soul" (Williamson, 2024). It's also a fair description of what Greene writes.

Science for the Sublime#

The Sublime pushed him into science more than any earlier book. "There's a lot of science in this book because I don't want to ground it in just your own emotions and your own imagination" (Williamson, 2026). The named sources:

• Steven Weinberg, The First Three Minutes. He mentions it in three interviews. To Kibrick in 2021: "a book that absolutely blew me away. It's a very technical book… written by a physicist" (Kibrick). To Williamson the same year he couldn't recall the name, "a great physicist, I forget his name, he died recently," and called reading about the Big Bang "one of the most earth-shattering experiences" (Williamson, 2021). It fed the first chapter, on the history of the universe.
• Anil Seth's chapter on octopuses in an anthology called Aliens, for the animals chapter. Greene retold its point that octopuses have neurons in their arms, "like 12 different centers of consciousness" (Theo Von).
• Stephen Jay Gould's book on the Cambrian explosion, with its pictures of "grotesque looking creatures" (Theo Von). He doesn't give the title. Gould's best-known book on the subject is Wonderful Life. Greene later says he wrote some of those creatures' names into the book (Williamson, 2026).
• Neuroscience, read since at least 2021 ("for my new book I'm going deep into neuroscience and the brain," Kibrick), including the UCLA researcher Jeffrey Schwartz on OCD (Williamson, 2024) and "a book about the infant's brain" (Williamson, 2026).
• Bone Games, a book about a climber's ecstatic escape from a mountain ledge, with "a lot of science in it" (Huberman, 2023).
• The Philosophical Baby, which he describes as "a book I didn't like but it had some interesting things" (Daily Stoic, 2026). It's a nice example of the X's and the generous reading in one sentence.

He reads science as a source of wonder, not of proof. The First Three Minutes passage to Kibrick goes on to say that in a hundred years the account "will be changed greatly," and that the point is that "these small little insignificant creatures" figured it out at all. That's the Sublime's argument in miniature. The science is there so the awe has something solid under it.

Scene research#

For some scenes he reads in a different way: not for a story or an idea, but for every physical detail of one moment. The best example is the 12-page scene in the Sublime that follows a woman walking to the Eleusinian Mysteries. To Holiday:

You have no idea the amount of research I went through to recreate that one story of 12 pages or whatever. So, I first had to look at maps of ancient Athens to figure out what neighborhood they came from… I read books about what women's life was like in ancient Athens… I had to read about the gate… Then I had to go through Google Maps.

The list goes on to flowers, shrines, torches, costumes, myrtle crowns, "the smells and the frogs." He calls it "a mosaic," and says "some of the details never made it in." For the dinosaur chapter he looked for recordings of what a mosasaur might have sounded like (Daily Stoic). For Sappho, he followed her love of the colour purple to the shellfish dye it came from (Daily Stoic).

This is reading for the "12 books per paragraph" line. A paragraph that puts you on the road to Eleusis needs maps, social history and botany, each for one sentence. Chapter 8 has more on how the stroke turned this into armchair travel.

Every day: the newspaper, and old favourites#

Two other kinds of reading show up as habits rather than research.

The newspaper. "I was reading the newspaper this morning, which I do every morning," he told Huberman in 2026, and went on to describe two science stories from that day's paper (Huberman, 2026). He makes similar references in other interviews ("I just read an article in the newspaper this morning," Williamson, 2026). Current events have made it into the books before. He says the Bitcoin craze, social media and the 2016 election fed into The Laws of Human Nature while he was writing it (Talks at Google). There's also a fragment at the start of the 2018 Brad Carr video, cut from somewhere later in the talk, in which he says "instead of reading the newspaper I have to read a biography and fit it in" (Brad Carr). Without its context it's hard to read, but it suggests the paper competes with the research for the same morning time.

Rereading. In the launch month of the Sublime he described a new habit. "I'm in this process where I'm rereading books that had a very big influence on me in my 20s. I don't know really why" (Daily Stoic, 2026). The book was a Dostoevsky novel, "probably my favorite novelist," about the nihilists whose movement fed into the Russian Revolution (Daily Stoic). To Perell he named it: Demons, also called The Possessed, which he's rereading "cuz I'm actually flirting with the idea of writing some fiction" and which he compares to "what's going on in America right now" (Perell, 2026). He also rereads Rilke (Letters to a Young Poet, the Sonnets to Orpheus, the Duino Elegies), in the college copies with his notes in them. "Whenever I read him, I feel like I'm 20 again" (Perell, 2026).

The early reading is part of the method too, in a longer loop. He names Nietzsche as "the writer that had the biggest influence on me," read at 18 (Williamson, 2026), and Castaneda's Journey to Ixtlan, read at 16, as the source of "a lot of the things in the 48 Laws of Power" (Theo Von). He traces the masks theme back to Richard Sennett's The Fall of Public Man, "a book that had a big impact on me many years ago" (Diary of a CEO, 2023). The 300 books per project sit on top of decades of reading that never went onto cards.

What he doesn't read, or doesn't say#

A few gaps are worth naming, because they shape what "reading" means in his case.

Archives. Nothing in the 22 interviews describes him working with letters, manuscripts, court records or other primary documents in an archive. His sources are books: biographies, histories, theory, science, plus some interviews for Mastery and the Sublime. The closest he gets to primary sources is reading memoirs, such as escaped slaves' narratives for the Tubman story, or the book "written by a woman whom he had seduced" that finally gave him Errol Flynn's secret (Diary of a CEO, 2023).

Languages. He studied French, Latin and ancient Greek, and Roland caught him with a French philosophy book. But the interviews never describe him reading sources in another language as routine research. Most of what he names is in English or in English translation.

A reading log. He doesn't mention one, and the Power of the Charlatan episode suggests there isn't one: he couldn't recall the author. The record of what he read is the marked-up books themselves and the cards made from them.

The internet. He uses it for lookups. "I use the internet, I use Google, I'm searching for like some factoid from my book, I find it, I use it" (Huberman, 2023). In 2026 he admitted using AI search on obscure queries: "you put in Scriabin and some weird thing that would take hours to find" (Perell, 2026). Chapter 9 covers his views on AI. For this chapter, the point is that the web finds the books and checks small facts. It doesn't replace the reading.

How this compares with research norms#

Greene isn't a historian and doesn't claim to be. But his books make historical claims on almost every page, so it's fair to ask how his reading compares with what researchers are trained to do.

Source type. Academic history leans on primary sources (documents from the period) and uses secondary sources (later books about it) for context and argument. Greene works almost entirely from secondary sources. That's normal for popular history and for the self-help genre his publisher files him under, but it means any error in a biography can pass straight into his book. Several reviewers of The 48 Laws made a related complaint: that the laws rest on isolated examples rather than tested principles (chapter 7 has the reviews). I found no historian's audit of his specific anecdotes.

Search order. Robert Caro, whom Greene mentions twice in the 2026 Daily Stoic interview as a model of patience, works in a fixed order: books first, then newspapers, then archives, then people, some of whom he interviews ten or twenty times (Paris Review). Greene's order is books, then more books found through the first books, with interviews added for Mastery and the Sublime. He stops at the first stage, and goes very wide within it.

Search method. A systematic review starts with a written search strategy, keyword searches across databases, and explicit inclusion rules, then adds snowballing on top (Wohlin, 2014). Greene's search is all snowballing, plus Amazon browsing and a "nose." His inclusion rule is drama and novelty, applied in the first pages. That would never pass a methods section. It's well suited to his goal, which is finding stories nobody else has used, rather than covering everything written on a question.

Reading depth. Here he exceeds the norm. Researchers under time pressure read abstracts and conclusions. Greene reads the good books cover to cover, with a pen, and then goes back to each one weeks later. He also reads adversaries on purpose and tries to argue their side in his head. That's closer to how a scholar reads the handful of books central to their field than to how anyone reads 300.

Cross-checking. For lead stories he reads a second, third or fourth biography. He doesn't describe that as checking, and he's honest that some numbers go in without much backing. On the claim that a large share of communication is body language, he told Steven Bartlett "I quoted the figure 95%, but who knows what it really is" (Diary of a CEO, 2023). And he says that for some chapters he writes first and researches after: "I'll be writing about Athena in one of my chapters and I'm not doing the research, I'm just writing off the top of my head, and then I go and do the research and realize you were right" (Danica Patrick). That's a confident reader trusting decades of background. It's also the order of operations that verification is meant to catch.

Disclosure. His books carry bibliographies, and he points readers to them. But nothing tells a reader which of the 300 books a given story came from, or which page. The cards have that information (book and page). The printed book mostly doesn't. Chapter 13 argues that this missing link is exactly what an AI version has to keep.

So the short version is: wide where a researcher would be systematic, deep where a researcher would skim, and light on the documentation that would let a reader check him. None of that is a scandal for a popular writer. It does matter for the second and third questions of this report, because the parts of his reading that are hardest to copy (the nose, the slow second read, the interpretation that no source states) are not the parts that make his claims checkable.

Known sources, by book#

This table lists every source Greene names in the interviews and written sources I read, grouped by the book it fed. It is not his bibliography. It's what he chose to mention on podcasts, which skews toward books he loved or found by luck. Where the captions or Greene give only part of a title, I say so.

A note on the 2014 row. When Rogan asked, Greene called his next book "a book about the Übermensch." It became The Laws of Human Nature. Whether Human, All Too Human made it into that book's cards, I can't tell from the interviews.

What this means for an AI version#

Most of this chapter's reading habits can be described as steps, and some of them can be handed to a machine.

• Footnote chaining is the easiest. Given a starting set of books, a script can pull bibliographies and follow references a few levels deep. The short post's pipeline already does this in its scout stage, three levels deep, with a human cutting the list.
• Triage on the first pages is harder but tractable. A model can read a book's first chapter and score it for the two things Greene looks for, drama and novelty, if someone defines them well enough to score.
• The margin pass maps onto marking passages with a page and a tag, which chapter 11 designs as its own stage.

Three habits resist that translation. The first is the out-of-print rule: the books Greene values most are the ones least likely to exist as clean digital text, which is a cost problem covered in chapter 10. The second is reading people you dislike generously, which only means something if the reader has views to set aside. The third, and the one that matters most, is the Napoleon move: reading five biographies until a reading forms that none of them states. A model can summarise five biographies. Whether it can come out the other side with "his genius was organizational," and whether anyone would trust the claim if it did, is the question the rest of this report keeps returning to.

Sources#