How I read

From Mohn und Gedächtnis, Anselm Kiefer, 2019-2020

---

1. Sometimes, I read for the pleasure of entering a mental world I hadn’t anticipated; sometimes, I read to be transformed. This essay is about the latter kind of reading.

2. When I want to be transformed, I chase my reading, to use Robin Hanson’s phrase. “Hunting has two main modes: searching and chasing,” Hanson writes. “With searching you look for something to chase. With chasing, in contrast, you have a focus of attention that drives your actions.” Searching is when I’m reading without a clear aim and continue to read even if I’m unsure about what the author is trying to achieve. Chasing is when I have a question I’m pursuing.

3. Chasing makes you more active and critical of what you read. This helps you learn more. As Hanson writes, “search-readers often don’t have a good mental place to put each thing they learn. [...] Chasers, in contrast, always have specific mental places they are trying to fill with what they read, so they better integrate new things they learn with old things they know.” When you chase, you continually ask yourself whether what you are reading “is relevant for your quest, or whether the author actually has anything new or interesting to say.” This means you drop books that don’t advance your understanding about the questions that matter to you, so you can find the books that do answer them and transform your thinking.

4. Of the roughly 300 books I start each year, I finish about 50. I skim a lot. Books are not sacred. I have to be ruthless in saying no to most of them (as well as to many other things in life) so I can spend an appropriate amount of time on the books (etc.) that really do challenge me and push the edges of my thinking. I once spent more than 100 hours thinking about 40 pages from Imre Lakatos’s Proofs and Refutations (at the end of serious meditation like that, I often find myself with notes for an essay about the ideas).

5. A surprising amount of the value of reading comes from stopping to think about what I’m taking in. I say surprising because I often feel like I get things right away. But with good writing, this is an illusion: it keeps getting more and more interesting the deeper I process it. As I read, I often do such things as close my eyes to visualize what the author is saying, reflect on how it connects to other things I know, and come up with counterarguments. I write in the margin. I dog-ear and return to important passages multiple times. A few days after finishing a book, I capture my thoughts in my notetaking system, which is organized along these lines. I also save quotes from every good book I read, as in a hypomnēmata. To be changed by my reading, reliably, I need to process what I read in an active way like this. If I read at a constant speed, page after page, I only half-understand most of it and retain little more than a fuzzy feeling from the book six months later.¹ (To retain more, I also make spaced repetition cards, which give you the option to remember something for the rest of your life for the price of ~5 min effort.)

Finding books

6. Needless to say, most writing does not deserve this amount of attention—I actually struggle to find 50 books a year worth reading. Sourcing good books is often my bottleneck, and I have had to put a lot of work into getting better at it. Here are some things I do:

7. When I get curious about something, I ask myself, “What is the best thing that has ever been written about this? Who has the deepest insight?” I’ll typically ask this in several places: I’ll ask on Twitter, I’ll email people who I think know about this problem, I’ll ask a few different language models—I gather a list of promising books.²

8. Then I interview the candidate books. Authors I deep-read will change me to some degree; they will restructure my brain. This is a big responsibility to give someone, so I think of picking a book as a recruitment process, and I do my due diligence to make sure they are up for the job. In How to Read a Book, Mortimer Adler gives a comprehensive overview of how to do this kind of evaluation before reading a book. What I do is some combination of (a) reading summaries (on Wikipedia, Goodreads, Stanford Philosophy Encyclopedia, etc.) to make sure the book is relevant; (b) checking what knowledgeable people have said about it, to evaluate if it is considered a high achievement in its field; (c) read the index of the book and the first few pages. That is, I spend 5–20 minutes gathering information that lets me evaluate if the book deserves to work on my brain. If I’m still excited, I order the book to my library and start reading it (without committing myself to finishing it if it turns out to be less useful than anticipated).

9. I often cluster my reading and read 5-10 books on the same topic to ensure I get a well-rounded perspective. When I read only one book, the topic seems easy to understand. The same is true if I read several books where the authors’s perspectives are closely correlated (sharing similar values, working in the same intellectual field, and so on). I get a false sense of clarity. But if I read multiple uncorrelated books about the same topic, I get a real sense of confusion. The authors contradict each other! They use different frameworks, so I don’t know if they are talking about the same thing! What is this!? The confusion hurts my head and forces me to wrestle with the material. I have to do such things as reconstruct the arguments and evaluate them. I have to translate between the different frameworks to understand if they are talking about the same things. In this process, I win my own understanding.

10. How do you ensure the books are uncorrelated? I tend to co-read books by authors from different disciplines. An economist, a philosopher, and a cultural anthropologist are going to reason about the flourishing of Ancient Greece in different ways, pulling on deep stacks of mental models and intellectual tools developed in their respective fields. I also prefer to read authors from different time periods. Jacob Burckhardt’s The Greeks and Greek Civilization (1898) has a very different view of Greek culture than H.D.F. Kitto’s The Greeks (1951), which has a different view than Robin Lane Fox’s The Classical World: An Epic History from Homer to Hadrian (2005). Modern authors have the advantage that they have access to more data. But I find their opinions easier to predict, so I often prefer older authors since they surprise me and therefore expand my worldview more.

11. Follow the citation trees. If I read something good, I always look at the index (or the links if it is a blog post) to see what they have been reading. Good writers and thinkers often have great taste. By moving along these intellectual graphs, you can find your way to some really interesting corners of our intellectual commons. As you map the connections, you tend to converge on the books, essays, and thinkers that are the highest peaks of their domain. I like to think of the first books that I surface when I get interested in an area as small streams that act as entry points to a river system. By tracing their influences, I can make my way upriver and into the mountains.

The habit of reading

12. Reading well is an endurance sport. I sometimes talk to people who want to become serious readers and so pick up Kafka’s The Trial or something like that—it is about as pleasant as running a marathon untrained. They often lose their enthusiasm for reading. You have to gradually ramp up your capacity to handle complex ideas and precise prose. I read a few hours a day, and I mostly read books that are comfortable for me to read, well within my range. It is more important to keep the reading experience easy enough that I keep going and going and going, than to always push myself to that edge. By reading within my comfort zone, I gradually build up my stamina and pick up more and more references, words, and patterns of thought, bringing more and more literature into my comfort zone. I remember reading Dostoevsky as a teenager, and I could do it, but it was a chore; these days his prose sounds like an email from a smart friend. It is thrilling when things that were beyond me become easy like that: the world cracks open. If you want to reach the deepest experiences literature provides, you have to put effort into building the stamina and conceptual understanding necessary for complex writing to become transparent to you.

13. I also do some reading where I actively try to push my limits: the type where I need to learn an entire new vocabulary and work hard to absorb new mental models. I find that this has gotten a lot easier since LLMs, large language models like ChatGPT, got good. These days if I read something that is outside my comfort zone, I often keep one or two language models open so I can ask questions as I read.³ I take pictures of paragraphs and upload them as context. If I’m confused, I ask the LLMs to explain what a paragraph says. This means I very rarely get stuck anymore. I can nearly always probe my way to an understanding that lets me keep reading if I spend a few minutes discussing it with an LLM.⁴ This has removed a lot of the emotional resistance I have had against reading very hard books, and has therefore expanded my range and made me a stronger reader.

14. The kind of reading I talk about requires plenty of time—I spend 10-30 hours a week reading. Where do you find that time? Well, we can work backwards from the fact that we’re awake some 112 hours a week. So the problem is rarely actual hours (though if you have kids, it sometimes actually is). But one of my blessings is that I never got into the habit of carrying a smartphone around; instead, I put a book in each jacket pocket and read if the kids are playing in the park or if I’m waiting for the ferry. I have a tendency to get stuck reading the internet if I have access. But if we switch off the router at home, as we aim⁵ to do if we are not doing something that requires internet access, I can carve out two hours at night after we’ve put the four-year-old to sleep. The late evenings alone are enough to read a book a week.

15. Why read with this intensity? Why turn it into so much work and make it another project to manage if you’re happy having it just as a hobby? I suspect this is partly a matter of taste. I value reading things for their own sake, too—without a goal in mind, for the pleasure of being present with a book. But when I treat books like something more than this, I get something deeper. By applying yourself slightly more—and if you are a reading person, what I’ve described in this essay is only slightly more work than normal reading—yes, by applying yourself slightly more, you can retain vastly more from those hours you already put in. Reading seriously changes your brain so that the world that comes at you grows more nuanced and interesting and filled with affordances that let you do things you didn’t know were possible. Serious reading compounds.

16. Good books are compressed thoughts. They are like soda cans left in a hot car: when you open them, the content explodes from their container, spraying and spraying—pouring out much more than it seems reasonable a little can like that could possibly contain. In seven hours, I can read a book of thoughts that someone spent two years thinking. There are few ways of spending seven hours that can compete with that.

Acknowledgments

Johanna Karlsson read and commented on drafts of this essay. Esha Rana did the copy edits. I made the mistakes.

And, as always, a big thank you to the subscribers who fund the production of the free essays.

1. I dislike audiobooks and podcasts for this reason: they make it hard to skim, and hard to stop and think hard about the important parts. They also encourage multitasking, which makes it even harder to process. I suspect people mostly like intellectual podcasts because it gives them the illusion of learning without the actual effort required to really learn.

2. Let me give an example of a question. In July, I was in London to record a podcast and took the chance to visit the British Museum. Having walked through the Indian, Chinese, and Egyptian collections, I entered the Duveen Gallery, where they house sculptures from Ancient Greece… it felt like watching aliens landing on earth. I had of course heard ten thousand times that the Greeks were special. But in the Duveen Gallery, I felt it for the first time: the sculptures that the Greeks were doing was on a level so much higher than what the other cultures were doing—I’m not sure we’ve yet surpassed it. The shift in quality and ambition was so remarkable that it felt impossible. So I decided I needed to figure out why the blossoming that happened in Ancient Greece happened.

Other examples of questions that have possessed me are: how does attention work? Are there any commonalities between how exceptional people were raised? What are some habits and ways of thinking that highly agentic people have in common? Why is Polish literature so good?

3. As of September 2025, my default is two run o3 and GPT-5 high in parallel. I also use Claude Opus and Gemini Pro 2.5, depending on the types of questions I have—they all have different strengths and weaknesses which you can infer by running them side by side and developing a taste for which model performs well on which task.

4. I’m not naive enough to think the LLMs always knows what they are saying. When an LLM is asked to summarize a really deep and subtle thinker, they will tend to tame the ideas and make them less interesting and radical—much like a university professor does when they summarize great thinkers. So there’s no way around having to wrestle deeply with the sources yourself, coming to your own conclusions. But having someone to talk to—even if they are ignorant and wrong—can get you unblocked. They might be wrong in a different way than you and so help shake you loose. Or they are wrong in a way that helps you articulate the truth better. And it is much, much easier to get this kind of sparring from an LLM than to hope to find someone else who’s interested in, say, the development of Polish poetry in the interbellum years (międzywojnie).

5. We fail a lot, but we get better and better at it. Once you have the intention of having no internet access, you can analyze each time you fail and figure out how you can update your routines to make yourself less likely to fail in that way in the future. Over time, that starts to converge toward better and better habits.