The Essay That Took Paul Graham Seven Months
In July 2023, Paul Graham published "How to Do Great Work." At roughly 11,800 words it's one of the longest things he's ever written, and by his own account it took about seven months. It grew out of a single paragraph in a different essay. The idea felt too big for a paragraph, so he cut it out and let it become its own thing. Shopify's CEO Tobi Lütke later called it "the perfect essay."
Graham frames the whole piece modestly. He says it's the advice he'd give his own kids, or the essay he wishes someone had handed him at the start. That framing matters, because "great work" sounds intimidating, like something reserved for Nobel laureates. The essay's quiet argument is that the path is more ordinary than it looks. It's available to anyone willing to be genuinely curious and to keep showing up.
The catch is that "simple" and "easy" aren't the same word. The recipe fits on an index card. Following it takes years, and most of the failure happens in the gap between knowing the steps and actually doing them. So the useful way to read the essay isn't as inspiration. It's as a set of instructions you can turn into habits, which is exactly what we'll do here.
This article explains the core of Graham's argument faithfully, grounds it in real examples like Darwin's notebooks and Feynman's, and then translates it into a workflow you can start this week. If you already keep highlights and notes, you're closer to step one than you think.
The Recipe: Four Steps to Great Work
Graham's central claim is that great work, across wildly different fields, tends to follow the same shape. He lays it out in four steps: "choose a field, learn enough to get to the frontier, notice gaps, explore promising ones."
The first step is choosing something you have both an aptitude for and a deep interest in. The second is learning enough that you reach the edge of what's known, the place where the textbooks run out. The third is the quiet, easy-to-skip move of noticing the gaps that open up once you're at that edge. The fourth is the work: chasing the promising gaps, even the ones nobody else seems to care about.
Here's the trap. Most people never finish step two. Reaching a real frontier in any field takes years of unglamorous learning, and it's tempting to stop while the material is still comfortable. But the gaps only become visible from the edge. As Graham puts it, frontiers look smooth from a distance and "turn out to be full of gaps" once you arrive.
| Step | What it means | Where most people stall |
|---|---|---|
| 1. Choose a field | Find work you have a natural aptitude for and a deep interest in | Picking what looks prestigious instead of what they're curious about |
| 2. Reach the frontier | Learn until you hit the limits of current knowledge | Quitting while the material is still comfortable |
| 3. Notice the gaps | Spot the questions and anomalies others overlook | Not slowing down enough to see them |
| 4. Explore promising gaps | Do the hard work of chasing outlier ideas | Abandoning ideas that aren't yet fashionable |
The reason this generalizes is that it describes how knowledge itself advances. You can only add to the edge of what's known once you've climbed to that edge. The same structure that produced Darwin's theory of natural selection produces a sharp startup idea or a genuinely new piece of writing. For the startup version of this exact pattern, see how to get startup ideas, which is really step three applied to markets.
Curiosity: The Engine and the Rudder
If the four steps are the map, curiosity is what moves you across it. Graham is blunt about this: "If you asked an oracle the secret to doing great work and the oracle replied with a single word, my bet would be on 'curiosity.'"
He calls it the key to all four steps at once. Curiosity picks your field for you, drags you to the frontier because you can't help wanting to know more, makes you notice the gaps because they nag at you, and then drives you to explore them. It's both the engine, the thing that powers the work, and the rudder, the thing that steers it. "Your curiosity never lies," he writes, "and it knows more than you do about what's worth paying attention to."
The practical test he offers is sharp. "What are you excessively curious about," he asks, "curious to a degree that would bore most other people?" That excess is the signal. Most people sand their curiosity down to fit what seems respectable or marketable. Graham's advice is to do the opposite and treat the weird, intense interests as data about where your great work might be.
This is where a habit beats a resolution. You can't manufacture curiosity by willpower, but you can notice it. Every time you stop to highlight a sentence, save a video, or scribble a question in a margin, you're leaving a trail of evidence about what genuinely pulls you. Tools like Glasp's web highlighter turn that fleeting reaction into a record you can review. Over months, the pattern in what you highlight is a more honest answer to "what should I work on" than any amount of introspection. The science behind why this works is covered in the neuroscience of curiosity.
How to Choose What to Work On
Graham says the work you choose needs three qualities: "it has to be something you have a natural aptitude for, that you have a deep interest in, and that offers scope to do great work." Aptitude without interest fizzles. Interest without aptitude frustrates. And both can be wasted on something too small to matter.
Then comes the line that reframes the whole problem of ambition: "Originality in choosing problems seems to matter even more than originality in solving them." Most people pour their creativity into solving the problem in front of them and accept the problem itself as given. The bigger lever is upstream, in deciding which problem to chase. A merely competent solution to a great question beats a brilliant solution to a boring one.
The hardest problems to spot, Graham notes, aren't the ones labeled unfashionable. They're the ones that just don't seem to matter as much as they actually do. They hide because everyone walks past them. Recognizing one requires knowing a field well enough to feel the wrongness others have learned to ignore.
So how do you find yours in practice? Two moves help:
- Pay attention to what annoys you. Friction, the thing that feels broken or clumsy in a domain you know, is often a gap in disguise. Save those reactions instead of forgetting them.
- Follow questions, not answers. When something makes you think "wait, why is it like that?", capture the question. A running list of your own open questions is a personal map of frontiers worth exploring.
This is also the link between great work and entrepreneurship. The best startup ideas come from noticing a gap you're positioned to see, which is why this essay pairs so naturally with the idea maze and with do things that don't scale. All three describe the same discipline of choosing the right problem and then doing unglamorous work on it.
Reaching the Frontier
Step two, reaching the frontier, is where ambition meets boredom. Graham doesn't pretend it's painless. He's candid that steps two and four "will require hard work," and that while you can't strictly prove hard work is necessary for great things, "the empirical evidence is on the scale of the evidence for mortality." That's his deadpan way of saying: it's about as certain as anything gets.
But there's a release valve, and it loops back to curiosity. The reason to work on something you're deeply interested in isn't sentiment. It's stamina. The frontier is far, and only genuine interest will carry you the distance without burning out. Interest is what turns the grind of learning into something closer to play.
Reaching a frontier doesn't only mean formal study. For most people it means active reading: working through the best sources in a field, arguing with them, connecting them, and pushing until you find the place where the experts disagree or go silent. That edge is where your contribution becomes possible. Passive consumption never gets you there because it never forces you to find the boundary.
This is also why the medium matters less than the engagement. A frontier can be reached through books, papers, or a deep run of expert talks and lectures. When the source is a video, the same rule applies: you have to interrogate it, not just watch it. Turning a dense talk into structured notes with YouTube Summary is one way to compress hours of material into something you can actually metabolize, then highlight the parts that push you toward the edge. For more on extracting real understanding from sources, see how to take smart notes.
Why Notes Are Where Great Work Begins
Here's the part Graham's essay implies but doesn't dwell on: the frontier isn't a place in the world, it's a place in your own head, and notes are how you build and hold it. The people who did the kind of work Graham admires were almost all relentless note-keepers, and not for the reason you'd guess.
Charles Darwin is the clearest case. Starting in July 1837, he filled a series of "Transmutation Notebooks." On page 36 of Notebook B he sketched a branching tree of life and wrote two words above it: "I think." That rough diagram, scribbled in a private notebook years before "On the Origin of Species," is one of the first traces of natural selection. The notebook wasn't a record of a finished idea. It was the place the idea got made.
Richard Feynman put it even more bluntly. When a historian referred to his notebooks as a record of his thinking, Feynman objected: "They aren't a record of my thinking process. They are my thinking process. I actually did the work on the paper." As a graduate student he once opened a fresh notebook titled "Notebook of Things I Don't Know About" and used it to take physics apart and rebuild it, hunting for the raw edges. The notebook was the frontier, written down.
The lesson for the rest of us is that capturing ideas isn't clerical work you do after the thinking. It is the thinking. Your highlights, marginalia, and saved questions are the substrate great work grows from. The difference between a passive reader and someone doing original work is often just that one of them keeps the trail and the other lets it evaporate.
| Approach | What you keep | What it gives you |
|---|---|---|
| Passive reading | Nothing durable | A fading sense you "read something good" |
| Highlighting only | Isolated quotes | Raw material, but disconnected |
| Highlighting + notes + questions | A living map of a field | The gaps and connections great work comes from |
Modern tools make Darwin's habit nearly effortless. Highlighting as you read across the web, then asking questions of your own archive with Glasp's AI chat, turns scattered reactions into a searchable thinking surface. The format changed; the method Darwin and Feynman used hasn't. For a deeper history of this practice, read the digital commonplace book.
Consistency and the Math of Compounding
Graham's most reassuring point is also his most demanding. "Writing a page a day doesn't sound like much," he writes, "but if you do it every day you'll write a book a year. That's the key: consistency." Then he names the mechanism: "If you do work that compounds, you'll get exponential growth."
Compounding is the quiet superpower behind great work. Knowledge built on knowledge, where each thing you learn makes the next thing faster to learn, produces a curve that looks flat for a long time and then bends sharply upward. The people who seem to leap ahead late in their careers usually didn't leap. They compounded, and the curve finally caught up to the effort.
Two consequences follow, and they're easy to get wrong:
- Daily beats heroic. A modest amount of real work every day outperforms occasional bursts, because compounding rewards the number of compounding periods, not the size of any single one.
- Don't break the chain. Each day of work is a small deposit. Gaps don't just pause the growth; they let the curve cool off, which is why momentum is so much harder to rebuild than to keep.
Notes compound especially well, because old notes feed new thinking. A highlight you saved two years ago can collide with something you read today and spark an idea neither could produce alone. That collision is only possible if the old material is still searchable and present, which is the whole case for keeping a durable archive instead of a stack of forgotten documents. The mindset behind this is covered in intellectual compound interest.
Staying Upwind and Working in Public
Few people can name their life's work at twenty, and Graham doesn't ask you to. His answer to uncertainty is a navigation rule he calls "staying upwind": at each stage, "do whatever seems most interesting and gives you the best options for the future." You don't need the destination. You need a good rule for the next step, and curiosity plus optionality is that rule.
Staying upwind means choosing the path that keeps the most doors open and teaches you the most, even when you can't yet see where it leads. Over time, those interest-driven steps accumulate into a direction that would have been impossible to plan in advance. It's the opposite of a rigid five-year plan, and in practice it works better, because it lets you update on what you learn instead of locking you into a guess you made before you knew anything.
The second amplifier is working in public. Graham notes that telling people what you're working on, and sharing what you find, tends to attract the collaborators, ideas, and opportunities that accelerate the work. Visibility isn't vanity; it's how the right people find you. When your notes and discoveries are out in the open, you become findable to exactly the people chasing the same gaps.
This is where the social layer of learning pays off. Sharing your highlights and summaries publicly, the way Glasp's community feed is built around, means your curiosity does double duty: it guides your own work and signals to others what you're exploring. You can even turn an archive of highlights into a finished piece with Hatch, shipping the public artifact that draws people in. For the case for doing this openly, see learning in public.
How to Apply It: A 30-Day Plan
The essay is abstract by design. Here's a concrete way to start living it, using habits you can run with any highlighting and note system.
Week 1: Map your curiosity. Highlight as you read, but add one rule: every time something genuinely surprises or annoys you, save it and add a one-line note saying why. Don't filter for what's "useful." At week's end, review the trail and look for clusters. Those clusters are candidate fields.
Week 2: Pick a frontier and start climbing. Choose the cluster with the most pull and find the three or four best sources in it: books, papers, or expert talks. Read and watch actively, summarizing each in your own words and highlighting the points where experts disagree. Disagreement marks the edge.
Week 3: Collect gaps. Keep a dedicated running list titled, in Feynman's spirit, "Things I don't understand yet." Every open question, contradiction, or "why is it like this?" goes here. Aim for quantity. You're prospecting, not yet committing.
Week 4: Commit to one gap and ship something small. Pick the gap that's both interesting and within reach, and produce one tiny public artifact: a written note, a thread, a short explainer. It doesn't need to be great. It needs to exist, because shipping turns passive interest into the compounding daily habit Graham describes.
| If you tend to... | The fix from the essay |
|---|---|
| Chase prestige over interest | Pick the field you're "excessively curious" about |
| Stop learning while it's comfortable | Push to the frontier where experts disagree |
| Forget what you read | Keep notes that are your thinking, not a record of it |
| Wait for the perfect project | Stay upwind: take the most interesting next step |
| Work in private | Share as you go so the right people find you |
Repeat the loop. Each month sharpens your sense of which gaps are worth your one and only career. Your highlights, notes, and questions are the same infrastructure Darwin and Feynman ran by hand. You just get to run it with search and AI on top.
Frequently Asked Questions
What is the main idea of Paul Graham's "How to Do Great Work"?
That great work follows a learnable recipe driven by curiosity. You choose a field you have an aptitude for and deep interest in, learn enough to reach its frontier, notice the gaps that only become visible there, and then do the hard work of exploring the promising ones. Curiosity is what powers and steers all four steps.
What are the four steps to doing great work?
Choose a field, reach its frontier by learning enough to hit the limits of current knowledge, notice the gaps that open up at that edge, and explore the most promising ones. Graham stresses that most people stall at step two because reaching a real frontier takes years of unglamorous learning, and the gaps are invisible until you get there.
Why does Paul Graham say curiosity is so important?
Because it solves the hardest part: knowing what to work on. Graham argues your curiosity "knows more than you do about what's worth paying attention to," so it reliably points toward problems you'll have the stamina to pursue. His test is to ask what you're curious about "to a degree that would bore most other people."
How do notes connect to doing great work?
Notes are where the thinking happens, not just where it's stored. Darwin developed natural selection in his private notebooks, and Feynman insisted his notebooks "are my thinking process," not a record of it. Keeping highlights, questions, and connections gives you a durable, searchable frontier to build on, which is why a tool like Glasp's web highlighter is a thinking aid, not just a memory aid.
Can anyone do great work, or is it only for geniuses?
Graham's framing is deliberately democratic. The recipe is available to anyone willing to follow real curiosity to a frontier and keep working consistently. Raw talent helps, but he argues that interest, persistence, and good problem selection matter more than most people assume, and those are largely habits rather than gifts.
How long is the essay and when was it published?
It was published in July 2023 and runs about 11,800 words, making it one of Graham's longest essays. He has said it took roughly seven months to write and grew out of a single paragraph in another piece that he decided deserved its own essay.
Conclusion: Curiosity, Compounded
Strip the essay to its bones and you get something almost embarrassingly simple: be curious, follow that curiosity to the edge of what's known, notice what others miss, and keep working. The difficulty was never in understanding the advice. It's in building a life that actually runs on it, day after day, while the compounding does its slow then sudden work.
The good news is that the core habit is one you can start today. Capture what genuinely pulls you, hold onto your open questions, push your reading to the frontier, and share what you find. That's the same loop Darwin ran with paper notebooks and Feynman ran with a pencil, now available to anyone with a highlighter and somewhere to think.
Start building your own frontier. Highlight as you read, summarize the talks worth understanding with YouTube Summary, ask questions of your archive with Glasp's AI chat, and share your trail with a community of people chasing their own gaps. Curiosity is the engine. The notes are the record of the journey. Great work is what's left when you keep both running long enough.