Paul Graham's Advice on Writing: The Best Writer in Tech

Paul Graham co-founded Y Combinator, which has funded Airbnb, Dropbox, Stripe, and Reddit. Before that, he co-founded Viaweb, sold it to Yahoo for nearly $50 million, and studied painting in Florence. Between 1993 and 2020, he published 188 essays on paulgraham.com, roughly 500,000 words. The site gets about 25 million page views a year. He is, by any reasonable measure, the most influential essayist in technology.

His writing advice comes from three essays that are themselves some of the best writing about writing on the internet: "How to Write Usefully," "Write Like You Talk," and "Write Simply." This post draws from all three, plus his conversation with David Perell.

The Formula

Useful writing has four components. They multiply together, so if any is zero, the product is zero.

Importance plus novelty plus correctness plus strength.

"Useful writing tells people something true and important that they didn't already know, and tells them as unequivocally as possible."

This is Graham's filter. Before you worry about sentence-level craft, ask: Is this true? Is it important? Is it new to the reader? Am I saying it as strongly as I can without becoming false? Most writing fails on novelty. It tells people what they already believe. The reader nods along, learns nothing, and forgets the piece by tomorrow.

Bold, but True

Graham wants strong claims. Not hedged, not vague, not buried in qualifications. Strong. The constraint is that they must also be correct.

"Useful writing makes claims that are as strong as they can be made without becoming false." This is a tightrope. Fall one way and you are wrong. Fall the other and you are boring. The skill is finding the edge where a claim is as bold as truth allows.

His analogy: "Pike's Peak is near the middle of Colorado" is better than both "somewhere in Colorado" (too vague) and "the exact middle" (false). Precision and correctness are opposing forces. The best writing finds the point where they balance.

Qualifications are not weakness. "Perhaps," "I think," and "in many cases" are tools. They express how broadly an idea applies, how confident you are, and where the exceptions live. Graham warns: "Don't underestimate qualification." The goal is not to eliminate hedges but to use them with precision.

Write Like You Talk

Most people shift into a different register when they write. Sentences get longer. Words get fancier. The voice stiffens. Graham calls this out: if you would never say "the mercurial Spaniard" to a friend, do not write it.

"Informal language is the athletic clothing of ideas."

His test is simple. After writing a sentence, ask: "Would I say this to a friend?" If not, replace it with what you would actually say. The harder the subject, the more this matters, because "the less you can afford to let language get in the way."

Three techniques for getting there. First, the filter method: review every sentence and replace anything that does not sound like conversation. Second, the read-aloud method: read it out loud and fix everything that does not sound like talking. Third, the explanation method: for severely formal drafts, explain the piece to a friend, then replace the draft with what you said.

Graham claims this puts you "ahead of 95% of writers." That sounds like an overstatement. It is probably correct.

Simple Language Exposes Bad Thinking

Writing simply is not dumbing down. It is a diagnostic. Complex sentences can hide muddled thinking. Simple sentences cannot.

When you write "the dynamic between the two characters is reminiscent of the broader societal tensions" you might mean something. You might also mean nothing. Strip that to simple language and you find out which. If the idea survives the translation, it was real. If it dissolves, it was smoke.

This is why Graham writes short sentences with common words. Not because he cannot use longer ones, but because short sentences force clarity. Every sentence must carry its own weight. There is nowhere to hide.

Write for Fifteen Years Before Publishing

Graham wrote what amounted to essays in notebooks for about fifteen years before publishing anything. He started with a narrow scope: things he genuinely knew about. He did not need a large audience. "If you only have ten readers who care, that's fine."

This is the opposite of most internet writing advice, which says to publish early and often. Graham's path was: think for years, write for years, then publish when you have something worth hearing. The fifteen-year apprenticeship is not a requirement, but the principle stands. You earn the right to write usefully by thinking deeply about a small number of things.

Reread Fifty Times

Graham rereads his essays roughly fifty times before publishing. Not fifty edits. Fifty reads. Each pass surfaces something the previous one missed.

"One of the most dangerous temptations in writing is to keep something that isn't right just because it contains a few bits or costs you a lot of effort." Sunk cost applies to sentences. The paragraph you spent an hour on might be the one killing the piece. Delete it anyway.

He reads passages repeatedly until nothing "catches his sleeve." "Mistakes seem to lose courage in the face of an enemy with unlimited resources." The enemy is the writer's willingness to revise indefinitely.

Surprise Means Contradiction

Graham defines surprise precisely. "A surprise is something that not only didn't you know, but that contradicts something that you thought you knew." Surprise without generality is gossip. Generality without surprise is a textbook. The essay lives where both meet.

He says curiosity about an apparently minor question is an exciting sign. The minor questions often lead to the biggest discoveries because nobody else bothered to investigate.

"If you don't learn anything from writing an essay, don't publish it." The essay should surprise you. If you sat down knowing what you would write and ended up with exactly that, the piece has no discovery. The best essays change the writer's mind during the writing. "Once you start exploring a new idea, your ideas are going to be largely mistaken." That is not a problem. That is the process.

Essays as Thinking

Graham wrote essays involuntarily. "I'm walking down the street and the essay starts writing itself in my head." The essays were not marketing for Y Combinator. They were how he thought. The startup advice he later formalized into YC's curriculum was first worked out in essay form.

This created a flywheel. He wrote about startups. The essays attracted people who wanted to start startups. Y Combinator funded them. Working with them taught him new things. He wrote new essays about what he learned. The writing was not separate from the work. It was the work.

For writers who want to build something beyond the writing itself, Graham is the clearest example. His essays did not describe his influence. They created it.

The Consequence of Useful Writing

Graham warns that this formula for useful essays "is also a recipe for making people mad." Novel ideas contradict cherished beliefs. Strong statements feel offensive. Simple language feels rude to those who disagree. "If there's anything that annoys people more than having their cherished assumptions contradicted, it's having them flatly contradicted."

He does not consider this a reason to soften the writing. The alternative is writing that is vague, familiar, and weak. Nobody gets mad at it. Nobody learns anything from it either.

Key Takeaways

• Useful writing is important, novel, correct, and strong. If any component is zero, the essay is useless.
• Make claims as strong as they can be without becoming false. Find the edge where boldness meets truth.
• Write like you talk. If you would not say it to a friend, do not write it.
• Simple language exposes bad thinking. Complexity hides it.
• Delete without hesitation. Only publish what is worth hearing.
• If you did not learn something writing the essay, do not publish it.
• Expect useful writing to make people angry. That is a feature, not a bug.

Graham and Sam Altman share the belief that writing is thinking made visible. Altman, who succeeded Graham at Y Combinator, echoes the same principle: clarity on the page reflects clarity in the mind. Both treat the essay as a tool for discovery, not a container for conclusions that already exist.

This post draws from Graham's essays "How to Write Usefully," "Write Like You Talk," and "Write Simply" at paulgraham.com, and his conversation with David Perell. Athens is an AI writing editor that shows every change as a tracked diff, so the simplicity Graham demands stays in your control.