TB: Paul Graham on Writing (Two Essays)

Core Thesis

Good essays emerge from genuine intellectual questions developed through a recursive writing-and-rereading process — and the quality of the resulting prose is both a tool for developing thought and evidence that the thinking was done well. Writing and thinking are the same activity, but only when writing is used to develop ideas rather than merely describe others'.

Key Takeaways

The essay as recursive question-and-answer

  • Essays begin with a question plus an initial insight — neither can be generated systematically.
  • Write a rough response; rigorously reread; follow the threads that show the most novelty and generality.
  • Each answer generates new questions. "In the limit case, the questions are the harder thing to get. That pool has no bottom."
  • The starting question is the hard part and the limiting factor — you can cultivate conditions for good questions but not manufacture them.

Breadth and depth are both required and genuinely distinct

  • Breadth comes from reading, talking, and seeing — consuming the world widely.
  • Depth comes from doing — direct, hands-on engagement with a subject.
  • Neither substitutes for the other.
  • John's annotation: "You can't learn everything, so getting breadth implies learning about topics that are very different from one another. Book topics should be a laundry list of unrelated subjects."
  • Unexpected connections across distant domains are the source of original insight.

Writing that sounds good is most likely to be right

  • John's annotation: "Writing that sounds good is most likely to be right." Being right (logically coherent) is different from being true (externally correspondent).
  • The writer is the first reader: readable prose helps the writer catch errors and gaps in reasoning.
  • Like a shaken container where objects settle into better order — any forced revision tends to improve both sound and substance simultaneously.
  • The aerodynamics analogy: Kelly Johnson ("if it looks right, it flies right") — if prose sounds right, the reasoning is likely right.

Writing develops ideas; describing others' ideas does not

  • The tight coupling between good writing and good thinking holds only when writing is used as a tool to develop ideas.
  • "The writing in textbooks and popular surveys can be bad for the same reason: the author isn't developing the ideas, merely describing other people's."
  • Writing through something (essay, grant proposal) is categorically different from writing about something (lab report, review article).

The timelessness paradox

  • Essays on timeless topics become "obvious" once their insights spread culturally.
  • The best essay at any moment is usually one describing a discovery not yet made — so the best essays always feel inevitable in hindsight.
  • To remain perpetually surprising, essays must address recurring mistakes, persistent institutional lies, or deep experiential knowledge that resists secondhand transmission.

Mental Models

  • First Principles Thinking — essays require first-principles insight; writing through ideas forces analysis; writing about ideas allows borrowed reasoning to slip through unchecked

See also

Sources

Paul Graham, paulgraham.com

Backlinks