I hated writing until I learned there’s a science to it (2024)

Skill, Taste, and the “Gap”

• Many relate to liking good writing yet hating their own; strong taste makes early work feel especially bad.
• The idea that you must “be bad for a long time” before quality and ambition align resonates across domains (writing, trades, drywall, electrical work, sports).

How to Get Better at Writing

• Repeated advice: write a lot on a schedule; expect years of mediocrity before things “click.”
• Deliberate practice ideas: imitate different styles, impose constraints, copy or translate novels, dissect stories, stop movies and continue the plot in writing.
• External feedback and writing communities are seen as essential; learning to both give and receive criticism is its own skill.
• Reading widely, including bad work, helps clarify what works and why. Some suggest lingering over texts and analyzing structure and word choice in detail.

Editing, Structure, and Academic Writing

• Many see editing as the real craft: start with a bad draft, then iteratively cut, tweak, and refine.
• Some love old, essay-like academic papers and dislike modern formulaic structures; others argue the structure supports skimming in an era of information overload.
• Scientific writing is viewed as hard but clarifying; the need to “step aside” as an author to reduce bias can make texts feel empty or joyless.

Talent, Luck, and Practice

• Several endorse “mostly sweat, a little talent, a little luck,” but others argue luck is a far larger factor.
• A counterpoint: “luck” often looks like being ready (skilled) when opportunity appears, after long periods of rejection and grind.

Skepticism About the “Science” of Writing

• Many expected actual scientific explanation and found only “write, then revise” advice.
• Some call the piece content-free clickbait, misusing the word “science”; others still find comfort in the normalization of iteration.

LLMs, AI, and Writing

• Comparisons are drawn between human practice and LLM training (reading plus “exercises”).
• Some claim AI prose has detectable stylistic fingerprints; others worry people will offload writing and lose its thinking benefits.