How I Write with Large Language Models (LLM Draft)
Back to the original:
OpenAI's ChatGPT 3.5 transformed my writing process when it came out. After years of experience using it, I've further refined my method of using LLMs. This post explains how.
ChatGPT 3.5 came out just over two years ago and sparked a surge in Large Language Model (LLM) development. Dozens of companies released their own models, and the state of the art advanced by the hour.
At the time, I wrote about how I used ChatGPT to write. My method was primitive. With years of experience and improvements in the models, I have refined how I use LLMs to edit text. Here is my updated method.
Drafting
I write the first draft entirely by hand. This helps me preserve my voice and prevents the writing from being overly influenced by the LLM. It also supports my main goal: clarifying my thinking. Whereas I used to write, edit, write, edit, and so on until I was nearly 100% happy with my work, I now stop earlier and let the LLM handle an editing pass. This is because the aspect I like least about my early drafts—how my paragraphs transition from one to the next—is something LLMs excel at fixing. They also catch spelling and grammatical errors that simpler programs might miss.
The Prompt
The prompt you use with the LLM is crucial because it shapes how the model edits your writing. The prompt keeps the machine from replacing my writing with overused phrases like “delve”, “showcasing”, and “underscores”. I currently use a slight variation of this prompt:
Help me edit this blog post I’m writing. Fix errors, make it clearer. Reword to make the arguments and sentences more coherent. Use the same sort of words I’m using, don’t substitute fancy synonyms. Maintain my voice. My work is below. Keep the formatting and wrap your output in ```.
Editing
Once I have the LLM’s version, I put it side-by-side with my draft to compare. Sometimes the edited version is spot-on, and I’ll adopt an entire paragraph as is. Other times, I’ll borrow ideas for restructuring a paragraph or crafting a transition, but I rewrite it in my own words. There are also cases when the LLM tries to fix something that’s already fine, and I simply ignore it.
After that, I go through another human editing pass to ensure my voice remains in every sentence. Occasionally, I will have the LLM focus on specific sentences or paragraphs that still need work. Once I’m satisfied, I commit my changes and publish.