Fellow tech folks often seem genuinely insulted when they learn that content is partially or fully AI-generated—from notes and/or a prompt—or that it was generated by AI and later edited by a human.
That reaction made me wonder:
Why is using an LLM to improve my writing treated so differently from using an LLM to improve my code?
In both cases, I have information I want to convey. And in both cases, I’m writing for you, the reader. I want to respect your time. You might be a very important person carrying a lot of mental load already. Respecting you means presenting something that’s clear, concise, and engaging—even if English isn’t your first language.
AI-enhanced content ≠ slop.
To those who get rather thorny at the idea of polished writing coming from someone who isn’t a professional writer by trade: look up almost any piece of great writing and check who’s listed as the editor. More often than not, it’s not the author.
Nobody seems to care about content being edited by editors—until that editor is a rock we tricked into thinking.
Using an LLM does not remove authorship any more than spellcheck, editors, or peer review do. Authorship is about who decides what is true, what matters, and what is said.
I run my blog posts through an LLM to critique flow, structure, and tone. Then I revise the post, feed it back in a few times and focus on specific paragraphs that I feel could use further improvement.
This is exactly how I use LLMs with code: to make it clearer and more maintainable before a human ever has to review it. And I do that because I respect the reviewer’s time—especially when that reviewer is a perfectionist world-class engineer with a packed schedule and a sassy attitude.
Over the past year, I’ve used Gemini to help plan and draft letters opposing the development of hyperscale data centers in America’s deserts, including the secret AWS-backed project, Project Blue. Doing all of this manually would have been effectively impossible, given the time constraints of being a full-time software engineer and partner.
In a world full of low-effort content, it’s our duty to put the best possible work on the internet. The real divide isn’t human vs. AI—it’s care vs. carelessness. The same carelessness that existed long before LLMs, Reddit, Stack Overflow, Yahoo Answers—now simply amplified.
If you’re bad at writing, use an LLM. If you’re good at writing, use one to get even better. This applies to any language—including code.
Just make sure you know what the f*ck you’re doing.