Lately, studying prompt engineering keeps reminding me that it’s not just about using AI well—it might also be a way to train our underlying communication skills at work.
This piece is about the experience that led me to that thought.
This blog’s layout was built with ChatGPT and Codex (OpenAI’s coding agent). I talk with ChatGPT about direction and improvements, then it turns those notes into prompts for Codex to act on.
One day, while looking at the mobile version of Silica Plant, I noticed the Book item in the top nav was slightly clipped. It seemed like a minor style issue, but after grabbing a screenshot and reviewing it with ChatGPT, the cause was clear: “Technology” was a bit too long on mobile screens.
The fix followed naturally: on mobile, “Technology” should become “Tech.”
I asked ChatGPT to draft a prompt for Codex based on that conclusion. It produced this:
Fix the mobile navigation label overflow.
Requirements:
- Prevent any navigation item from being visually clipped or misaligned
- On mobile, shorten “Technology” to “Tech” to maintain a single-line nav
- On desktop, keep the full label “Technology”
- Do not solve this by shrinking font size or wrapping
- Maintain balanced visual rhythm across nav items
Deliverables:
- Minimal HTML/CSS changes
- Mobile-first approach with a clear desktop override
What struck me was how much clearer this prompt was than my initial vague idea—especially the line, “On desktop, keep the full label ‘Technology’,” which I had assumed but never stated.
If I’d handed the task to a person, they might have asked:
“What about desktop?”
That question wouldn’t be odd; it would just show the original request was a bit under-specified.
This made me think: practicing good prompts for AI isn’t so different from learning to brief a teammate well.
Digging into prompt engineering, I read the article below:
26 Prompt Engineering Principles for 2024It summarizes good prompt principles as:
- Clear and concise
- Contextual
- Aligned with the intended task
- Supported by examples
- Free of unnecessary bias
Sound familiar? These are the same things we need when explaining a task to a colleague or asking for help.
Maybe studying prompt engineering is really “practicing how to organize what’s in my head so others can understand it without confusion.”
If AI practice also sharpens how we communicate with people, that alone makes it worth learning—like pro Go players getting stronger by studying AI games.