Why writing is one of the best skills to have as a developer

A few months ago, at the Edpuzzle Engineering team we began to use AI coding tools in our daily work. We haven’t been in the early adopter group, so we faced a steep learning curve in introducing these tools into our day-to-day and codebase. Everyone on the team is still finding what works and what doesn’t, but one thing has stood out to me over the past months: engineers accustomed to writing are having better results when they incorporate it.

Why writing?

In Software Engineering, we know good documentation helps us do our jobs. I’ve heard “we need more documentation” or “We don’t spend enough time on documentation” many times, coming from people I manage or mentor. When I dig into what the person saying the phrase truly means, it’s not that they want documentation by itself; they need more clarity. Clarity in how things work, why they work that way, and the decisions behind the code they have to work on. We generally crave the clarity writing provides because it helps us work better.

The part of writing for clarity that is often overlooked is that it’s a two-way street; it gives clarity to the reader but also to the writer. When you write something down, the piece of writing puts a mirror to your ideas. This act of externalizing your thoughts forces you to confront gaps and inconsistencies that your mind might otherwise gloss over by its internal biases. The final text does not carry the same blind spots our minds do. As you attempt to articulate concepts, you quickly discover where your understanding is solid and where it’s shaky, clear as day.

This is why writing can be such a valuable learning tool. It compels you to organize your thoughts and logic, often revealing the true depth—or superficiality—of your understanding. Asking yourself, “How can I explain this clearly?” or “How can I structure this to make it simpler to understand?” leads you to deepen your own comprehension.

But how is this related to AI tools?

As I said at the beginning, the engineers who are getting significant gains from using AI tools are the ones who have always practiced writing to understand what to do. And once you think about it, it makes a lot of sense. Before coding everything by hand, many programmers jumped straight to coding, figuring things out as they went. This figure it out as you do the thing with AI is no longer a possibility if you want to get incredible results. Today, if you can’t clearly express your intent and logic, the AI won’t understand you.

Another way of seeing it is that the lone-wolf developer who didn’t collaborate well before is a severe disadvantage with AI tools. AI tools require us to think about basic things of collaboration, like how you should communicate things, think of what the other party knows and doesn’t know, structure that communication to make it easy to understand, and many more. It forces us to, as strange as it sounds talking about AI, to empathize with the one receiving the information.

As our roles and industry change, we must write more to improve clarity and collaboration. Focus on what matters: write to understand, not just to code.