A communications person opened the terminal and she is okay

Six weeks ago, I'd never written a line of code. My background is in communications (and before that, wedding photography). I'm Eliana, and I'm on the contributor communications team at Outlier. I write Inside Outlier, our weekly newsletter, and most of the emails and platform messages you (hopefully) read. Today, I also build internal software tools that my team uses every day.

I didn't learn to code. I learned to describe what I needed clearly enough that the right thing got built. And I think the reason that works is the same reason I'm good at my actual job: I know how to communicate clearly. It turns out that skill transfers to AI tools just as well as it does to people.

What "vibe coding" means in practice

The term "vibe coding" has been getting a lot of attention lately, but in practice it's simpler than it sounds: you describe what you want a tool to do, it gets built, you react to what you see, and you describe again. The whole cycle takes minutes.

I recently built a content calendar app that tracks everything our communications team publishes across email, community, social media, and SMS. I didn't sketch wireframes or write specifications. I just said "I want a weekly view where I can see every piece of content going out, organized by channel" and kept going from there.

The building blocks of software development haven't changed, but who can use them has.

Clear communication turns out to be the hard skill

Here's the part that surprised me: the thing that makes vibe coding work isn't technical knowledge. It's the ability to describe a problem clearly.

My day job is translating complex information into something a general audience can understand. I take detailed input from engineers and project teams and turn it into clear, useful messaging for millions of contributors. When I'm describing a feature to an AI tool, I'm doing a version of that same thing: figuring out the clearest way to say what I mean, being specific about what matters, and leaving out what doesn't.

I think that's worth paying attention to. The skills that people in non-technical roles use every day, clear communication, understanding your audience, simplifying complexity, may be exactly the skills that matter most in this new era of building with AI.

More tools doesn't mean better tools

Because building is fast, the temptation is to build everything. My first version of the content calendar had too many features, too many views, too many options, and the feedback I got was immediate: "This is over-engineered."

Speed doesn't replace judgment. The most important questions are still the human ones: What problem is this solving? Who is it for? What does "done" look like? If anything, the speed makes those questions more important, because you can build the wrong thing before you've even figured out what the right thing was.

The real barrier isn't skill. It's willingness to start.

The tools can be intimidating. The one I use runs in the terminal on my Mac, that black window with the monospace font that looks like it was designed to keep non-engineers out. I still haven't figured out how to change the font to something less hostile, so if anyone on the Outlier team knows how, please tell me.

But the intimidation wears off faster than you'd expect. You'll see error messages you don't understand and the interface will feel unfamiliar, but that's fine. The willingness to describe what you need, see what happens, and adjust, that's the whole process.

You don't need a computer science degree or to understand what's happening under the hood. You need to be able to think clearly about what you're trying to accomplish and communicate it precisely. That's it.

What this means for the future

The line between "technical" and "non-technical" is blurring faster than most people realize. The ability to build software is no longer limited to people who learned to code, it's expanding to anyone who can think clearly and communicate well.

At Outlier, we see this every day: people with all kinds of expertise contributing to AI in ways that didn't exist a few years ago. Vibe coding is just the latest example of that shift. The tools are getting more accessible, the barrier to entry is dropping, and the people who can ask the right questions are going to be the ones who build what comes next.

Eliana Melmed is on the contributor communications team at Outlier. She builds internal tools, writes newsletters, and is still trying to change the terminal font.