I teach AI to editorial and PR teams for a living, and if there’s one thing that excites and engages them more than any other, it’s vibe coding. The highly visual and interactive projects my students create with vibe-coding tools often turn me into the person taking notes.
Vibe coding is definitely having a moment. It’s arguably the most impactful thing to come out of the field of generative AI in the past year, at least as far as applied AI goes. Broadly, vibe coding is the practice of using AI to create not just “content,” but webpages, apps, and experiences—software people can actually do things with. And you don’t need to know a lick of code: The AI will take your plain-language prompts and do all the programming for you, whipping up pages or even entire websites in minutes.
The thrill of the first click
The feeling you get the first time you vibe-code something is similar to what you probably felt the first time you asked ChatGPT to write an essay. You feel incredibly empowered, and maybe even a bit fooled. “It can’t be that easy,” is a common thought. And you’d be right: Vibe-coded experiences are visually and technically impressive, but they are almost always one-offs: Turning them into stable tools you can use on an ongoing basis typically requires a wider set of software and developer skills.
Nonetheless, vibe coding has the potential to be transformative for storytelling, newsrooms, and the media at large. At last, the people crafting content are no longer constrained by the tools forced upon them by their organizations. I remember the publication I worked at in the early days of blogging didn’t even have a gallery tool for readers to quickly scroll through images. Today, even absent AI, there are many platforms and plug-and-play tools to choose from, but they rarely have all the features you want. In any case, incorporating new software is typically a lengthy process in organizations.
With vibe coding, creators can now build experiences that are tailored to the content, not the other way around. Like I mentioned at the outset, this often ignites an enthusiasm in storytellers and domain experts, which is leading to a fantastic uncorking of creativity as more journalists dabble with vibe coding, like an interactive explorer of Newark’s municipal service data or a webpage that turns wildfire point data into Datawrapper-ready hexagon maps.
The challenge for media organizations is to translate that enthusiasm into deeper audience engagement, and to do it in an ongoing way. That, however, requires an approach that goes beyond simply giving reporters and editors permission to experiment. It requires new skills, specialized tools, and above all a cultural shift.
Turning vibe coding into a team sport sport
The skill of vibe coding isn’t that different from “normal” AI skills, which is to say structured prompting and understanding how to collaborate with AI will get you a long way. But to get the most out of vibe coding, it helps to think first about what inputs you need (beyond your story) and find examples of other interactive experiences that are similar. Most importantly, think about what your audience wants and how you expect readers to interact with what you’re creating. Data journalists will probably have an advantage here, but it ultimately comes down to thinking a bit more like a product manager than a writer.
