Rise of the vibecoding journalists

In 2025, advances in AI-powered code generation led to the ability to “vibe code,” i.e. create apps through prompting without even looking at the code. These vibe-coding tools lowered the cost and technical expertise required to build prototypes. After spending a year building media tools with LLMs, I predicted more dynamic and interactive interfaces for telling news stories. Expanding this further, over the next year we’ll see more custom tools and applications actually created by journalists, editors, and other non-technical staff in newsrooms.

Ten years ago I organized a series of hackathons bringing together journalists and developers. In my work going to developer conferences and newsrooms I found developers interested in making impactful things but not sure what to work on, and journalists with ideas for tools they would immediately use, but without the resources to build them. Hacking Journalism brought these two groups together to build. Now a decade later, many of these ideas could be quickly vibe-code prototyped. Instead of hackathons, I’ve shifted to throwing vibe-code jams — casual events where participants, regardless of technical expertise, rapidly build out 2-3 ideas within an hour. I see the most potential in niche areas of expertise, hyper-individualized tools just for you and your workflow.

Vibe coding is not a replacement for the way we build digital products today. The resulting vibe-coded apps can be buggy, difficult to maintain, and have security issues. But consider that not all apps need to be launched at scale, and not all apps require such a high a bar for quality to be useful. These could be tools for an individual or only a handful of users. There’s a parallel here between mass media and zines. Zines opened up a new way to create, publish, and share media. They lowered the bar of what could be printed and shared, welcoming all sorts of ideas and authors. There are ideas that would have never been explored and published outside the format, especially in niche areas. You could make a zine so unique, it’s for one person, and all the more meaningful because of that.

With vibe coding, it’s now cheaper to build interactive digital things, and so many niche ideas that would otherwise be too expensive can now be built. The next year will only see code-generation tools improve. I’m most excited to see what is possible with vibe coding when in the hands of domain experts. In this space of news technology, what will journalists, editors, and newsroom staff make that otherwise wouldn’t have been feasible? In 2026, the most significant newsroom innovation won’t come from dev teams — it’ll come from journalists who can now create their own tools.

There are numerous tools to use for vibe coding. Here are three I use, with some quick notes. I’ve ordered these tools based on skill and outcomes. Generally the tools that require more skill will also have higher flexibility with what you can make.

• Gizmo works right on your phone and is super fast. It’s a fun feed of interactive media, kind of like digital toys. It’s a great way to get started with vibe coding. Here are my projects.
• AI Studio has a Build feature to vibe-code apps. You can make entertaining toy-like things, but also useful apps. Google’s text and image generation APIs are built in, so you can build apps leveraging these tools. Here are some quick instructions on getting started.
• Claude Code is incredibly powerful, but also requires the most skill. It uses the command line. If you find yourself using the vibe-code tools and curious about what you could make with a more powerful one, try this out.