Vibe coding before the trend
This practitioner report documents early 2025 classroom experiments with "vibe coding" across diverse student cohorts in the Netherlands and South Africa, revealing five key patterns where AI tools shifted learning focus from syntax to higher-order thinking, transformed skills from memorization to evaluation, and fostered a collaborative partnership mindset rather than replacement fears.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine you are teaching a group of people how to build a house. For years, the only way to do this was to force every student to memorize the exact shape of every brick, the chemical formula for the cement, and the precise hand movements needed to lay a single tile. If they made a mistake in the formula, the whole wall collapsed.
Now, imagine a new tool arrives: a magical, super-smart construction assistant. This assistant can instantly mix the cement, cut the bricks, and lay the walls if you simply tell it, "I want a cozy kitchen with a big window." You don't need to know the chemistry or the brick shapes anymore; you just need to know what you want the house to look like and how to describe it.
This is exactly what two teachers at Fontys ICT in the Netherlands discovered when they let students use these new "AI coding" tools in early 2025. They didn't wait for a perfect study or a government policy; they just handed the tools to four different groups of students and watched what happened.
Here is the story of what they found, broken down into simple terms.
The Experiment: Four Groups, One Magic Tool
The teachers gave four different groups of students a challenge: build a digital tool to solve a problem. They used a specific AI tool called Cursor (think of it as a very smart, automated construction crew).
- The Tech Pros (Open Learning): 54 students who already knew a bit about computers. They built complex data tools about Pokémon evolution.
- The Marketers: 24 students who were good at business and ads but were scared of computers. They built games and data charts.
- The Communicators (South Africa): 22 students studying communication in South Africa. They were nervous about the technology but managed to build prototypes.
- The Journalists: A small group of reporters who wanted to build tools to help them tell stories better.
What They Saw: The "Vibe Coding" Surprise
The teachers expected the students to struggle with the AI. Instead, they noticed something they later called "Vibe Coding." This is a fancy way of saying: You can build working software just by describing the "vibe" or the feeling of what you want, without memorizing any technical rules.
Here are the five big things they noticed:
1. The "I'm Falling Behind" Panic (But a Good Kind)
Almost every student felt a sudden rush of urgency. They realized, "If I don't learn how to talk to this AI, I'm going to be left behind in my future job." But here's the twist: they weren't scared the AI would replace them. They were excited that it would change their job. They saw themselves not as brick-layers, but as architects or conductors of an orchestra, directing the AI to do the heavy lifting.
2. Syntax is Gone, Thinking is King
Before, students got stuck on "syntax" (the tiny spelling and grammar rules of computer code). With the AI, the computer handled the spelling. The students could finally focus on the big picture: What problem am I solving? One student said, "I stopped worrying about typos and started thinking about how the whole system works."
3. From Memorizing to Checking
The most important skill shifted. It wasn't about memorizing facts anymore; it was about critical thinking. The students realized their new job was to be the "editor." They had to look at what the AI built and say, "Yes, that's right," or "No, that's wrong, try again." They became judges of quality rather than just producers of code.
4. The "Plumbing" Problem
Here is the funny part: The AI was amazing, but the students still got stuck on basic computer stuff. They didn't know how to save a file, where to put a folder, or how to install a program. The teachers realized the hardest part wasn't talking to the AI; it was navigating the computer itself. It's like having a Ferrari engine but not knowing how to turn the key in the ignition. The students had to learn the "plumbing" before they could enjoy the "vibe."
5. The Door Opens for Everyone
The most surprising result was for the non-tech students (the journalists and marketers). For the first time, they could build a real software product using their own ideas. A marketing student could build an app without needing to be a programmer. It felt like a locked door had suddenly opened, letting them into a room they thought was only for "tech people."
The Big Picture: Then vs. Now
The teachers wrote this report in March 2026, looking back at their experiments from early 2025. They noted that the world moved incredibly fast.
- Then (2025): The tools were a bit clunky. Students had to fight with installation errors and confusing menus.
- Now (2026): The tools are smoother. Students can now finish a whole semester of coding classes without writing a single line of code by hand. The focus has shifted entirely to: "Can you take ownership of what the AI made? Can you explain why it's good?"
The Lesson for Teachers
The teachers have three main pieces of advice for other educators who are nervous about AI:
- Just Do It: Don't wait for the perfect plan or official permission. Start messy. You will learn more by trying and failing than by waiting.
- Watch the Interface: Don't just throw a complex tool at beginners. If the tool looks scary (like a cockpit with too many buttons), students will get stuck. Use simpler tools first.
- Let Them Choose the Problem: Don't force everyone to build a Pokémon game. Let the journalists build a news tool and the marketers build an ad tool. When students work on things they care about, they do their best work.
The Bottom Line
This paper isn't a scientific study with perfect numbers; it's a collection of stories from the "trenches." The main takeaway is that AI is changing the game. It's not about replacing humans; it's about lowering the barrier so that anyone with a good idea can build a digital product. The students aren't afraid of the future; they are realizing that the future is something they can finally build themselves.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.