For most of my working life, I’ve been fascinated by how ideas become real. In the last century, when I taught mass communications and videography, the creative process was equal part excitement and endurance. We storyboarded ambitious scenes, imagined sweeping camera moves, and dreamed in colour. We then we spent hours in the editing suite wrestling with sync, sound, and timing. The gap between what we envisioned and what we produced was often wide, and closing it required patience, skill, and a fair bit of stubbornness. That has since changed.
I followed a similar path with early computing. Tools like WordStar, HyperCard, and QuarkXPress opened new creative possibilities, but they also demanded a working knowledge of code. When the web arrived in the 1990s, I used those skills to build simple pages for students. But as blogging platforms and user‑friendly tools emerged, the need to hand‑craft everything faded. Even when I later managed large coding projects, I left the actual code to the coders.
Today, that landscape around coding has shifted again and dramatically.
The Rise of Vibe Coding
We are now in an era where AI allows us to be more ambitious than ever. “Vibe coding” ,using natural language to shape an application into existence, it represents a genuine paradigm shift. Instead of wrestling with boilerplate or scaffolding, we can focus on the logic, the user experience, and the problem we’re trying to solve. The AI does the magic and handles the heavy lifting.
Whether you’re automating a manual workflow, prototyping a weekend idea, or simply tidying up a webpage, vibe coding lets you move from concept to working prototype with astonishing speed. It’s powerful even if you never intend to write a line of code yourself. You can test ideas, explore possibilities, and hand a validated proof‑of‑concept to professional developers when you’re ready.
I’ve been experimenting with this through the Glasgow No Code Initiative, which has taken me “back to college” in the best possible way. While the classes use Replit, I’ve tried a range of tools, and the experience has been both a refresher and a revelation. It's been good too to catch up with some fellow like minds from Glasgow's digital community who have signed up for the course.
I’ve even built a simple app that checks a community hall’s availability via Google Calendar and allows users to make a booking, something that would once have required a small team, a long timeline and/or would have cost money if we'd bought an off the shelve tool.
The Golden Rules of Prompting
Regardless of the tool, your results depend on the clarity of your instructions. There are four principles that consistently produce better outputs from large language models like Replit , Claude or ChatGPT:
1. Be Clear and Concise
Avoid unnecessary detail. Tell the model exactly what it needs to know.
2. Prioritise Requirements
Put the most important constraints at the top of your prompt. Structure matters.
3. Frame Requests Positively
Tell the AI what to do, not just what to avoid. Positive instructions reduce ambiguity.
4. Use Precise Language
Specificity narrows the “hallucination gap” and leads to more reliable results.
And if the model gets it wrong? Refine your prompt. Iteration is part of the process — just as it always was in the editing suite.
A New Creative Cycle
What excites me most is how familiar this all feels. The tools have changed, but the creative impulse hasn’t. We’re once again at a moment where imagination leads, and technology follows. The difference now is the speed: the distance between idea and prototype has collapsed.
If I was back in the classroom I'd be showing all learners how to use these tools to develop apps for themselves.
For educators, community builders, and anyone with a spark of curiosity, this is an extraordinary opportunity. If you have an idea, even a small one, give it a go. You might be surprised by how quickly it becomes something real.
And if the Glasgow No Code Initiative runs again, it’s well worth signing up.
