Vibe Coding

The latest buzz phrase from the software development sector is 'Vibe Coding', but what it is? Is it any good and should I be doing it?

What is vibe coding?

The term was coined by computer scientist Andrej Karpathy early this year (2025). Vibe coding has the potential to revolutionise how software development is approached in the future with it's use of artificial intelligence (AI) to transform and write code.

Vibe coding, or vibecoding, is an AI-dependent programming technique where a person describes a problem in natural language to a large language model or LLM - think ChatGPT for example. Instead of manually writing code, the programmer provides a prompt, and the AI generates the software. This method shifts the programmer's role from hands-on coding to guiding, testing and refining the AI generated code.

How does vibe coding work?

Imagine you want to create a website that shows local coffee shop ratings based on user reviews. Traditionally, you;d need to learn HTML, CSS, JavaScript and various other technologies. With vibe coding, you simply describe you ide to the AI, such as: "Create high-end, functional website where users can rate coffee shops and see averages for their city." The AI then generates the code, sets up the structure, and may even suggest a database schema.

The benefits of this approach are clear:

1. Accessibility: vibe coding lowers the barrier to entry for software development, making it accessible to those without extensive technical expertise.

2. Efficiency: IT speeds up the development process by allowing programmers to focus on strategic direction rather than the minutiae of coding.

3. Creativity: By freeing developers from syntax and debugging, vibe coding encourages more creative and innovative solutions.

What are the drawbacks?

It's widely known that AI generated code is frequently buggy, and the improved accessibility will mean more people with zero coding knowledge will be attempting to create applications without understanding how to identify those bugs, or even to be able to attempt refining the code or the identify potential security issues.

Is there a way to navigate these drawbacks?

Yes, the obvious way is to equip yourself with at least some knowledge of coding, but this is unlikely to be a quick solutions.

AI will only improve as more people use it, reducing the potential error rate over time.

At Cloudplex Solutions, we leverage the awesome power of Xano for all our backend development and in a recent update, Xano now comes with AI assistance to support the creation of your database and APIs, but there's an important difference. Rather than AI taking your prompts and developing a solution in a black box environment where you are unable to scrutinise the code or rationale, Xano alongside it's existing visual development tools keeps you in control, you're able to test and revert and work alongside AI suggestions rather than relinquishing control to it.

So what's the verdict?

Vibe coding isn't just a trend, it marks a significant shift in the software development landscape and will democratise people's ability to develop applications. It will only improve as AI improves. However, the risk is that control is handed over from the individual to AI which for first iterations of any application will probably be or small concern, but as apps develop, grow and require scaling and more complex infrastructure the average person will likely not know how to accurate articulate what they are seeking for AI to generate a satisfactory response. The solution is to use AI as a tool, to interrogate it's suggestions, to improve understanding as you continue to iterate.