What Defines Vibe Coding?
Rather than starting from a blank file, developers use prompts like “Build an Illustrator script for placing a file and swatching the colors,” and the AI outputs the relevant code. The emphasis shifts from technical syntax to clearly describing software behavior. The goal is to communicate intent, what you want the software to do, rather than how to write it. Routine or repetitive tasks are handled by AI. Developers spend more time on architecture, and logic. Because AI bridges the technical gap, people with limited coding experience can now contribute directly to building functional applications, and tools for efficiency. You can prototype, revise, and test ideas way faster than with traditional methods.
Tools That Enable Vibe Coding
- Replit – A cloud-based development environment with built-in AI support for code generation. We also looked at Replit in and episode of °FAI.
- GitHub Copilot – A widely used AI assistant that offers contextual code completions and suggestions. GitHub is great, although in my experience, it does usually require some developer experience, even with AI help.
- Windsurf – A Visual Studio Code fork optimized for AI-assisted workflows. I have not tried this but I do bounce back and forth with Visual Studio Code and Chat GPT, this implies it is all in one place.
- ChatGPT– I have really come around to ChatGPT because it handles so many things, I can save on subscriptions in one space. Art, ‘vibe’ coding, etc. Like I mentioned above, I use this with ChatGPT.
Real-World Example
Creating Illustrator Scripting for Separations – my whole reason for writing this week, many people ask me for AI seps. This is how you do it. There are a series of AI seps on Ink Kitchen for free, the sep script, the instructions, all of this was created using this method.
OR building an AI pricing calculator
Benefits of Vibe Coding
- Speed – Faster from concept to working prototype.
- Accessibility – Lowers the barrier to entry for people without coding experience.
- Focus – Allows experienced developers to focus on higher-level logic and design.
- Cost Efficiency – Reduces time spent on repetitive tasks, potentially lowering development costs.
Limitations of Vibe Coding
- Quality Control – AI-generated code can have bugs or inconsistencies and inexperienced individuals may not know how to correct them. Conversely, I have noticed the more experienced developers get hung up in the AI mistakes, which slows the process overall. If you decide to try this you have to let the AI build the code. If you try to build the code for the AI or correct the AI, you will get caught in a loop, it took me awhile to learn that. You can clean up the code at the end if you want, but if its working, why worry about it?
- Security – AI Generated code must still be reviewed for vulnerabilities, and it can be more vulnerable due to lack of knowledge. (I avoid these spaces mostly)
- Getting Stuck – you can fall into a rabbit hole quickly with this method that you can struggle to escape, meaning you make an error OR it’s not following your natural language. I have had to learn to save iterations after each step, so I can easily step back (common for any development) but if I cannot progress I may have to start over feeding the AI the last successful step, to break its current progression.
- Transparency – The code may work, but understanding why it works may be lost.
In Summary
Vibe coding is a practical evolution to building software. It emphasizes speed, intent, and collaboration with AI. While it won’t replace developers, it will change how they work—and who gets to participate in building software. I highly recommend trying one of these platforms, building something simple and moving into something more complex, that you’ve never tried to build before.
Comments