I created a custom Linux desktop setup and used an AI assistant while learning how to configure it. Along the way, I learned Lua, Bash, JSON, and CSS, used documentation and tutorials, debugged problems, and generated some Bash boilerplate as a starting point. The overall design, configuration, and final implementation are mine, and I understand how the pieces work. I use the setup every day for university and can modify or troubleshoot it without relying on AI. Can I reasonably consider the project my own, and does this count as "vibe coding"?
3 Answers
Based on your description, this sounds like using AI as a learning and productivity tool, not blindly accepting generated code. If you understand the configuration, can debug it, and can recreate or adapt the important parts using documentation and other resources, calling it your own project is reasonable. The key distinction is whether you are directing and understanding the work rather than merely prompting until something works.
It is fair to say you made the setup, especially since you chose the design, integrated the pieces, and maintain it yourself. However, “100% mine” can mean different things. Some snippets may have been generated or influenced by the AI, so it is more precise to say that it is your project and implementation, created with AI assistance. That is common in modern development.
The important skill is knowing when the generated answer is incorrect or unsuitable. Documentation and testing still matter, because an AI assistant can confidently suggest broken, insecure, or outdated code.
Nothing in your description sounds like vibe coding. That term usually refers to accepting large amounts of generated code without understanding, testing, or being able to maintain it. Using an assistant for explanations, boilerplate, or small pieces while you learn and remain responsible for the final result is a different workflow.
You do not need to memorize every command or write everything without looking anything up. Real developers use documentation, search engines, examples, and other tools. What matters is knowing what you need, evaluating the result, and being able to fix it when it is wrong.