The most important step is setting up a clear project structure. Don't even think about writing code yet.
When working with AI tools, I've found that jumping straight into coding often leads to confusion and rework. The AI needs clear context about where things should go and how they fit together. By spending time on structure first, both you and the AI have a shared understanding of the project's organization.
- Break down your project into logical components before writing any code
- Use descriptive folder names that reflect their purpose
- Include a place for documentation from the start
- Consider future scale - will this structure work when the project grows?
- Trying to figure out structure as you go along
- Putting everything in one directory because "it's a small project"
- Not having a designated place for tests and documentation