Finishing Side Projects with LLMs and Cursor

This blog post details how the author dramatically improved their side project completion rate using LLMs and the Cursor IDE. Facing time constraints, they leveraged AI to refine project specifications, bootstrap code generation, and iterate effectively. A habit tracker website serves as a case study, walking through the process from initial spec (refined with ChatGPT) to deployment via GitHub Actions. Key strategies include using Vite for project setup, Cursor's agent mode for code generation, iterative development with a divide-and-conquer approach, and providing ample context to the LLMs. The author emphasizes the importance of choosing the right LLM and Cursor mode for different tasks, ultimately delivering a functional v1 of the habit tracker.