How We're Using Cursor at Neon

Neon's engineering team shares their experience using Cursor AI coding assistant across TypeScript, Go, Rust, and Python codebases. They highlight successful workflows including Cursor Tab for faster coding, Agent Mode for refactoring and debugging, understanding unfamiliar code, and modernizing legacy patterns. However, they identify limitations like context gaps leading to incorrect suggestions, reliability issues in large files, and the cognitive overhead of model selection. The team suggests improvements including better context awareness, version-specific syntax suggestions, automated testing workflows, and open-sourcing the editor.