Why Cursor
AI-native development environment
The Problem
Traditional IDEs were designed for humans typing every character. AI-assisted development is fundamentally different—you're directing, reviewing, and refining, not writing from scratch.
Bolting AI onto a traditional IDE creates friction. Copilot in VS Code is a suggestion engine, not a collaborator. You're still doing most of the work; AI just autocompletes.
The paradigm shift is from writing to directing.
When AI can generate entire functions, files, or features, the bottleneck shifts. It's no longer "how fast can I type?" but "how quickly can I express intent and verify output?"
We needed an IDE that:
- Treats AI as a first-class collaborator, not an add-on
- Enables natural language interactions with the codebase
- Understands project context, not just the current file
- Makes verification and refinement as fast as generation
Current Options
| Option | Pros | Cons |
|---|---|---|
| VS Code + CopilotIndustry standard editor with AI autocomplete. |
|
|
| CursorVS Code fork rebuilt around AI. Context-aware, conversational. |
|
|
| JetBrains + AI AssistantProfessional IDEs with integrated AI capabilities. |
|
|
Future Outlook
The IDE of the future is a conversation interface backed by code intelligence.
AI will handle more, humans will direct more.
Current AI assistance is primarily generative—it writes code you describe. The next phase is analytical—AI that understands your codebase well enough to answer questions, identify bugs, and suggest architecture improvements.
Cursor is positioned for this future. Its codebase indexing and context management are infrastructure for increasingly capable AI. As models improve, the IDE that best integrates them wins.
The risk is lock-in. Cursor's advantage is its AI integration, not its editor. If VS Code (or another editor) matches that integration, switching costs drop. But for now, Cursor's head start is significant.
Our Decision
✓Why we chose this
- Codebase-aware AIAI understands your entire project, not just the current file
- Integrated chat and editConversations flow directly into code changes
- Multi-file operationsRefactoring across files in a single AI interaction
- Familiar foundationVS Code base means existing knowledge and extensions transfer
×Trade-offs we accept
- Subscription costPro features require monthly payment
- VS Code fork lagNew VS Code features may arrive late
- Learning curveNew workflows take time to internalize
Motivation
We switched to Cursor because it changed how fast we ship. Features that took days now take hours. Refactoring that required careful manual work now happens in a conversation.
The key insight is that AI-assisted development isn't just faster typing—it's a different activity. You're specifying intent, reviewing output, and iterating on refinements. Cursor's interface is designed for that workflow.
For business-critical systems, this speed matters. Faster iteration means faster bug fixes, faster feature delivery, faster response to customer needs. The cost of Cursor is trivial compared to the productivity gain.
Recommendation
Try Cursor for two weeks. The learning curve is short—it's still VS Code underneath—and the productivity gain becomes obvious quickly.
Key workflows to learn:
- Cmd+K for inline AI editing anywhere
- Cmd+L to chat about the current file/selection
- @codebase to ask questions about your entire project
- Composer for multi-file changes
Keep your existing VS Code settings and extensions—most transfer directly. Cursor maintains compatibility while adding AI-native features on top.
Examples
# Cursor AI Workflow
## Quick Edit (Cmd+K)
Select code, press Cmd+K, describe the change:
"Add error handling for network failures"
## Chat (Cmd+L)
Open chat, ask questions about selection:
"Explain what this regex does"
"Are there edge cases this doesn't handle?"
## Codebase Search (@codebase)
Ask questions about your entire project:
"@codebase where is user authentication implemented?"
"@codebase what patterns do we use for error handling?"
## Composer (multi-file)
For changes spanning multiple files:
"Create a new API endpoint for user preferences
with tests and type definitions"Cursor workflows are conversational. You describe what you want; the AI proposes changes; you accept, reject, or refine.