Hello everyone,
I’m on the hunt for the most effective workflow with AI that can maintain the context of my entire codebase, as of March 2024. My experience with Copilot in VS Code and occasional use of GPT-4 in chat has been positive, but I find they fall short in keeping up with the full context of my codebase.
As it stands, it seems we have two solid options:
-
GitHub Codespaces: I’m intrigued by the possibility of using Codespaces with Copilot integration. Has anyone used this combination extensively? I’m particularly interested in whether it effectively maintains context across all project files consistently. The idea of a live, streamlined production-only environment (free from the constraints of local VS Code configurations and dependencies) like Codespaces is appealing to me.
-
Cursor : This tool has also caught my eye, especially after seeing some positive reviews on Reddit. However, I’ve come across a few discussions about codebase indexing issues on the Cursor forums. I’m torn between user testimonials and forum reports and would appreciate insights from those with firsthand experience.
Devin doesn’t seem to be fully available and fined-tuned yet so I didn’t include it as a serious and faultless option for now. Nor did I include the other competing products as they seem to fall short.
So my question is: has anyone here had the opportunity to use both Codespaces with Copilot and Cursor, ideally benchmarking them both on the same project?
I’m looking for feedback on your experiences, particularly regarding the ability of these tools to maintain an understanding of your entire codebase.
Your input will be invaluable to me as I consider whether to invest in Cursor Pro.
Thank you for your insights!