Hey everyone! 👋 I'm relatively new to the whole AI assistant scene for coding, and I've been diving in headfirst. I've seen a lot of threads comparing assistants individually, but I'm curious about something a bit different.
I'm currently using **AWS CodeWhisperer** at work (mostly for SQL and Python ETL scripts), and I've been really tempted to try out **Cursor** for its chat-based approach and deeper IDE integration. My question is: **has anyone actually tried running both of them simultaneously in the same editor?**
I'm wondering about a few practical things:
* Is it even technically possible, or do they conflict with each other?
* If you can run both, how do you manage the suggestions? Do you toggle one off, or do you get overlapping/completing suggestions?
* For those who've done it, was there a clear "best for" scenario you found? Like using Cursor for refactoring or data model questions, and CodeWhisperer for quick line completions?
I'd love a detailed walkthrough of your setup if you've made it work! I'm especially interested in how it plays with data tasks—like building a dbt model or a Looker block. Any beginner-friendly tips or gotchas would be super appreciated!
Ooh, this is such a good question. I've been experimenting with something similar! I haven't run both in the *same* editor exactly, but I have used CodeWhisperer in VS Code and Cursor separately on the same project.
For your first point about conflicts, I'd be really worried about them stepping on each other's toes with inline suggestions. You'd probably have to disable the autocomplete for one of them entirely. The bigger headache might be the shortcut keys clashing.
> was there a clear "best for" scenario you found?
This is where I'm at too. I find myself using Cursor for the "what should this function even do?" questions and planning, and then I'll switch over to CodeWhisperer to quickly fill in the patterns it's good at, like the SQL joins you mentioned. It's less of a simultaneous setup and more like task-switching between tools. Have you found the context switching jarring, or is it smoother than I'm imagining?
Just my two cents.
> The bigger headache might be the shortcut keys clashing.
This is the part that makes it practically untenable in a real work environment, beyond just being a technical annoyance. Your editor becomes a battlefield of popups and conflicting keystrokes. It's not just jarring, it actively slows you down.
Your task-switching approach is actually the sane path. Use one for the high-level design chat, then a separate, clean session with the other for the boilerplate generation. Trying to run them in parallel feels like chasing a "magic" workflow that just creates friction.
I'd be curious if anyone has tried this and actually measured a net productivity gain, versus just feeling like they're using more "AI".
Trust but verify