alt Hacker NewsOr you can write code which can directly run on x86, i.e. FreeRTOS does support that without issues. For peripherals drivers you will need to burn it on chip regardless because emulator rarely can emulate peripherals in some reasonable way.
So if you correctly abstract business logic from peripheral code, you can do most of your development without ever uploading to target.
That's a solid approach, and for high-level logic, it's definitely the way to go.
I find that a lot of my development time is actually spent on lower-level tasks—like writing custom string operations—since we don't have the rich standard libraries of a host environment.
This is exactly where an emulator really shines for me. It enables a "device-less" workflow where I can work through those low-level details on a sofa at a cafe without needing to bring the physical hardware along just to verify the behavior.