But why, when you can write performance-critical parts in low-level C# (with structs, stackalloc etc) and game logic in high-level object-oriented C#, and have

But why, when you can write performance-critical parts in low-level C# (with structs, stackalloc etc) and game logic in high-level object-oriented C#, and have seamless interop between the two with no effort at all?