Sort of - ARC's behavior will vary based on the architecture, and in some cases how the emitted code behaves will vary based on other parameters, but the code emitted will be consistent.
ARC is not really about when to call retain and release at the correct time though - it's more about managing object lifetimes and the guarantees related, it actually goes to a lot of trouble to avoid calling retain and release.
You are right though - this post focuses more on the fundamentals of how retain/release is implemented, and made efficient, but since this is only really used in ARC and not in manual reference counting, it's applicable to understanding ARC.
The second half of this will include details around how ARC uses these optimized retain implementations to actually perform the reference counting operations, how it determines things to remove vs keep, etc.
