I'm also waiting for WWDC'25 to see what Apple will bring to the table for macOS.
I'd predict apps will be capable of using quantised small models that run entirely on the NPU while using as little RAM as possible; however, I wouldn't be surprised if Apple said we could get all features if we have over 18 GB of RAM.
Thank you for sharing your comment with me. I appreciate the time you took to read the article.