Great that you gain motivation from interest in a feature . I think the highlight reconstruction is a nice thing but it’s unlikely that I shoot double size raws for the potential possibility to get better highlight under very special circumstances. However, even a very rough depth map would be a feature I see real use cases. There I know before that I require it and would switch on dual px raw. E.g. for greenbox purposes.
Next time I use the camera I will try some dual raw shots, but that won’t be soon.
Edit: It seems that other dual px tech is used for depth computation: https://petapixel.com/2017/10/12/samsungs-new-dual-pixel-sensor-help-spread-portrait-mode/. Don’t know if the tech is comparable.