You guys are really missing the point here. This is not user error from my side. Cloud computing has minimal impact on a local computer.
This is about Apple having silly ass background processes like "mediaanalysisd" that was eating at some point 8GB RAM.
At least for once you gave an example. Here's the deal, as you do not seem to understand how macOS actually works. This is fine of course, in the end we are users, but when starting to complain it's sure worth it to actually dig a little deeper. As Apple publicly on stage explained several years (could be as far back as macOS Lion?!) RAM in a computer basically consumes the same amount of energy no matter whether used or not. Thats why apple started to use caching in many parts of the OS, to try to preload files which it assumes are gonna be used next into RAM, so some user tasks end up being executed faster. Also these processes, while "clobbering" up RAM, immediately unallocated RAM when a user tasks needs it. Regarding theses mechanics high RAM usage is actually a feature.
Reading these media scanning processes: These take place for several reasons, and are local rather than cloud based like google and others. Also the database constructed by it's findings remains local as well, and even is excluded from any kind of macOS/iOS native backup solutions, thus needs to be recreated when setting up a new machine but also when coping files from external volumes, downloading content, new safes and so on. The features this enables include people recognition of your contacts (photoanalysisd) and contents analysis so media can better be found from spotlight. It enables features like being able to find documents by content (and fast!), rather than by document name among other things.
Also you cannot compare memory usage among different OS easily. For one all OS are setup differently and enable different use cases. What's more easily comparable are specific scenarios, where the same applications exists for different systems, but even than those applications might (and should, if they are optimized!) rely on different system APIs. So basically to evaluate an OS ask yourself, if it enables you to do what you need it to do, and wether it does so in a timely manner prioritizing metrics important to you like ease of use, energy consumption, time, reliability, and joy of use. It has never lead to anything comparing devices by GHz, RAM, MP, mAH or whatever hardware metric. It's about what the resources you purchased enable you to do.