Nobody asked, but I do have a small contribution to the 8GB RAM on MacBook Pro in the year of our lord 2023 discourse. (Coherence not 100% guaranteed.)
My first Apple silicon machine was a Mac mini M1 with 8GB of unified memory. I was quite pleased with its performance (and probably still would be!) especially compared to the 2015 MBP it replaced. However, I definitely ran into bottlenecks with my “pro” workflows with that little memory available. And when I hit those bottlenecks, it hurt: the system would become unresponsive, with beachballing and even the cursor becoming rather jumpy. It got to the point where I’d have to shut down Docker in order to run OBS, for instance…and I just grew weary of that limitation pretty quickly. (I’d at least feel like I could hit them, though, without feeling like the computer was about to launch into outer space, unlike the Intel machine!)
I traded it in for a Mac mini M2 Pro with 32 GB RAM when that model was introduced earlier this year, and I’ve not hit a hiccup since. I’ve been comfortable editing relatively complex GIS files, Docker containers, OBS…you name it, it runs it, and it runs it well. The Pro chip probably helps, but I know everything can fit because it’s got plenty of memory to work with.
I think there are a surprising amount of things that folks who get a base M3 with 8GB of RAM can do. I don’t totally poo-poo the notion that Apple Silicon can more efficiently manage memory, either – that M1 could certainly do quite a bit more multitasking more responsively than the Intel equivalent with the same RAM. And, of course, there’s the disclaimer that my workflow is generally out of the ordinary. But I still think that “Pro” means more than a base level of memory. Apple Silicon is pretty cool, but it can’t defy the foundational principles of computer science, either.