FYI, other then the 16G RAM limit the current mini is very fast. For many work loads even with 16G it is fast (I think the SSD being so fast helps it perform even with what for modern systems is a pretty limited amount of RAM)
Rosetta 2 is shockingly fast. For some things it is faster then running on a real x86 (atomic memory operations are dramatically faster on the M1). I think the real problem with relying on something that needs Rosetta2 is R2 requires some hardware assist, and at some point Apple will likely stop putting that in the CPUs and stop supporting R2. (the hardware assist being maybe just having a MMU bit to support the x86-64 style memory reordering in hardware so you get x86-style enforcement of what does and does not get covered by locks and other atomic instructions, but it may also have an 80-bit FPU as well)
I do a lot of compiles, and while clang-llvm as far as I know only uses one core most build systems will launch multipole instances at once. Apple’s linker however is single threaded (there are actual linkers more or less drop in comparable that get multicore speedups, but Apple doesn’t currently ship one, and they are a wee bit on the experimental side…).
I don’t know if I am the actual target audience though. I kind of expect it would handle my work load faster then the iMac Pro I currently have…
