For real world problem solving, a 32 or 64-bit processor is easier. But for learning it doesn't make any difference. In fact the limited size of 8-bit is an educational advantage as the programmer is forced to deal with overflow rather than just ignoring the possibility. You can work on interesting problems like "write the code to multiply two 16-bit values". I find these sort of problems an excellent way to start learning assembly language.
While 8-bit processors take multiple cycles per instruction, the number of cycles is usually fixed. On superscalar pipeline processors it is very difficult for the programmer to predict the performance without profiling.
Looking back at what I studied in my EE degree in the early 90s, my first introduction to assembly language and hardware was with the 6809. Then I moved on to the 68000. I am probably biased but I think starting with 8-bit then moving to a simple 32-bit architecture is a good introduction to computer hardware.