> There are a lot of computations where 256 is too small of a range but 65536 is overkill

wasnt this more to do with cost? they could do arbitrary precision code even back then. its not like they were only calculating numbers less than 65537, ignoring anything larger