Good teachers experiment - maybe something they taught last time didn't go over too well, how can that be improved? Can they make the material even more relevant this year?
Also, it makes teaching it more interesting, rather than regurgitating lessons. Teachers are human after all.
Education isn't like the naive "encoding/decoding" model of communication, where the subject matter is simply "transmitted" from teacher to students. Even if the subject matter remains stable over time, many other things do not: changes in the media of communication, signal interference (say, from the average classroom size drastically increasing over time), all kinds of changes in the teachers and students themselves, changes in society's expectations of what constitutes success or failure (e.g. rote learning is now widely seen as having many shortcomings), changes in what students actually need to move forward (a career in the trades may well require a much lower standard of understanding in the age of the pocket calculator), ... this list could go on and on. Teaching is not a "solved problem" like that.
It is true that 50 years ago a fraction of kids were learning just fine, but more recently the goal has been to make that fraction larger, in a society that actively devalues learning.