I had a CS professor as an undergrad who would teach a couple of advanced seminars in his own research area. His approach to those simplifications was to announce, "I'm going to lie to you now, but just go with it and I promise that later we're going to learn the real truth." I liked that as a compromise, to make some practical progress, but not to mistake the simplification for full understanding. (And he wasn't rigid about it -- if somebody would ask a deeper question he'd happily answer it to some level and then get on with his plan.)