When scientists want to work on cutting edge problems they must first study all the knowledge achieved before. Kids need to study calculus even though calculus already exists. Then another and another topic.
Given enough time the amount of study required to master a field will exceed a lifespan, and then future research is impossible.
How can this problem be structurally tackled? Hyperspecialization is not a good solution since it also involves learning many topics just to reach the specialty.