The issue is that a) it can be greatly discouraging or impossible to fit in the subject's schedule, b) a waste of time and effort if the requirement isn't that hard, c) there may be a better way: the a, b, a, b pattern, which is to say: do A until you need to do some of B, then go back to A until you need more of B, and so on.
The a,b,a,b pattern is the right answer in at least some cases.
What sort of efficiency do you think this optimizes for?
In general I do think that B-before-A is a very useful approach, but it cannot be the only approach. For one, it doesn't scale, not as the chain of Bs grows longer over time. We teach a lot more math to students today than 100 years ago, but almost certainly we're taking shortcuts that essentially amount to a,b,a,b -- how else can we manage the otherwise ever-growing cognitive burden of the sum of our knowledge?