Of course. What I meant is that if you have a dataset for which you can give an answer by loading it all into memory on a single machine -- it's not a big problem. That isn't to say that every problem which is better solved using a distributed system is a big problem.
I agree. If you can solve the problem in a reasonable amount of time on current high-end hardware, it's not a big problem. If the only way to meet demand is by distributing the solution, then you can start calling it a big problem.