I can see how it might get old hearing that a lot, but when I look at the suite of tech screen questions we have here, several of them are basic under the hood (tree-building and traversal, string manipulation, etc). However, they all do have something about them that made them interesting to me when I did them to calibrate my interviews.

Things can be interesting without being hard or novel, and as Aphyr's entertaining "Hexing the interview" series shows, you can often find something interesting for purely personal reasons that are separate from the question itself. The interesting part might be that a normal data structure isn't enough, or that the difference between the naiive solution and the performant one is substantial. It might just be that we might be doing a task that we normally use a library for (string manipulations) in our normal work, so it's refreshing/challenging/interesting to look at things from a different level of abstraction.

> Seems that a lot of people start with that, robotically.

Can't tell if you're trying to be an ass or not.

People can use whatever they want. The point, as many others point out, is to come up with something you'll use to not say, "ummm..." so much.