For real use, you are right. But to teach list programming to beginners, I think this is necessary. For instance, I would explain how to process String without regular expressions.
I still argue that this is a bad idea. To an extent, functional programmers tend to focus too much on lists as the be-all and end-all of data structures. Whilst lists are great at what they do and generally serve as a good "intermediary" data structure (e.g. converting a Map k a to a Seq (k,a) ), we should not be abusing them to do _everything_.
I don't think that's a realistic attitude at the moment. tailDropWhile was one of the first things I put in my personal library, along with various string functions like strip, split, and join, and I still use them all the time on strings. Certainly the performance sensitive parts of code use ByteString or Text, but the vast majority of strings are not in critical sections or are quite small and using Text everywhere would be awkward syntactically for no measurable gain. Also a lot of libraries only accept and produce Strings.