It's become quite infamous in technology circles as an example of people not realizing how simple problems for an one audience (engineer in this example) can solve incredibly difficult, intractable problems for another audience (plain Jane & John Doe).
But it also highlights how this feature became so necesary for its target audience even when alternatives exist.
For instance, you could buy hard drives (~\(50 at the time of writing), thumb drives (~\)140), or iCloud storage (~$9.99/mo for twice as much storage) and get the same thing.
But it's not quite the same thing is it?
People have their stuff on Dropbox and moving to another one is hard. Dropbox is either free or cheap enough to not make worth the effort.
Similarly, Dropbox just works on a lot of platforms. I've used it forever and it's native linux client works well. Other solutions are limited by platform or require manually uploading. This isn't quite as nice.
It's easy as an engineer to forget which parts of your software are the most useful for your customers. That's why Y Combinator gives the standard advice of "go talk to your customer." Using a tool like HotJar reveals that people don't use your product like your expect. Monitoring this is paramount to providing value to customers.
Or, in other words, everything is a feature until its necesary.