While I think I agree with most of this post, Design Sprints can be much more broad than a single user story. I've seen them be successful in determining whether or not to bring a new product to market, to build a new mobile version of an existing product, for growth opportunities or even adding more personality to a product.
So, while yes, they are more focused than "Discovery", I would argue that they're perfectly suited for testing the viability of a new product or feature. As for "what they're good for", IMO it's really all about validating ideas and gaining confidence for when you begin to build.
I think the obvious point here is that you cannot test and validate the universe in 4 or 5 days. It is best for a narrowed focus or user story. Sometimes that focus can encompass an entire product, as the article says, but that is more rare when the products you're designing are complicated and high risk.