It always strikes me as odd when a company claims it’s working on the discovery problem — for music, books, social media follows, or any other content domain. I’m not convinced a single solution exists, or even makes sense. A better company might claim it’s working on yet another vector for discovery.
We’re foragers by nature. I tend to value the albums I’ve collected in large part as a function of the circumstances of their discovery. The more unique the path to discovery, the higher a value I place on them. This is independent of obscurity; I might take a circuitous route to discover a well-known title. If my musical world collapsed down to a single recommendation algorithm, or even a single friend, I imagine I’d ultimately find none of it very interesting.