“The collection of 300 terabytes of audio and metadata targeted in the scrape.”