I over-simplified the premise. We want to identify interesting and unique content, produced by individuals/publications that may not have any real visibility. This would serve as a win-win for the creator and sharer. Obscure content - that isn't fortunate enough to be on HN or some equivalent - needs a way to be identified and considered for digestion besides popularity. I never said users wouldn't read it before sharing it. While companies could certainly abuse it, that's not the purpose. If I had a smarty-pants blog, but no blog followers, I'd be happy that a deep-learning service classified my theoretical physics blog for consideration/syndication somewhere based on substance of the posts...