887 karma · joined November 15, 2018
These people are training on a completely different set of features than what Spotify is using. For example, listen time is likely to be a important feature where a short time or a skip will be a pretty good proxy for a dislike.
In other words, the authors conclusions are conditioned on only using the features outlined in the article, like, dislike, and super like. And I’m not sure it will generalize for Spotify.
Inference tasks isn’t like software testing where the states are well defined.
In a way, I feel the American government also wants FB to do the same as it advances the US centric world order.
If there is a global monopoly. American wants it. This is evident in America’s position in negotiating the digital tax with the EU.
Talks about it in this article