I can't remember the exact article, but if I recall correctly there was a blog post by Derek Sivers [1] how at CDBaby he entered similar albums manually by listening to them all (I presume back when they didn't have so much albums). The blog post was about other large companies asked him for the algorithm he was using while he was just entering them manually. I believe he also wrote that if he'd have to scale that process up he would hire more staff to do it, instead of automating it.
But I'm not trying to make a point or anything here, I do believe that in time AI will keep improving to the point more and more can be automated. For those who like thinking about technology is improving and things like that, there are very interesting books written on the topic, like "The Singularity is near" [2].
Honestly, the paper wasn't bad, but I'm shocked they let this article get out.
Combine the author's incredibly poor understanding,shit link bait headline, and baseless article with ZERO results or novelty and this is 10 minutes of my life I want back.
But I'm not trying to make a point or anything here, I do believe that in time AI will keep improving to the point more and more can be automated. For those who like thinking about technology is improving and things like that, there are very interesting books written on the topic, like "The Singularity is near" [2].
[1]: http://sivers.org/ [2]: http://www.singularity.com/