Sure, there are realistic applications of Machine Learning with great results.

But the article has failed in its headline goal ("ML is easier than it looks") if it chooses an example that is mathematically more complicated and still less effective than a naive alternative.

The first difficult task is to identify an effective method.