Incorrect formular for "Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change" paper?

[mapa17]

Could it be that the rotation matrix described in the paper summary of "Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change", in the embedding chapter of the NLP course, is incorrect?

It states that the rotation would be the maximize the distance between the two projected vectors. Having a look at the paper looking at equation 4, i think it should be a minimisation.