Would be nice to see some more concrete data/references. This is a big claim. I mean, if it is "basically the same" but "faster and requires fewer examples" you

Would be nice to see some more concrete data/references. This is a big claim. I mean, if it is "basically the same" but "faster and requires fewer examples" you can just say that it's "strictly better" than K-means, and stop using K-means ever. However, everybody is aware of K-means and personally it's the first time I hear about "Columb exemplar algorithm".