I think this is a very good and very underrated point. I'd like to see some evidence why there's any reason to assume repetitive movement with less stretching is somehow less harmful than repetitive movement with more stretching. I've been typing on the qwerty layout for about twenty-five years at this point, and I'm up to around 110WPM, so it would really take a gigantic amount of evidence for me to be bothered with trying to rewrite all of that muscle memory. There seems to be a tacit assumption that these alternate layouts improve something that qwerty is doing wrong, but as it stands I'm not convinced that premise has any merit.
Still an interesting article, none the less.