This whole subject of ice sheet effects on Earth's rotation is something that I have been asking about for over 2 decades, but mostly my questions have been ignored.
The current article about ice loss affecting day length is trivial compared to what must have happened when the 2-mile-thick ice sheets in the northern hemisphere melted rather rapidly about 20,000 years ago, raising sea level by around 325'. The Earth is still a little "pear shaped" as the rock slowly adjusts to having been depressed under the continents in the northern hemisphere, and some land is still rebounding upward, while the areas that were not glaciated are still slowly sinking back down.
So, what is happening now is only "unprecedented" in that humans are measuring and recording it. Much larger effects have already happened in human experience.
But, not only should the interglacial period melting have affected the day length by a lot more than what his article is talking about, I am thinking that it should have also affected the rate of "wobble" of our spin axis, just like pressing on a spinning top makes it wobble differently. Wouldn't that have changed the rate of precession and made different frequencies of the Milankovitch cycles? Could taking that into account help make the global climate models better match the geological records for the actual ice age frequencies?