The timescale, I think. If the ice at the poles melts, the water immediately (days?) distributes itself along the geoid, the earth's moment of rotational inerti

The timescale, I think. If the ice at the poles melts, the water immediately (days?) distributes itself along the geoid, the earth's moment of rotational inertia increases, it slows down. But, once the weight is taken off the poles, the crust rebounds over a period of thousands of years, making the earth less oblate, its moment decreases, it speeds up.