I see no paradox here. Yes, the rubrics change over time, making morality relative, but the motivation (empathy) remains constant, meaning you can evaluate morality in absolute terms.
A simple analog can be found in chess, an old game that’s fairly well-defined and well-understood compared to ethics. Beginners in chess are sometimes confused when they hear masters evaluate moves using absolute terms — e.g. “this move is more accurate than that move.
Doesn’t that suggest a known optimum — i.e., the most accurate move? Of course it does, but we can’t actually know for sure what move is best until the game is near its end, because finding it is hard. Otherwise the “most accurate” move is never anything more than an educated guess made by the winningest minds/software of the day.
As a result, modern analysis is especially good at picking apart historic games, because it’s only after seeing the better move that we can understand the weaknesses of the one we once thought was best.
Ethical absolutism is similarly retrospective. Every paradigm ever proposed has flaws, but we absolutely can evaluate all of them comparatively by how well their outcomes express empathy. Let the kids cook.