Yes, those implementation are slow compared to classical floating points. The goal, as I understand it, is to let people experiment with Posits in order to prove that they bring something to the table.