Remarks on the Kuramoto-Sivashinsky equation