The slowdowns are reduced in number and slowliness. If they occur now, they can be explained by heavy disk activity due to other processes. (That was not the case two weeks ago ! I am sure.) Is it (under lax superstition filters) possible that the DVD writer needed a few tries to get in working shape ? All problems were with media that were used at a very early moment in its life.
I can imagine that unit slows down when it "steps" on virgin block, one which was never written to. In such event unit has to "resume background formatting," which involves an update in lead-in area. Naturally there should be only one such slowdown per recording [assuming that we still to sequential recording, such as those performed with growisofs]. However, no such drastic slowdowns were actually observed with other units as you report... In other words I suppose I don't have explanation for the phenomena you're describing... A.