I had to look at that video 4 times and try it out for myself to notice something different and even then it is very minor. When picking your END point - after typing END, before actually snapping to an endpoint - it does look like your dynamic input is set to cartesian coordinates instead of polar. But as soon as you clicked on the snappoint it reverted back to the polar coordinates. So while functionally it makes no difference, technically it is different from my experience where the dynamic input stays in polar coördinates while using a snap override. But you use a snap override specifically because you want to click on the screen to determine the input. So whether the dynamic input is in polar or cartesian coordinates while doing it, it doesn't make a difference.
I can get the same behaviour if I set the DDSETTINGS to this. So if you don't want it, set it to polar coördinates.

I made a .gif to show how the setting changes the behaviour.
You can see after typing END in the first case the dynamic input prompt is cartesian and in the second try the dynamic input prompt is polar (using ">" to denote the angle input)
EDIT: setting DYNPIFORMAT to 1 (cartesian) should always make second points and next points be shown as x,y inputs, but somehow dynmode reverts to polar input, so this might be a bug. I'd have to check in older version if this is also the same.