As I understand it, there (at least) two ways to use the GPS system. The primary way most of our consumer-grade GPS appliances (cellfone, car, yacht nav system, etc.) operate is using the doppler shift of the GPS signals to compute a position - this has the advantage of being very fast, leading to only tiny delays in depicted location, and the “jitter” of position errors can be filtered out fairly well.
The second mode is, in my experience, called “survey mode”, using the actual timing offsets of the satellites and since SA (selective availability) was turned off, can result in very accurate position info - at the expense of significant computational load, which takes computational capability (which costs power) and time . This method is capable of measuring things like continental drift and land mass motion (millimeters per year) though there must be significant filtering taking place for sub-millimeter measurements.
For something like a car or a speedboat, the “fast with jitter” mode makes the most sense - but the slower and more accurate method may be more practical for applications (like ship heading or rig position) where high accuracy is desired.
There may well be other technology in use here - I haven’t kept up - - but it is good to know that there is more than one way to derive position info from the GPS constellation (including the use of GLONASS and the other sat systems up there).