Let's put rolling resistance (RR) in perspective. Per Wilson and Schmidt, 2020, Bicycling Science, 4th ed., the coefficient of rolling resistance (Cr) for a typical commuter tire on smooth, hard pavement is around 0.006. Which makes the corresponding RR equal to the slope resistance you'd encounter on a grade of 0.6%.
This is true at any total weight. And for many riders, a grade like that would be hard to detect — with or without assist.
So if you're not riding the Tour de France, and don't have to squeeze every last mile out of your battery, you can reasonably let the RR be the motor's problem.
Per Wilson and Scmidt's data, smooth hardpack has little effect on RR. (Loose, soft, and rough surfaces are of course a very different story.)
As Rene Herse's research shows, RR on pavement has a 2nd broad minimum at lower pressures. So an ebiker on smooth pavement or hardpack might reasonably pick tire pressure on desired ride feel alone, with rock bottom set by the pressure needed to maintain an acceptanble risk of pinch flats on the surfaces actually ridden.
Question is, how do you determine that rock bottom?