For the most part, I agree with you. Bed levelling done correctly, once should never need to be adjusted again. With a properly rigid frame and something to keep level consistent, it certainly should not matter. I won’t be going that route, less because I don’t think that it’s of benefit, and more because I don’t want to deal with the hassle of using 3 drivers for the Z axis.
My CoreXY designs have involved 3 screw bed support, but I was going to lock the screws into sync with belts, and drive them from 1 Z motor. If done well, and locked into synch with the belts, this should never need to be adjusted. Put good Z limit switch in place, and it shouldn’t need a Z probe, either, though with thermal expansion it seems to matter what you set the bed temperature to where exactly the zero point is.
My first attempt at a 3D printer, the mechanisms all worked great at 20-25°C but I started getting binding and the bed started warping at 70-75°C because it didn’t have enough allowance for thermal expansion. Everything ended up being over-constrained. Using 3 drivers allows for a little more looseness to the design, which should then reduce the necessary constraints, and in turn allow some freedom from the support rods, which will reduce problems like Z wobble.
Mostly though, I think it’s just become one of those things that a “premium 3D printer” needs to have as a feature, or else people will think that you’re making compromises to be cheaper at the expense of better. It does certainly allow use for those with less technical ability, so I guess that it does bring the 3D printer closer to the consumer electronics “press print, then wait.” state.