Having a good understanding of the minimum error in the disc cutting system (ie. just how flat a frequency response could be achieved, and how accurate are the test discs used to calibrate the cutting systems) will help make specifying minimum RIAA accuracy for reproduction less arbitrary. If disc cutting systems were accurate to 0.1 dB of RIAA from 20 to 20 kHz when properly set up, then I think the Neumann constant is worth looking into more deeply. If disc cutting systems were accurate to 1 dB of RIAA, then the Neumann time constant is a far smaller consideration.
I do believe it is a slippery slope to say that just because there are many other elements in the reproduction chain that introduce far bigger errors, we should ignore the potential influence of the Neumann time constant - especially if the Neumann time constant could be easily compensated for.