No point really. The Nyquist sampling theorem says that 44.1kHz is overkill, much less 48kHz or anything beyond. You only need twice the sample rate of the highest frequency to be reproduced, and human hearing generally goes up to 20kHz (less for almost all adults). Accordingly, many production recording equipment won't even bother with frequencies approaching 20kHz. The only conceivable point is that you don't need to resample files in higher sample rates, which saves you a tiny bit of cpu time I guess.
absolutely nothing outside of the recording studio. It's useful when handling intermediate s when you're mixing several recordings. Once the mix is done, it's useless
Dynamic processors (e.g. compressors, limiters, peak detectiors) are more accurate at higher sample rates (and bit depth). Also, less latency at higher frequency. Lastly, it greatly improves editing including "modern" processing such as time streching, pitch correction etc. I am not sure what the effects on "spatialization" are ...
This is why I put playback in brackets, where it makes no difference at all. On DAC's it can be useful, because you can use them for other stuff, but in headphones and IEM's its completely useless.