Comment by lukeplato

I remember reading that for k-d trees to be able to split on k dimensions the dataset needs to be > 2^k, which becomes unwieldy pretty quickly