Unfortunately, they also indicate that the desired substitution will be very hard to achieve — crucially, the "crush and separate the composite" suggestion a few commenters have made is unlikely to work, since the heterogeneity exists within the same crystal, depending on whether Cu substitutes into Pb {1} (good) or Pb {2} (bad) crystal sites. This may be why the very oblique synthetic approach favored by the authors — reacting copper phosphide with lanarkite, giving lots of copper sulfide byproduct — was needed to give interesting properties. Now that we have an idea of what to look for, though, it may be possible to derive other synthetic approaches with a clear idea of what particular sort of copper substitution should be achieved, and a way to determine if it was achieved.
Of course, there remains a possibility that all of this is a big mistake, but now it would have to be several correlated mistakes. It may be a long time until the necessary selective substitution is achieved in a high-quality bulk sample, so, going on this hypothesis, I do not expect a sudden rush of new technologies in the near-term.