This study is mainly basic science on the properties of XMRV's reverse transcriptase enzyme (RT) without drawing conclusions about the life cycle of XMRV itself. They did comparative analyses between XMRV RT, HIV-1 RT and MoMLV RT, in order to add to the knowledge base about RTs in general; this study does not attempt to say anything about XMRV as a pathogen.
They found that, in vitro, XMRV RT showed relatively better ability to avoid 'mismatch' errors while creating DNA strands (as compared to HIV-1 or MoMLV), but not in vivo (as compared to HIV-1 and related amphotropic MLV); therefore this study is inconclusive about how relatively mutation-prone it really is.
They looked at the known mutations for HIV-1 that make it resistant to AZT and tenofovir and engineered XMRV RT clones with mutations at the equivalent sites. Their engineered RTs did turn out to be less susceptible to AZT and tenofovir. That suggests two things - that XMRV has the potential to develop such resistances if the appropriate mutations evolve, and that the same mechanisms by which HIV-1 develops resistance to these drugs can be used by a relatively distantly related retrovirus like XMRV. Again, the latter point is more about understanding RTs via comparison than anything else.
It is useful to know as much as we can about XMRV RT, particularly what drugs might inhibit it, as it does indeed have to replicate in certain tissues and at some points in the infectious cycle. But as for their conclusions about likely mutation rates or recombination rates, I think the differences in this study's in vivo results suggest that they can't extrapolate yet from their in vitro ones. And of course, as G pointed out, anyone trying to predict mutation rates of MLVs in the wild via in vitro-observed mutation rates has to be aware of the likelihood that there will be little if any variation in the blood compartment, where replication takes place primarily via clonal expansion (where there will be no mutation) rather than RT activity. In blood at least, any relative differences in RT fidelity are likely to be irrelevant; the bulk of wild variation of the virus will not show up in blood isolates unless the actively replicating virus in tissue reservoirs actually makes it to the blood.