The observation in our preprint, that the D614G mutation was associated with higher viral loads in subjects, but not associated with greater disease severity (indicated by fewer PCR cycles needed for detection, Fig. 5C and 5D of Korber et al., based on clinical data from Sheffield), has recently been repeated by the Bedford lab in Washington: https://github.com/blab/ncov-D614G. So some progress on this issue, of course more work will need to be done.
The global pattern of repeated shifts over time from the D614 to G614 variant continues to be supported as data accrues, in dozens of regions in parallel throughout the globe. There is an interesting exception in California, where sequences of the original D614 are currently dominant. That could be explained by a bolus of available sequences from Santa Clara county in late April; D614 is the local epidemic form in Santa Clara county. To see what is happening in California at the county level go to the Tracking Mutations tool, but here is a summary as of today: California, a county breakdown: May 11, 2020
In our original bioRxiv preprint, we noted that a mutation in Spike at position 943 seemed to be accruing. This position turned out to be a sequence processing artifact, which we have corrected in our bioRxiv preprint.