Comparing 2014 1-year Supplemental Estimates to Earlier 3-year Estimates

I'm curious to know if anyone is contemplating using the recently-released ACS 1-year Supplemental Estimates for geographies of 20,000+ residents in comparison to the earlier 3-year releases.

In particular, is this statistically kosher? My sense is that it's OK because sampling time periods are non-overlapping, but a colleague has doubts.