Yes, we follow Dgraph pretty closely. It is a cool project, but it is distributed first with all the challenges (and advantages) that brings. My general impression is that they are google (or similar) engineers that are focused on solving google type problems, whereas 95 to 99% of databases or data projects don't need sharding etc. and we are in the area. In fact, we think that as main memory becomes larger and cheaper the advantages of in-memory, on even big DBs, is going to become clearer.