In that context, I don't think "hot" was referring to the blockchain's "Hot" feed. Hivemind is a database that sits in between the blockchain and the Steemit front end, and it's that database where I think the gap probably exists. If I understand right, "hot queries" in that post just meant that data for frequently used queries will be copied to and retrieved from a temp table in the hivemind database that's smaller and faster than the one the original table. The bigger/slower one is unchanged, but the goal is to improve performance by using the big one less frequently.

For some reason, though, the "created" feed is not retrieving the older data in communities that don't have much traffic. I assume that the change was probably unintentional, so hopefully it'll be fixed.