Whenever learned database indexes come up, people get worried about worst-case performance compared to the predictability of B-trees. So, to preemptively clarif

Whenever learned database indexes come up, people get worried about worst-case performance compared to the predictability of B-trees. So, to preemptively clarify: the B-tree alternatives used in SageDB are able to give the same logarithmic upper bounds on query time, but can often do much better in terms of size and speed when there's exploitable structure in the data.