Yeah bloom filters aren't used by default my bad, by default you get no indices at all! I was thinking about the tokenbf/ngrambf indices. These help a ton in improving more sparse queries.

I will need to check out the materialised views. :)