If you have the date in the where clause of the query, it would be rather trivial to partition the table by day and significantly reduce the size of your working indexes.

Shared block lists are a great feature.

~23.4mm according to bsky.jazco.dev/stats

Take a look at the article posted here (and the rest of the website for more info on the protocol):

atproto.com/articles/atp...

For the code, I'd start at their Github org - github.com/bluesky-social

ORMs are great at what they do. That being said, I do all DDL in handwritten SQL, and schema generation is one of the first things I disable with any ORM that does it.

I'm a pretty big user of Hibernate/Spring Data and SQL Alchemy myself.

Were? I still go there regularly.

And don't forget:
fark.com for offbeat news
thedailywtf.com to learn what not to do when coding