even though it “wouldn’t be datomic”, immutable-style postgres usage is very valuable and scales nicely. I’ve used it to implement ML-cooperative transactional data that had to mix with the ML-inferred data and deal with user transactions in the light of predictions to update the models.

I’m really not sure what the comment about queries is, but I gather that most developers have never really mastered query generation and struggle with it as though it were code. It’s not code; metaprogram it.

Honeysql 0.9.1 is now available

not much change from 0.9.0, there's a helper you can use to ensure values are inlined rather than parameterized

Also, more honeysql discussion in #honeysql