All of these endless debates seem to boil down to two different groups who work in two different problem domains talking past each other. I've personally never

All of these endless debates seem to boil down to two different groups who work in two different problem domains talking past each other. I've personally never seen an ORM lead to success in the long run. But I also work in a space where queries frequently end up involving something that ORMs typically don't handle well: merge statements and pivot statements, window functions, management of the lock escalation policy to fine-tune performance, temp tables... The list is endless. What I have not ever worked on is a relatively basic CRUD datastore. Which I realize is what most people are using databases for. So at this point, I'm putting my money on ORMs being a hole in one for that application. Because, otherwise, I just can't reconcile a statement like, "99% of the time, an ORM is fantastic" with the reality I'm living in. In my career, 100% of the time, when an ORM was present, it was invariably the single biggest piece of technical debt.