🙂 Sounds very similar to my prior proposal from when Metrics were being discussed: https://github.com/dbt-labs/dbt-core/issues/6379#issuecomment-1338199055 https://github.com/dbt-labs/dbt-core/discussions/4621

👑 - nice reply by Callum

This is either the birth or the death of dbt and I have no idea which

There's a lot of prior art available for semantic models (SQL Server, Mondrian, and MicroStrategy were very strong in this area), and I'm slightly concerned with the piecemeal approach that doesn't seem to match with a robust long-term plan.