Is there any update on this? Basically you cannot access Materialized Views with ML Clsuters. To copy all tables for our Data Scientists seems like a really unnecessary step. Also they cannot profit from the advantage of the incremental table updates like others that can use shared cluster or SQL warehouses.