npx @buildinternet/releases get dbt-labsdbt login command and the bundled dbt-state plugin (including --manage-state/manage_state and the DBT_ENGINE_STATE_* env vars), reverting PRs…Dependencies declared inside an overloaded UDF's overload body are now carried onto the root function node, fixing dropped DAG edges. Parse-time memory and macro-generation overhead reduced via virtualized macros and lazy no-copy manifest views.
The Open Semantic Interchange (OSI) project has been accepted into the Apache Incubator and renamed to Apache Ossie (Incubating). The specification, community, and mission remain unchanged; all references to OSI are historical and will be removed over time.
Flags introduced in dbt-core 1.9 and 1.10 now default to true. Added catalog_database field to CatalogWriteIntegrationConfig so catalogs.yml can override the database name for any catalog type, and fixed partial parsing ghost-duplicate errors and unrendered config kwarg display.
The experimental dbt login command and bundled dbt-state plugin have been removed, including the --manage-state/manage_state flags and DBT_ENGINE_STATE_* environment variables. Also fixes an intermittent parsing error when unit testing versioned models caused by non-deterministic filesystem parse order.
DuckDB now supports read-write access to Iceberg REST and Unity catalogs via catalog-aware materializations, plus write support for Horizon and Unity attachments. Fixed state:modified selector incorrectly flagging seeds, snapshots, and data tests when their configs use environment-aware Jinja expressions—a significant correction for multi-environment workflows.
Stopped emitting spurious deprecation warnings for state model configs (pre_clone, lag_tolerance, require_fresh_data_from) by adding the ModelState schema to bundled jsonschemas.
We've seen this pattern often enough to name it. A team migrates to dbt, spends six months, and ends up with a dbt project that looks exactly like their old workflow, just with Jinja templating instead of drag-and-drop. The data model still has the same problems. It's only…
AI agents do not fail because of a weak model. Rather, they fail because of using infrastructure that was never meant for autonomous decision-making.
When it comes to getting their data AI-ready, many organizations start with cleaning and structuring their data and then simply stop. This is an important first step, but it's not the last step, because AI-ready data relies heavily on context: the layer of meaning that explains…