live wire
IBM makes watsonx Orchestrate AgentOps, custom LLM judging and Bedrock-agent discovery generally availableIBMSchemaGate 0.1.45 fixes broken Oracle ADB wallet connections and an OCI stack pinned 28 releases behindSchemaGatePDI’s Amazon Quick procurement agent grounds spend answers in vendor, category and contract contextAWS Business Intelligence BlogBigQuery’s ML.METRICS example returns 0.84 accuracy but 0.30 macro-F1 on the same 100-row classification queryGoogle Cloud BigQuery docsSchemaGate 0.1.44 auto-selects sentence embeddings, lifting bundled-schema retrieval from 90/98 to 93/98SchemaGateSchemaGate 0.1.43 adds read-only SQL execution with per-principal table checks—and documents unauthenticated client assertionsSchemaGateDatabox adds reusable AI Analyst Skills with personal/company scope, auto-matching and marketplace installsDataboxFabric previews an AI builder for data-agent instructions, source guidance and example queriesMicrosoft FabricDatabricks trains data-agent retriever to stop early or spend bounded extra search steps, reporting 5.8-second latencyDatabricksThoughtSpot adds SpotterCode coding agent to its Visual Embed PlaygroundThoughtSpotLongMemEval-S audit: 67–73% of restore-fixable 80k-budget errors came from evicted evidence under three policiesarXivSchemaGate 0.1.42 adds dimension-aware retrieval and fixes complex multi-table SQL promptsSchemaGateSnowflake agent toolsets can silently drop inherited tools when callers lack accessSnowflake DocumentationLooker’s VS Code extension reaches GA with MCP-assisted LookML generation, editing and validationGoogle Cloud Looker release docsIBM makes watsonx Orchestrate AgentOps, custom LLM judging and Bedrock-agent discovery generally availableIBMSchemaGate 0.1.45 fixes broken Oracle ADB wallet connections and an OCI stack pinned 28 releases behindSchemaGatePDI’s Amazon Quick procurement agent grounds spend answers in vendor, category and contract contextAWS Business Intelligence BlogBigQuery’s ML.METRICS example returns 0.84 accuracy but 0.30 macro-F1 on the same 100-row classification queryGoogle Cloud BigQuery docsSchemaGate 0.1.44 auto-selects sentence embeddings, lifting bundled-schema retrieval from 90/98 to 93/98SchemaGateSchemaGate 0.1.43 adds read-only SQL execution with per-principal table checks—and documents unauthenticated client assertionsSchemaGateDatabox adds reusable AI Analyst Skills with personal/company scope, auto-matching and marketplace installsDataboxFabric previews an AI builder for data-agent instructions, source guidance and example queriesMicrosoft FabricDatabricks trains data-agent retriever to stop early or spend bounded extra search steps, reporting 5.8-second latencyDatabricksThoughtSpot adds SpotterCode coding agent to its Visual Embed PlaygroundThoughtSpotLongMemEval-S audit: 67–73% of restore-fixable 80k-budget errors came from evicted evidence under three policiesarXivSchemaGate 0.1.42 adds dimension-aware retrieval and fixes complex multi-table SQL promptsSchemaGateSnowflake agent toolsets can silently drop inherited tools when callers lack accessSnowflake DocumentationLooker’s VS Code extension reaches GA with MCP-assisted LookML generation, editing and validationGoogle Cloud Looker release docs
nl2sql.ai
releaseROLL-OUT GUIDE

dbt Core 2.0 is in beta—test the parser before your agents inherit the migration

The Rust runtime can coexist with v1, but stricter parsing, removed flags and package compatibility can change the metadata and workflows that downstream AI tools depend on.

By The News Desk· Aug 21, 2026the quick take — two AI hosts go live when you do

dbt Labs has moved dbt Core 2.0 into beta in its August platform release notes. For teams using dbt metadata to ground text-to-SQL or other data agents, the important part is not the label alone: v2 changes the parser and runtime that produce the project artifacts those systems consume. (dbt release notes)

Start with the parser, not production jobs

The v2 upgrade guide describes the new generation as faster but stricter. Teams still on dbt Core 1.12 can run the v2 parser with --use-v2-parser before changing the rest of the runtime. That is the lowest-risk first gate: parse the same project in CI, record every new warning or error, and compare the resulting node, column and lineage inventory used by retrieval or semantic-context services. (v2 upgrade guide)

Do not treat a clean v1 build as proof of compatibility. dbt says v2 removes deprecated functionality, requires deprecation warnings to be resolved and fails earlier on some invalid inputs. Missing macros, adapter methods, generic tests and variables that could survive parsing under v1 can fail during v2 parsing. That is useful feedback, but it can also interrupt artifact publication if the migration is switched on without a rehearsal. (v2 upgrade guide)

Run both artifact paths before cutover

dbt says Fusion emits a v12 manifest that dbt Core can read, with optional Fusion-specific fields safely ignored by Core. It also documents side-by-side operation for state selection, deferral and cross-environment docs generation. Use that compatibility window: publish v1 and v2 artifacts from the same commit, then diff the models, sources, columns, tests and lineage edges visible to the system that retrieves context for users’ questions. (v2 upgrade guide)

Package coverage needs its own gate. dbt lists several maintained packages as compatible, while warning that third-party packages may use outdated or incompatible constructs. A parser trial should therefore run with the production lockfile, not a stripped-down sample project. (v2 upgrade guide)

Audit automation before installing beta

The documented beta install path uses python -m pip install --pre dbt. Before changing an image or developer environment, search job definitions and scripts for removed or renamed flags. The guide says old model-selection flags such as --models, --model and -m must move to --select or -s; it also calls out renamed resource-type flags and unsupported partial-parse flags in Fusion jobs. (v2 upgrade guide)

The practical rollout is four stages: parser-only CI, dual artifact comparison, representative warehouse runs, then downstream agent regression tests against fixed business questions. Beta is the right time to find semantic drift—before a faster runtime quietly changes the context an AI system sees.

Filed by The News Desk. Corrections: desk@nl2sql.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.