Databricks tested a stronger model against its multi-step agent on hybrid queries. The stronger model still lost by 21%.
When queries span databases and documents, better models don't fix the problem. Databricks research shows why architecture does.
Sean Michael Kerner
Intuit compressed months of tax code implementation into hours — and built a workflow any regulated-industry team can adapt
No pitch, no IRS forms, hard deadline: inside the AI tooling Intuit built to ship accurate tax code under pressure.
Sean Michael Kerner
Amazon S3 Files gives AI agents a native file system workspace, ending the object-file split that breaks multi-agent pipelines
Gartner and IDC analysts say S3 Files removes the last friction point between enterprise data lakes and autonomous AI.
Sean Michael Kerner
Oracle converges the AI data stack to give enterprise agents a single version of truth
Matt Kimball, vice president and principal analyst at Moor Insights and Strategy, told VentureBeat the data layer is where production constraints surface first.
Sean Michael Kerner
Ai2 releases MolmoWeb, an open-weight visual web agent with 30K human task trajectories and a full training stack
Ai2 releases the first open-weight browser agent with its full training stack, letting enterprises audit and fine-tune without a proprietary API.
Sean Michael Kerner
Enterprise AI agents keep operating from different versions of reality — Microsoft says Fabric IQ is the fix
Microsoft opens Fabric IQ's business ontology to any AI agent via MCP, targeting the fragmented context problem breaking multi-agent deployments.
Sean Michael Kerner
Nvidia BlueField-4 STX adds a context memory layer to storage to close the agentic AI throughput gap
Nvidia's new reference architecture targets the storage bottleneck slowing agentic AI — and most major storage vendors are already co-designing on it.
Sean Michael Kerner
Agents need vector search more than RAG ever did
Qdrant raised $50M arguing agents made vector search harder, not obsolete. Two production deployments show where general-purpose databases break down.
Sean Michael Kerner
The team behind continuous batching says your idle GPUs should be running inference, not sitting dark
FriendliAI's InferenceSense turns idle neocloud GPUs into an inference revenue stream — and could put downward pressure on token prices.
Sean Michael Kerner
Databricks built a RAG agent it says can handle every kind of enterprise search
Databricks built a RAG agent it says can handle every kind of enterprise search. Here's how reinforcement learning did what distillation couldn't.
Sean Michael Kerner
ServiceNow resolves 90% of its own IT requests autonomously. Now it wants to do the same for any enterprise
ServiceNow resolves 90% of its own IT requests autonomously — 99% faster than humans. Its new role automation framework bakes governance into the execution layer from day one, so AI agents inherit permissions instead of reasoning past them.
Sean Michael Kerner
IBM's $40B stock wipeout is built on a misconception: Translating COBOL isn't the same as modernizing it
IBM just had its worst single-day stock drop in 25 years because of an Anthropic COBOL announcement. Analysts say Wall Street confused translation with modernization. They're not the same thing.
Sean Michael Kerner