Of course, the LLM is deciding which tooling to use and what params to pass, etc. So there is still some room for it to mess up. But it is better than the LLM alone.

I meant that your comments about whether you can trust the numbers coming from the LLM and things like that. Those numbers are coming from the analytics MCP Server tooling not from the LLM. I think the tooling makes it more trustworthy as it is not just statistically more frequently the answer.

It seems like maybe your comments are more focused on the LLMs ability to do these things on its own rather than the tooling providing the analytics based results. Thanks for sharing though, very cool!

Secure Data Sharing and Interoperability Powered by Iceberg REST Catalog | Cloudera Blog Many enterprises have heterogeneous data platforms and technology stacks across different business units or data domains. For decades, they have been struggling with scale, speed, and correctness requ...

Notes from Knox-ville - Apache Knox providing authentication and secure access tokens for Iceberg Tables through the Iceberg REST Catalog API. blog.cloudera.com/secure-data-...

What in the Bay Doors, HAL! Come on, people!

Cloudera AI Inference Service Enables Easy Integration and Deployment of GenAI Into Your Production Environments | Cloudera Blog Welcome to the first installment of a series of posts discussing the recently announced Cloudera AI Inference service. Today, Artificial Intelligence (AI) and Machine Learning (ML) are more crucial th...

Notes from Knox-ville - Apache Knox leveraged as an istio external authorizer in Cloudera AI Inference Service architecture! blog.cloudera.com/cloudera-ai-...

along with first bluesky typo, FTW!

Not sure why I thought of this as my first blusky post but here it is.