Databricks Highlights DBRX: Cutting Edge Open Source AI

In late March, when Databricks Inc. announced DBRX, a general-purpose large-scale language model, it marked another milestone in the company's strategy based on three big bets when it was founded in 2013. .

“One, organizations will move to the cloud, two, successful companies will use open source, and three, machine learning will truly be at the forefront. Chris D'Agostino (pictured), Global Field Chief Technology Officer, Databricks. “DBRX is the culmination of decades of strategy.”

D'Agostino spoke with theCUBE Research analysts John Furrier and Savannah Peterson during Google Cloud Next 2024, broadcast exclusively on theCUBE, SiliconANGLE Media's live streaming studio. They discussed the integration of DBRX with MosaicML and how machine learning is being adopted across the Databricks portfolio. (*Disclosure below.)

Databricks leverages MosaicML retrieval

DBRX can be combined with Mosaic AI, a toolset for building and deploying AI models that is the result of Databricks' acquisition of MosaicML Inc. for an impressive $1.3 billion last year. MosaicML provides generative AI tools that Databricks customers can use to train and fine-tune models at low cost.

“We acquired MosaicML and there has been a lot of press about the price,” D'Agostino said. “DBRX is the first step in that journey. We will be able to build open models that perform as well as our own models, allowing our customers to do more with their own data in a highly secure way. ”

Databricks' CTO said the company is also focused on integrating machine learning in ways that enable it to offer products and services that go beyond building and training language models at scale.

“In addition to creating an environment where people can build and train their own LLM and other types of machine learning models, we are embedding machine learning within the Databricks platform,” D'Agostino said. states. “As new data comes in, you can train the model on the existing data, and as new data arrives, you can compare and contrast whether the new data is consistent.”

Below is the full video interview. Part of coverage from SiliconANGLE and theCUBE Research. Google Cloud Next 2024:

(*Disclosure: TheCUBE is a paid media partner for Google Cloud Next 2024. Neither Google LLC, the primary sponsor of theCUBE's event coverage, nor any other sponsors have editorial control over theCUBE or SiliconANGLE content. )

Photo: SiliconANGLE