As large language models continue fueling the artificial intelligence fire, the need for open LLMs, enhanced customization and adaptability — which boosts customer centricity — is growing.
Since open LLMs can be fine-tuned and adapted for domain-specific tasks, Teradata Corp. has set things in motion through its Bring Your Own Large Language Model feature, designed to drive faster enterprise return on investment with AI by deploying real-world generative AI use cases with flexibility, ease and trust, according to Hillary Ashton (pictured), chief product officer of Teradata.
Teradata’s Hillary Ashton talks to theCUBE about why open LLMs are a force to be reckoned with.
“I think that our customers who are in the global 10,000 market really want to be able to manage their own environments,” Ashton stated. “They’re not going to be putting data out there in the public domain. They want to be able to bring their own large language model. They want to be able to customize it and tailor it, maybe make it small or medium size. Customer context really matters in the era of AI. BYO Large Language Models like or whatever works.”
Ashton spoke with theCUBE Research’s Rob Strechay and co-host Savannah Peterson at Teradata Possible during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed why organizations need open LLMs, such as BYO LLMs, to create a more diverse, customizable and adaptable AI landscape. (* Disclosure below.)
Starting the ball rolling with open LLMs
Given that leveraging open LLMs is a fast-emerging alternative and disruptive force, Teradata has joined hands with Nvidia Corp. This move is meant to foster AI usage and research, according to Ashton.
“Teradata is known for so many things, but one of the biggest things that we’re known for is our ability to scale with massive amounts of data and analytics, and who better than Nvidia and Teradata to come together to help our customers with large language models in this new era of AI?” she said. “Nvidia and their graphics processing units are going to help our customers scale beyond what they’re able to do today with large language models.”
With data being the backbone of AI, aspects such as data meshing are important since they make AI better. This is because data meshing offers a decentralized approach to data management, and Teradata makes this a reality through QueryGrid, Ashton pointed out.
“We have a product called QueryGrid, which lets our customers get data wherever it is,” she explained. “If you want to do remote push-down processing, you can do that, which is very efficient. You don’t have to relocate the data to do that … [you can] extend that out to gen AI use cases, and I think it’s even better because a lot of that unstructured data is sitting elsewhere. It’s not in a highly structured environment.”
Here’s the complete video interview, part of SiliconANGLE’s and theCUBE Research’s coverage of Teradata Possible:
(* Disclosure: TheCUBE is a paid media partner for Teradata Possible. Neither Teradata Corp., the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)
Photo: SiliconANGLE