A new partnership between Nvidia and Snowflake will allow companies to build generative AI assistants trained on their own data.
what it does Nvidia offers Snowflake customers GPU-accelerated computing and its underlying Large Language Model (LLM) called NeMo. A customer can use this along with their own business data to develop his LLM for advanced generative AI services such as chatbots, search and summaries.
Let’s dig deeper: 6 marketing automation use cases where AI can help with data quality
why do we care The biggest weakness of generative AI is its tendency to make up facts. This is especially true for systems such as ChatBot, Bing, and Bard that have been trained on very large datasets. One of his ways to mitigate these “hallucinations” is to train the system on smaller, limited amounts of data. Nvidia uses the NeMo Guardrails program to make that even less likely. We are absolutely in favor of anything that makes data more reliable.
“Nvidia and Snowflake are jointly building an AI Factory that will help companies transform valuable data into custom-generated AI models to drive groundbreaking new applications, directly from the cloud platforms they use to run their businesses. We will,” said Nvidia CEO Jensen Huang. statement.
This isn’t the first service to allow companies to use generative AI in this way. Last month, Microsoft launched Azure AI Studio. This will allow customers to create AI-powered custom apps called Copilots. These can take many forms, such as running as chatbots.
Get Martech! every day. free. It’s in your inbox.
