Amazon AWS and Microsoft will compete in GenAI field

In an interview with VentureBeat on Monday, Swami Sivasubramanian, Amazon’s Vice President of Data and Artificial Intelligence at AWS, said that he oversees all AWS databases, analytics, Machine learning and artificial intelligence services, and briefly outlined Wednesday morning's keynote address and Tuesday morning's keynote address from AWS CEO Adam Selipsky

The main theme around GenAI is that enterprise There is a desire to have the flexibility and choice to use different models from different suppliers rather than being locked into a single supplier or platform, however, he added that the models themselves may not be enough to provide a competitive advantage as time goes on. , they are likely to become commoditized, so a key advantage for businesses will be their own proprietary data and how they integrate that data with models to create unique applications.

To support this vision, Sivasubramanya said Amazon is focused on emphasizing two aspects: first, invention, providing various GenAI models that customers can access through its basic services, and more Good, seamless data management tools for customers to use to build and deploy their own GenAI applications How to not only benefit from data, but in turn enhance and improve databases and data systems

Siva Subramanian revealed some highlights at Re:Inventent. Just two weeks ago, Microsoft showed off their plans to go all-in on GenAI at rival Ignite conference. In less than a minute: AWS's Bedock, launched in April, is a completely A managed service that allows customers to use underlying GenAI models provided via API. Sivasubramanya said bedrock is becoming easier to work with. He will feature some customer stories showing how easy and fast it is to build applications on Bedrock, with some examples taking less than a minute. He said customers such as Booking.com, Intuit, LexusNexis and Bridgewater Associates are among the companies using Bedock to create impactful applications.

BedRock has given enterprise customers more LLM options, including companies like Its own pre-trained foundation model Titan, as well as foundation models from third parties such as AI21’s S Jurassic, Anthropic’s Claude, Meta’s Llama 2 and Stable Diffusion. Expect to see more action in the future, including more about Amazon's partnership with OpenAI rival Anthropic, following a major investment in the company in September. "We will continue to invest heavily in vehicle selection," Sivasubramanian said.

Vector database expansion: Another area where GenAI models can help is vector databases, which support cross-image, Semantic search of unstructured data such as text and video. By using GenAI models, Vector Database can find the most relevant and similar data to a given query, rather than relying on keywords or metadata. In July of this year, Amazon launched a vector database feature-Vector Engine for its OpenSearch Serverless in preview mode. Sivasubramanya said the vector engine has gained impressive traction since its launch and hinted that it may soon become ubiquitous. He also hinted that Amazon might expand vector search capabilities to other databases in its portfolio. "You're going to see that as part of Bedrock we're going to make this easier and better, but also in a lot of other areas," he said.

In terms of Gen AI applications, Sivasubramanian hinted at some announcements related to the application layer of the enterprise GenAI stack. He mentioned some examples of applications that are already available and integrated with GenAI models, such as Amazon QuickSite. It is a serverless tool that allows customers to create and share interactive dashboards and reports. Additionally, there is Amazon HealthScribe, which automatically generates clinical notes by analyzing patient conversations with clinicians. These applications are designed to be convenient for users who may not have any GenAI or coding knowledge or experience

For enterprises with complex data needs, integrating data from disparate sources and formats is a key challenge. Traditional extract, transform, and load (ETL) processes are cumbersome and expensive, requiring moving data from one database to another, as well as performing data transformations and transformations. To avoid this friction, some cloud providers are developing "fabric" technologies that use open and standard formats for data exchange and interoperability. Microsoft has been touting its Fabric plans, but some analysts say Amazon has an advantage over Microsoft and Google with zero ETL. Sivasubramanya said Amazon has been trying to give developers a choice of databases and will continue to invest in its zero-ETL vision, which it began last year with the integration of some of its own databases, such as Aurora and RedShift. Enterprises want to store and query their vector data and other business data in their databases. Sivasubramanya said Amazon will continue to improve these services, while mentioning that Amazon's Aurora MySQL recently added support for vector searches. It is reported that Aurora MySQL is a cloud-based relational database. Amazon will make greater and more meaningful progress toward zero ETL in the future

During Selipsky and Sivasubramanya’s keynote, several AWS customers shared their Stories about how GenAI-based models can be customized through further training or fine-tuning to meet their specific needs and domains. However, they do so without compromising the security and privacy of the data, as the data remains stored in the customer's own virtual private cloud (VPC), a secure and isolated part of the AWS cloud. Sivasubramanya said that this is one of the important differences between AWS and other cloud service providers. At the same time, to ensure customer data security, AWS provides secure GenAI customization services

GenAI chip innovation: Finally, Amazon has been developing its own silicon solutions to support GenAI. Sivasubramanya said AWS will provide some new updates on the performance and adoption of its Nitro Hypervisor and Graviton series of chips, which are designed to deliver high performance and low cost for cloud computing. He will also introduce the company's Tradium and Inferania chips, which are designed for GenAI training and inference respectively

The above is the detailed content of Amazon AWS and Microsoft will compete in GenAI field. For more information, please follow other related articles on the PHP Chinese website!