AWS Bedrock Generally Available—Amazon’s Generative AI Launch
AWS Bedrock Generally Available—Amazon’s Generative AI Launch
Yesterday, AWS announced the general availability of Bedrock, its generative AI service initially launched back in April that offers a variety of foundational models (FMs) from AI companies AI21 Labs (Jurassic), Anthropic (Claude), Cohere (Command), Meta (Llama 2), Stability AI (Stable Diffusion), and Amazon (Titan). Up until now, Amazon Bedrock has been available to a limited set of users through limited preview.
Flexibility in selecting Foundational Models and Fine-Tuning with Your Data
Amazon Bedrock allows users access to any of the FMs through a common set of APIs, allowing flexibility in choosing a model that works best for a particular use case without making extensive code changes. It supports fine-tuning of the models by using your own proprietary data while ensuring that the data is kept private and secure, all without writing any code.
Additional Refining Completed at Runtime
Amazon Bedrock uses a concept named retrieval augmented generation (RAG) that allows additional data to be passed to the foundation model to generate more relevant, content-specific, and accurate results without the need to retrain the model. RAG gleans information from a knowledge base and provides relevant information to augment the input prompt at runtime.
Amazon CodeWhisperer—New Capabilities Coming
Amazon CodeWhisperer was generally available at the time of the April Bedrock announcement, however there are some additional features coming soon. Amazon is adding the capability to allow customers to customize CodeWhisperer suggestions to incorporate their private code base, enabling auto-generated code to use internal APIs, libraries, packages, and classes as well as better adhere to a company’s individual coding standards.
Bottom Line:
AWS has entered the Generative AI battle with some solid offerings that should position it well, especially for those users wanting to experiment with different foundation models. Additionally, AWS is providing the ability to easily and securely enable their users to extend the models with their private data, which should yield much more relevant generated results.
Policies and AI: A Guide for Enterprise Policies on AI and Its Usage
With the rapid adoption of AI, practical challenges inevitably arise. This extends not only to designing products and services that harness AI but also to the consumption of AI by both enterprises and individual knowledge workers. In this webinar, we will delve into the critical issues surrounding the deployment and management of AI, addressing both the application and policy aspects, both at the individual and enterprise levels.
Join Aragon’s expert analysts, Jim Lundy and Adam Pease, will provide insights into key considerations for enterprises in the realm of AI. Topics to be covered include:
- Why effective AI management is crucial for enterprises.
- The essential policies that need to be put in place for AI.
- Strategies for enterprises to harness AI’s power while safeguarding the enterprise’s integrity.
Have a Comment on this?