AWS Summit New York City - Matt Wood's Keynote on Generative AI

Welcome and Opening Remarks

• Speaker: Matt Wood, VP of AI Products at Amazon Web Services
• Greetings to attendees in New York City and those joining via live stream.
• Highlight of event: learning, networking, and innovation.

Main Topic: Generative AI

• Generative AI: Identified as a major shift in interaction with data, information, and peers.
• Comparison to the early internet's impact.
• Generative AI drives growth in AI and ML workloads on AWS.
• Amazon Bedrock: Fast-growing service for building generative AI applications.

Customer Adoption and Industry Impact

• Growth: Multi-billion-dollar annual run rate for AWS AI and ML services.
• Enterprise and Startup Usage: High adoption amongst startups (e.g., 96% of AI/ML unicorns on AWS) and major enterprises.
• Regulated Industries: Fast adoption in financial services, insurance, healthcare due to robust data strategies compliant with regulations.

Why Customers Choose AWS for Generative AI

• Security: AWS ensures data confidentiality, privacy, and security without compromising generative AI success.
• AI Capabilities: AWS offers a broad set of AI tools, including foundational infrastructure, application-building services like Amazon Bedrock, and user-friendly applications like Amazon Q.
• Rapid Feature Updates: 326 generative AI features launched in 2023, surpassing other cloud providers.
• Sustainability: AWS matches 100% of its electricity usage globally with renewable energy.

Generative AI Stack

• Foundation Layer: Infrastructure for building, training, tuning foundation models.
• Middle Layer: Services like Amazon Bedrock for application-building with generative AI.
• Top Layer: Applications enabled by generative AI (e.g., Amazon Q).

Trends in Generative AI

• Cloud Foundation: Essential for AI workloads, infrastructure necessary for large-scale model training and inference.
• Access to Diverse Models: No one model fits all—importance of having a variety of models to suit various tasks.
• Amazon Bedrock: Provides a wide selection of models from multiple partners.

Enhancements in Generative AI on AWS

• Fine Tuning: Now available for Anthropic’s Claude 3 models using Amazon Bedrock.
• Guardrails for Safety: New contextual grounding checks to reduce hallucinations and filter harmful content.
• API for Guardrails: Apply security and safety protocols to models outside Amazon Bedrock.

Business Examples and Customer Success Stories

• NASDAQ: Leveraging AWS for cloud innovation and AI to improve market operations.
• Bayer: Using generative AI to advance agriculture technology and accelerate data science solutions.
• Ninja Tech: Successful startup utilizing AWS infrastructure for cost-effective generative AI solutions.

Generative AI and Automation

• Agents: Enable multi-step tasks, actuation with memory retention and code interpretation for advanced analytics.
• Amazon Q: Powerful assistant for software development, providing high-quality code suggestions, and now customizable for internal APIs and documentation.

New AWS Service: AWS App Studio

• Description: Generative AI-powered low-code application building service.
• Functionalities: Builds secure, scalable applications in minutes with natural language inputs.
• Deployment: Easy publishing, integration with other systems, and management of user access and data security.

Conclusion

• Future Outlook: Generative AI will integrate into many applications and processes, transforming productivity and innovation.
• Encouragement to explore and leverage AWS's generative AI capabilities for transformative outcomes.

Thank you for joining the AWS Summit New York.