re:Invent 2023 - Day 3

▪️ AI Service Cards advance responsible AI
▪️ Amazon Q generative SQL is now available in Redshift Query Editor | Preview
▪️ AWS Marketplace
      ▫️ APIs for sellers
      ▫️ Partner CRM Connector
      ▫️ SaaS Quick Launch
▪️ Bedrock
      ▫️ batch inference
      ▫️ Claude 2.1 foundation model from Anthropic | GA
      ▫️ Evaluate, compare, and select the best FMs | Preview
      ▫️ Llama 2 70B foundation model
      ▫️ Stable Diffusion XL 1.0 foundation model from Stability AI | GA
      ▫️ Titan Image Generator foundation model | Preview
      ▫️ Titan Multimodal Embeddings foundation model | GA
      ▫️ Titan Text models—Express and Lite | GA
▪️ Clean Rooms Differential Privacy | Preview
▪️ DocumentDB vector search
▪️ MemoryDB for Redis vector search | Preview
▪️ Neptune Analytics | GA
▪️ OpenSearch
      ▫️ OR1 instances
      ▫️ zero-ETL integration with S3 | Preview
      ▫️ Serverless Vector engine  | GA
▪️ Partner Network
      ▫️ AWS Resilience Competency
      ▫️ Built-in Competency Partner software automates Installation for customers
      ▫️ new Amazon EKS Ready Specialization
      ▫️ Request a Cyber Insurance Quote
      ▫️ Solution Building Enablement
▪️ Redshift
      ▫️ Apache Iceberg support | GA
      ▫️ metadata security to simplify multi-tenant applications
      ▫️ row-level security enhancements | GA
      ▫️ Serverless enhanced manageability and usability features
▪️ SageMaker
      ▫️ better SDK tooling and UX for model deployment
      ▫️ Clarify now supports foundation model (FM) evaluations | Preview
      ▫️ HyperPod, a purpose-built infrastructure for distributed training at scale
   ▫️ new inference capabilities to reduce costs and latency
      ▫️ Canvas Leverage FMs for business analysis at scale
      ▫️ Canvas natural language instructions for data preparation
      ▫️ Model Training  smart sifting of data | Preview
      ▫️ Notebook create jobs API
      ▫️ Pipelines simplified DX for AI/ML workflows