AWS re:invent 2023 Recap
By: Date: 09/12/2023 Categories: AWS Tags:
  1. Amazon Q AI assistant: GenAI-base chat tool and AI assistant where knowledge workers can query via NLP using their data.
  2. Amazon Q integration w/ Amazon Connect: Automatically create post-call summaries and support workers in real-time with natural language prompts.
  3. Amazon Q Code Transformation: enable developers to complete Java language upgrades in a “fraction of the time”. Devs at AWS carried out 1,000 Java upgrades in just two days.
  4. Amazon Transcribe updates w/ GenAI: Adding over 100 spoken languages for speech recognition and call transcription allowing for programmatic use of speech-to-text capabilities to their apps.
  5.  Amazon S3 Express One Zone: Delivering single-digit millisecond data access for your most frequently accessed data and latency-sensitive applications, improving data access speeds by 10x and reduce request costs by 50% (compared to S3 Standard) while scaling to process millions of requests per minute.
  6. Zero ETL (4 types): quickly analyze data w/o building and managing complex ETL data pipelines. Amazon Aurora, DynamoDB, and RDS for MySQL integrations w/ Redshift simplify access to transactional data from multiple relational and non-relational DBs in Redshift using Amazon OpenSearch Service to perform full-text and vector search on DynamoDB data in near real time.
  7. AWS version of Starlink (Project Kuiper): $10B plan to build a network of 3,236 satellites in low Earth orbit providing internet access anywhere in the world, starting with breaking ground on a $120 million pre-launch processing facility in Florida.
  8. Amazon DataZone AI recommendations: powered by Bedrock, generates detailed descriptions of data assets and their schemas, and suggests analytical use cases, dramatically decreasing the amount of time needed to provide context for organizational data, automating the traditionally labor-intensive process of data cataloging.
  9. Amazon CodeWhisperer update: enable generative AI-powered code suggestions to help users remediate security and code quality issues, highlighting a concerted focus to weed out hard-to-find security vulnerabilities that typically evade built-in security scans.
  10. Amazon ‘AI Ready’ initiative:  project to provide 2M people with AI skills training by the year 2025. AI Ready will consist of three new initiatives aimed at adults, young learners, utilizing existing AI training programs.
  11. Guardrails for Amazon Bedrock: define and enforce boundaries for LLMs to deliver safe and relevant user experiences by preventing LLMs from responding to irrelevant or inappropriate questions by setting out-of-bounds topics, filtering offensive content, and excluding PII data.
  12. Agents for Amazon Bedrock: build and configure autonomous agents in your app. so end-users can complete actions based on organization data and user input – orchestrated between FMs, data sources, SW apps, and user conversations, calling APIs to take actions and invoke KBs.
  13. Titan family of FM: Titan Text Lite, Titan Text Embeddings Model, and Titan Text Express, and access to other FMs including AI21abs, Meta Llama2, Cohere, and Anthropic (Claude 2.1) – which effectively reduced hallucinations by 50% while outperforming GPT 4.0 in token capacity.
  14. Amazon EC2 Capacity Blocks for ML: enables scale to hundreds of GPUs within one cluster, providing capacity when they need it, and only when they need it. 
  15. S3 Glacier Deep Archive storage class: secure long-term storage for large amounts of data at a price w/ off-premises tape archival services – stored across 3+ AWS AZ w/ retrieval in 12 hours or less. VNA replacement?
  16. Amazon Detective w/ AI Updates: remove the “heavy lifting” for security analysts and improve the speed and efficiency of investigations and incident response using GenAI to create finding group summaries to “enrich” security investigations.