Paris Keynote
summary
Keynotes list concrete product launches and hard numbers: Amazon Bedrock in Paris region with Mira and upcoming Anthropic Claude 3 models, Bedrock exposing 80 CloudTrail events and serving 10,000+ customers; customer wins include Lonely Planet cutting itinerary generation cost by ~80% and OpenFold speeding protein structure work 4–10×. Infrastructure claims: provisionable ultra‑clusters up to 20,000 GPUs, a NVIDIA‑AWS supercomputer at ~414 exaFLOPS, and AWS chips (Trainium/Inferentia) cutting training costs up to 50% — mostly announcements and metrics rather than live demos.
AI-written summary · about
chapters (31)
- 0:00Welcome and introduction
- 11:59AWS in France and innovation
- 13:24Technological revolutions history
- 15:12AWS cloud transformation
- 16:14AWS services and examples
- 18:01Generative AI new era
- 19:28Generative AI impact and use cases
- 21:14AWS generative AI accessibility
- 22:16Air Liquide case study
- 32:06Enterprise AI foundation
- 33:31AWS AI stack explained
- 34:33AI infrastructure and training
- 35:56GPU innovation and cost
- 37:00Project Stargate with NVIDIA
- 37:42AWS custom silicon benefits
- 38:48Hugging Face collaboration
- 39:52Amazon Bedrock in Paris
- 41:19Model choice and flexibility
- 42:01Responsible AI with Guardrails
- 43:28Hugging Face partnership
- 49:46Data foundation for AI
- 51:09Modern data strategy
- 51:33Adobe Firefly and AI
- 53:20Data transformation in AI
- 55:47Amazon Bedrock enterprise features
- 56:49Security and auditability in AI
- 57:53TF1 streaming transformation
- 1:08:50Amazon Q for developers
- 1:12:45AWS training and community
- 1:13:48Future impact and call to action
- 1:14:53Q&A with Anthropic
