Keynote
summary
Keynote demos a new change feed mode all versions and deletes that records inserts updates and deletes into the same log used for point‑in‑time recovery and shows a live Python demo using the Python SDK to stream game events; claims dynamic autoscale saves up to 80% and is adopted by 80% of usage; cites Cosmos DB as main DB for OpenAI during growth to 400 million weekly active users; showcases DocuSign moving vectors into Cosmos (60× cheaper than in‑memory) and H&R Block hitting 400–1,000,000 RUs at peak.
AI-written summary · about
chapters (18)
- 0:00Welcome to Cosmos DB conference
- 1:22Cosmos DB powers modern apps
- 2:05Cosmos DB change feed with versions
- 3:10Building with Cosmos DB change feed
- 5:59Demo: change feed dice game
- 8:05Cosmos DB for MongoDB API
- 8:48Dynamic autoscale cost savings
- 9:51AI refactoring with schemaless data
- 10:35Vector and full text search in Cosmos DB
- 11:17Solving $2T agreements problem
- 13:05DocuSign search with Cosmos DB
- 14:51Why DocuSign chose Cosmos DB over alternatives
- 15:53H&R Block apps on Cosmos DB
- 16:38Modernizing tax systems
- 18:45AI Tax Assist with Cosmos DB
- 20:07SQL Server to Cosmos DB
- 21:31Cosmos DB tax season performance
- 23:21Closing keynote: future of Cosmos DB
