Opening Keynote
summary
Temporal unveils concrete features: Workflow Streams for real-time, durable LLM and tool output; Standalone Activities as a durable job primitive (public preview, five language runtimes); and serverless workers (AWS Lambda now, GCP Cloud Run soon) so workers scale to zero. Claims 35 billion workflows a day on Temporal Cloud; OpenAI reports 60x growth in 2025 using Temporal for long-running agentic workflows. Live demo shows an agent controlling a robot arm, streaming outputs, and surviving injected Lambda crashes via automatic durable retries.
AI-written summary · about
chapters (20)
- 0:00Welcome and keynote opening
- 1:18Temporal's origin and mission
- 6:02AI systems and durability challenges
- 8:53Temporal's role in AI infrastructure
- 10:18Major announcements overview
- 12:08Workflow streams for real-time AI
- 13:30Standalone activities for durable jobs
- 14:58Expanding AI ecosystem partnerships
- 16:47OpenAI's scale and Temporal use
- 23:54OpenAI's 60x growth with Temporal
- 25:41Future of agentic products
- 33:05Live demo introduction
- 33:33Hardware-controlled AI agent demo
- 36:43Adding workflow streams to UI
- 39:06Deploying agent on serverless workers
- 41:12Audience voting app demo
- 47:41Serverless recovery and scaling
- 49:29Treasure chest retrieval with AI
- 51:57Demo recap and key takeaways
- 54:06Building resilient production systems
related talks
01h 19mOpenAI + @Temporalio : Building Durable, Production Ready Agents
42mUsing Temporal Workflows for Seamless Dev + Research Collaboration & Continuous Algo D
33mFrom Prototype to Production: Making Agentic AI Systems Reliable
24mDurable Agents: long running AI workflows in a flakey world
01h 00mReliable Agentic Systems need Durable Execution 🤖
31mTemporal Substrate Architecture: Building Persistent Identity for Autonomous Agents
56mOpening Keynote
28mAn introduction to the Workflow Development Kit