Future of AgenticAI: Architecting the Global Control Plane (Keynote)
summary
Company-grade platform reorients agentic AI around a distributed control plane not single agents, shipping an SDK, cluster runtime and event‑sourced persistence that supports 5 billion tokens/sec on an 18,000‑core cluster, active‑active multi‑region replication across 30+ regions, and immutable audit logs with zero‑trust networking. Claims include LLM accuracy at ~76–82% (18–24% failure), actor units ~4 KB enabling massive concurrency, DICE domain integrated context engineering to reduce context rot, and up to 95% less cloud compute vs certain Python frameworks.
AI-written summary · about
chapters (15)
- 0:00Introduction and company background
- 2:00What is Agentic AI
- 5:12Challenges of Agentic AI systems
- 8:23Barriers to production adoption
- 10:10Building trust in Agentic AI
- 11:31Six steps to trustworthy systems
- 14:46AA’s Agentic AI platform
- 16:53Why distributed systems fit
- 17:15Guaranteeing elasticity and resilience
- 19:25Domain integrated context engineering
- 20:09Real-time processing with streams
- 20:53Multi-region active-active replication
- 21:57Production metrics and use cases
- 23:43Live demo and next steps
- 24:04Q&A session
related talks
13mAgentic AI for Autonomous Micro-Frontend User Interfaces
17mAI TechOps: Data and Infra for AI Foundation to Scale and Manage 1000s of Cohesive Agents
06mAI Agents, Sovereignty & the Next Decade of Innovation
40mIn AI We Trust? Securing the Future, One Agent at a Time
06mThe Building Blocks of Agentic Al
24mAI Agents in Production: Why the Platform Matters
30mBuilding the Agentic Enterprise: Turning Data into Autonomous Business Systems
43mArchitecting the Future: Modern Solution Design with Agentic AI