Jan 16
Loading video player...
Episode 1 of our 5-part "Platform Engineering 2026 Look Forward Series" tackles the question keeping platform engineers up at night: Will AI agents replace us or amplify us? AWS Frontier Agents can reason across 30+ steps and orchestrate infrastructure changes autonomously. The MLOps market is hitting $129 billion by 2028. Netflix's AI triage cuts MTTR by 40%. But where are the hard limits? We introduce the 60/30/10 Framework for deciding what to hand off to AI agents vs. what to guard. Timestamps: 00:00 - Series Introduction: Platform Engineering 2026 Look Forward 02:15 - What is Agentic AI? (Copilots vs Agents) 05:30 - AWS Frontier Agents Architecture: Reasoning + Tool Use + Memory 08:00 - Where AI Agents Excel: Netflix, Uber, Honeycomb Examples 11:30 - The Hard Limits: What AI Can't Do 14:45 - The 60/30/10 Framework (Delegate/Augment/Guard) 17:30 - Five Action Items for 2026 20:00 - Career Implications & Next Episode Preview Resources: - Episode page: https://platformengineering.org/podcasts/00067-agentic-ai-platform-operations-2026 - AWS Frontier Agents: https://aws.amazon.com/ai/ - Gartner MLOps Report: $129B market by 2028 - Netflix Engineering Blog on AI Incident Triage Key Points: - 60% Delegate: Log analysis, runbooks, cost optimization, alert correlation - 30% Augment: Incident response, capacity planning (AI suggests, human confirms) - 10% Guard: Architecture decisions, security posture, novel failures, stakeholder comms - The 20% AI can't do is 80% of the VALUE - Platform engineering evolves from "doing" to "orchestrating" #AgenticAI #PlatformEngineering #MLOps #DevOps #SRE #AWSFrontierAgents #CloudNative #Kubernetes #2026Predictions #AI #MachineLearning