Fireside chat
summary
Argues that agentic AI adds a new interface but not new fundamentals: platforms must expose real APIs to legacy systems via things like Model Context Protocol so LLMs can act on scoped user tokens, use RAG truth stores and guardrails, and require agentic observability and agentic debugging for predictable behavior. Predicts 80–90% of existing platform skills apply, warns of huge hosting costs (Nvidia‑scale hardware), and urges equal investment in training human engineers and structuring RCA data.
AI-written summary · about
chapters (9)
- 0:00AI trends and platform impact
- 1:28Fundamentals of platform engineering
- 5:17Hosting and managing AI models
- 8:35Investing in human intelligence
- 10:52Technical reality of AI agents
- 13:43Operationalizing human knowledge
- 16:52Developer and platform culture
- 20:01The evolution of platform engineering
- 25:53Agency and self-development
related talks
38mPlatform Engineering in the Age of AI Teammates
44mAI, Corporate Responsibility & Democratic Legitimacy
06mThe Next Transformation?
33mOperationalizing GenAI Applications with DevOps Practices
10mLightning Talk: Building AI That Ops Teams Actually Trust
03mInterview
54mArtificial Intelligence seen from the software development lifecycle perspective
01h 02mPlatform Engineering in the age of Generative AI