Hybrid cloud, intelligent infrastructure, and the future of AI with IBM and HashiCorp
summary
Claims that agent proliferation creates a blind spot—current ratio ~1:50 agent-to-human ratio and code-generation accuracy only 30–60%—and demonstrates concrete controls: using Consul as an agent registry, Terraform Actions to run pre-change runbooks, and Vault-issued delegated tokens (one-time, time-bound) for agent access to payroll, SAP, CRM and mainframe systems. Shows real failures—agents with trivial passwords exposing job applications—and argues for identity-based, zero‑trust governance across hybrid clouds rather than IP‑based controls.
AI-written summary · about
chapters (8)
- 0:00Introduction and speaker background
- 1:28Hybrid cloud infrastructure overview
- 5:43Application lifecycle management with HashiCorp
- 7:56AI agents and security risks
- 13:11Managing non-human identities
- 16:20Zero trust for AI agents
- 19:11Standardizing AI governance across hybrid cloud
- 20:37Key takeaways and closing
related talks
40mIn AI We Trust? Securing the Future, One Agent at a Time
13mKeynote: Agentic AI Is Running Your Infrastructure
20mCIAM for AI: Authn/Authz for Agents
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…- 24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
39mManaging Trust Between Multiple Clouds and Non-human Identities Without Passwords
21mSecurity in the Age of AI & Copilot
01h 11mKeynote - Securing AI: Building Trust in the Era of AI