
You built the platform. Now something else is using it.
And it does not pause. It does not ask. It does not notice when something looks wrong.
There is a 1968 film that made this exact point.
HAL 9000 was not evil. He was efficient. He had a mission, the access to pursue it, and no built-in sense of when to stop. The danger was never his intelligence. It was confidence without limits inside a system that never defined them.
John Willis, one of the original voices behind DevOps, used that exact reference in his keynote at DevOpsCon Berlin. Not as a sci-fi metaphor. As a warning for how we are building right now.
His framing: unlimited knowledge at machine speed, and every micro-delegation creates a blast radius.
Your agents are not HAL. But the architecture problem is the same.
Permissions scoped for humans. Ownership that lives in someone's head. No explicit boundary between what the agent can touch and what it should never reach.
That is what changes with AI Platform Engineering. Identity, permissions, governance, and accountability become the foundation, not an afterthought. The platform stops being delivery infrastructure for developers and becomes the structure that decides how autonomous action is allowed to happen at all.
Read the Full Article : devopscon.io/blog/governing…
Learn More about AI Platform Engineering Day : devopscon.io/ai-platform-en…
#DevOpsCon #PlatformEngineering #DevOps #CloudNative #SRE
English





















