Comment by dijit

Fair point, but I think you’ve actually illustrated my argument perfectly: you didn’t eliminate the need for specialists, you outsourced them to your cloud vendor. Those underlying nodes being “patched and replaced automatically” by AWS/GCP/Azure? That’s their SRE teams doing exactly the work your sysadmin used to do, just at massive scale. The control plane managing your deployments? Cloud vendor specialists built and maintain that.

And I’d wager you’ve still got people on staff doing operational work, they just don’t have “sysadmin” in their title anymore. Someone’s managing your K8s manifests, debugging why pods won’t schedule, fixing networking issues when services can’t communicate, handling secrets management, setting up monitoring and alerting. That work didn’t vanish, it just got rebranded. The “DevOps engineer” or “platform engineer” or “SRE” doing that is performing sysadmin work under a different job title.

Managed K8s can absolutely reduce operational overhead compared to hand-rolling everything. But that’s not democratisation, that’s a combination of outsourcing and rebranding. The expertise is still required, you’ve just shifted who pays for it and what you call the people doing it.