AI tooling is making Site Reliability Engineering more mechanical and less skilled, not…

Site Reliability Engineering emerged in the late 1990s as a distinct discipline within technology organizations, sitting at the intersection of software development and infrastructure operations. Understanding how AI tooling is reshaping this role requires understanding what SRE was designed to do, how expertise accumulated in the field, and what happens when certain tasks become more automated and less knowledge intensive. What SRE actually is Site Reliability Engineering is a job category that treats infrastructure operations as an engineering problem. Rather than following a traditional operations manual of reactive firefighting, SREs apply systematic approaches to things like system reliability, deployment safety, incident response, and capacity planning. The field emerged within Google and was later formalized in industry literature, creating a recognized set of practices and skills around keeping large systems stable at scale. SREs typically work on problems that blend software development with systems management. They write code to automate operations, design systems to fail gracefully, build tools to monitor distributed infrastructure, and develop playbooks for responding to failures when they inevitably occur. This means the role has traditionally required both deep technical knowledge and judgment about tradeoffs. The expertise that shaped the discipline Becoming effective in SRE has historically required accumulating substantial knowledge over years. An experienced SRE develops intuition about how systems fail, learns to recognize subtle patterns in logs or metrics that signal emerging problems, understands the architectural constraints of their specific infrastructure, and builds mental models of how different components interact. This expertise was earned through incident response, postmortem analysis, observing how other experienced engineers solved problems, and deliberate study of systems thinking and reliability principles. Much of this knowledge was difficult to systematize. It lived in the heads of experienced practitioners and was transmitted through mentorship and learning from others' mistakes. The value of an SRE partly came from this rare, hard won expertise. What modern AI tooling changes AI driven tools are making certain SRE tasks more accessible to engineers with less experience. Pattern recognition across logs, triage of alerts, drafting initial responses to common scenarios, and generation of documentation can now be handled by AI systems. These capabilities automate away some of the judgment calls that previously required years of experience to make well. The concern is not that these tasks disappear entirely but that they become more procedural and less valuable as sources of expertise development. An engineer following AI generated guidance through a common incident learns less from it than an engineer who had to reason through the problem independently. The deskilling tension This creates a paradox