A study by ClickHouse found that large language models (LLMs) aren't currently capable of replacing Site Reliability Engineers (SREs) for incident root cause analysis, despite advancements in AI. LLMs can be helpful tools, but require human oversight.
