@mitchellh Designing human experiences and deciding on tradeoffs is the last thing I want agents doing, it's the part I love most about my job. But that doesn't mean I need to read all the code. Great instructions = human. Validating they've been followed = agent
An SRE I know spent 6 months evaluating AI root cause analysis tool: "We still can't decide if it's worth it."
Wrong question. Instead of "Is it accurate?" ask "What's the CAIR score?"
CAIR = Value ÷ (Risk × Effort to Fix)
New post: https://t.co/Sdgm2a2pvc
@bwplotka@dnsmichi@dev_el_ops 😂 I know, I'll take the blame for that.
Hopefully we'll soon be able to answer the 2nd question too: "...and what did we change?"