The whole book is full of little surprises. Here’s Ohno on ‘productivity increasing tools’, for instance — basically, increased productivity MUST lead to cost savings. If a tool makes life easier, this isn’t worth anything, because you’re still paying for the worker.
Conversation
To adapt this to a different field: cost savings from k8s is only worth it if you can fire one worker from your DevOps team.
Oh, is that screaming I hear?
Guess k8s doesn't qualify, then!