Only one kernel and scheduler for all the containers, that doesn't have a lot to do if most of the processes are idle.
And I am not your bro.
And I am not your bro.
And presumably the person running k8s is probably not running them to have them idle. They are technically knowledgeable enough to be aware how much resources do they consume, and be excited for such comparison.
When you have kubernetes on your laptop, that is to test your code alongside a set of other microservices functionnally representative of a prod deployment. That doesn't mean your containers will have much load appart from your own punctual testing.