Or because you respect the perspective that introducing new skills brings. I'll pick the example everyone loves to hate; Java developers. I used to work at a shop where most of their people had invested in "stable tech", which for them largely meant writing Java code that would get shipped onto Tomcat or WebSphere or something similar.
It was hands down one of the worst places I have ever worked from a technical perspective. Anything that didn't fit neatly into that universe was a no go. We had some version of WebSphere that was like 12 years old because people had gotten "deeply into it". High availability was basically non-existent because everyone was "deeply into" a toolkit that didn't provide it natively. Everything was built on "stable tech" that generally didn't support HA (like putting files into CIFS without any locking, because that was the stable tech we had chosen).
I'm not saying everyone needs to chase the latest and greatest thing, but the ecosystem moves because needs change. Sometimes learning new tech is less about how to use the tech, and more about understanding what problem that tech is solving. Perhaps the newest tech isn't a perfect solution to that problem, but there is always value in understanding what problem it is attempting to solve. Kubernetes is not a perfect solution to a problem, but avoiding it to a degree where you don't even understand the problems it is trying to solve will devalue you as an engineer.
Always keep your running shoes around your neck.