If you know something can be started and stopped if it doesn’t work, then you must start it.
For example, taking software to production, but if nobody uses it, you can always stop the server and just move on and build something different.
So taking risks when you know about the consequences and its effects is okay if you can revert to it as it was.