Ask HN: How do you migrate from static to dynamic typing

I've been using statically typed languages my entire life. I've been working with clients on some dynamic languages projects lately though. In a static language, I'm used to being able to e.g. change variable names, change class/record structures, etc and have all that either caught at compile time or (better) have the code change propagation fully automated by tooling.

With dynamics that obviously doesn't exist. I'm not complaining, just looking for coping mechanisms. Unit testing is obviously one, though I've found it's just as likely if you forget to update a function to use the new record format, you'll forget to change the test too, so lots of stuff gets through.

Do you essentially stop refactoring when doing dynamic code, or does it change what you refactor e.g. less "structural" or "naming" refactoring? Or do you find yourself using longer variable/function/class names such that a global text search will be less likely to produce collisions? Other things?

Mostly interested in opinions of people who have successfully/happily made the switch from static languages, especially if they previously liked static typing and were similarly disgruntled initially with dynamics. I'm curious how you've found your programming style has changed.

I recognize many benefits of dynamic languages; that's not what this post is about nor is it meant to start a flame war.