Is There a Difference Between Biased Reference and Deferred Reference Counting?

In Python, each object has a thread owner. The process for what reference counting is for the owner is different than the process for the other owners. This means that some objects may not have to be locked or accessed by all threads at once. It's part of how we do reference counting and how our object model is. In C++, you'd totally know when the destructor was going to happen. But in Python, we don't really talk about destructors too much.