“Fearless concurrency” in future versions of free-threaded Python

Overview

With the new free-threaded or "no-GIL" versions of Python, it's now possible to run Python code with full parallelism on multiple CPU cores. It also means "data race" bugs are easier than ever to create, where it's hard to guarantee the sequence in which a given object will be altered by code in different threads. One of the discussions at the recent Python Language Summit confronted this problem, and in this video I unpack their proposed solution in detail, and what implications it has for the rest of the language.
For details about the talk itself: https://pyfound.blogspot.com/2025/06/python-language-summit-2025-fearless-concurrency.html

Register Now