July 10, 2026
Threads, shade, and a Python flex
Show HN: Runloom – Go-style coroutines for Python free-threaded
Python fans are cheering, doubters are side-eyeing, and everyone’s asking: why not just use Go?
TLDR: Runloom claims to let Python handle huge numbers of tasks at once using normal-looking code, with speed tests bold enough to invite comparisons to Go. The community reaction is split between hype that this could be a big moment for Python and suspicion that the project is too huge, too new, and too hard to trust yet.
A new project called Runloom just crashed into the tech conversation with a huge promise: write normal, old-school “blocking” Python code and still make it fly across many CPU cores at once. In plain English, it’s trying to give Python the easy feel of regular code with the massive multitasking swagger people usually brag about in Go, a language famous for this kind of thing. The creator even posted benchmark numbers showing it can go toe-to-toe with Go on speed in some tests — which is exactly the kind of claim that gets the comments section foaming.
And oh, the comments delivered. One camp was instantly dazzled. People called it “seriously impressive” and started wondering if Python was about to get its own “Project Loom moment” — basically, a big turning point where doing lots of tasks at once suddenly gets much easier. Others were far less ready to pop the champagne. The loudest skeptical reaction came from readers staring at the project’s sheer size and complexity, with one commenter practically clutching their pearls over 12,303 lines of C, 244,740 lines of Python, and the fact that it all seemed to land in one giant commit just hours earlier. Translation: cool demo, but can we trust this thing yet?
The funniest mini-drama was how fast the community split into tribes. Team “This could change Python” battled Team “Why not just use Go?” while another person brought up gevent, the older Python concurrency tool, like an ex being awkwardly mentioned in the middle of a flashy new romance. The overall vibe? Excited, suspicious, and deeply entertained.
Key Points
- •Runloom is a Python library that provides Go-style stackful coroutines and a blocking-code programming model for free-threaded CPython.
- •The article claims benchmark parity with Go in several runtime metrics, including connection churn and keep-alive echo throughput, while reporting faster pure-C spawn rates.
- •Runloom uses a hand-written assembly context switch, a C M:N work-stealing scheduler, per-goroutine PyThreadState snapshots, and cross-platform netpoll backends.
- •The library supports monkey-patching of blocking standard-library modules so existing synchronous code can run cooperatively, and it also offers an async bridge via `runloom.aio`.
- •The article states that multi-core gains require free-threaded CPython 3.13t/3.14t, while memory use per parked fiber remains significantly higher than Go.