March 17, 2026
Snakes, speed, and side‑eyes
Python 3.15's JIT is now back on track
Python speeds up—cheers, side‑eyes, and “just use PyPy” chants
TLDR: Python 3.15’s new JIT finally shows real speed gains, but free‑threading isn’t ready yet. Comments split between “just use PyPy,” skepticism about big changes, and calls for clearer explanations—making this a win with a side of drama that matters to anyone who wants faster Python.
Python’s new just‑in‑time compiler (a feature that speeds code by turning it into machine code while it runs) is finally delivering real wins: about 11–12% faster on Mac ARM chips and 5–6% faster on Linux PCs, per doesjitgobrrr.com. The blog from core dev Ken Jin frames it as a hard‑won comeback after funding drama and months of “is this worth it?” Yet the crowd isn’t just popping champagne—this is tech Twitter meets Reddit: half hype, half heckle. The loudest chorus? The PyPy gang asking, “Why not use the faster, already‑JIT‑ed Python?” Others demand plain‑English explainers for the JIT jargon, complaining that discussions are buried in dense pull requests. Free‑threading (letting Python use multiple CPU threads freely) isn’t ready yet, and that’s triggering anxiety and curiosity—one commenter linked an interview about how hard it’ll be to change the whole ecosystem. There’s also a spicy “I’m not your employer’s guinea pig” vibe from folks wary of big corporate pushes. And the meme crowd? Loving the “does JIT go brrr” chart and posting snake emojis for speed. Bottom line: progress is real, but the community is split between cautious optimism, “use PyPy,” and “please explain this like I’m five” energy.
Key Points
- •Python 3.15 alpha’s CPython JIT meets early performance targets: ~11–12% faster on macOS AArch64 vs tail‑calling interpreter and ~5–6% on x86_64 Linux vs standard interpreter.
- •Benchmark figures are preliminary geometric means; individual tests range from ~20% slowdown to >100% speedup (excluding the unpack_sequence microbenchmark).
- •Proper free‑threading support is not yet available in the JIT, with support targeted for Python 3.15/3.16.
- •Earlier CPython JITs in Python 3.13/3.14 often underperformed and the Faster CPython team faced funding loss from its main sponsor, creating uncertainty.
- •A community‑led plan set goals (5% by 3.15, 10% by 3.16) and reduced bus factor by adding maintainers across JIT stages; middle‑end contributors grew from two to four.