Explore the new sampling profiler in Python 3.15, dive into no-impact Python profiling with the tprof library, see how Mojo 1.0 is taking shape, and more in our latest Python Report.
The first beta release of Python 3.15 has arrived. That means it’s time to discover all the best new features in Python, and especially its low-to-no impact profiling tools. Plus, with the recent debut of Mojo 1.0, you can see for yourself how the one-time Python challenger has set out on a path of its very own.
Top picks for Python readers on InfoWorld
First look: Mojo 1.0 mixes Python and Rust
What was once pitched as a potential Python replacement is evolving into a synthesis of Python’s easy syntax and package ecosystem and Rust’s native speed and memory safety.
The best new features in Python 3.15
Lazy imports, faster JIT compilations, better error messages, and smarter profiling make Python 3.15 a compelling upgrade. And you can try out beta 1 right now.
Hands-on with the new sampling profiler in Python 3.15
Another can’t-miss Python 3.15 feature: its low-impact statistical sampling profiler, for getting visibility into your programs without slowing them down.
Profile Python functions with zero extra code
For another powerful way to do no-impact profiling, the third-party tprof library makes use of low-level hooks in Python 3.12+ to garner runtime statistics without needing to instrument your code.
More good reads and Python updates elsewhere
Why is Python slow? An Anaconda engineer explains
In an episode of Beyond the Commit, Anaconda principal engineer Antonio Cuni, developer of PyScript and PyPy, explains why Python is slower than statically-compiled languages like C and Rust.
PyWry: Webview rendering in Python
A “build once, render anywhere” library for Python, PyWry lets you “prototype in Jupyter, deploy via FastAPI, and compile to a stand-alone desktop executable”, using PyTauri instead of Electron.
Flet Studio
Flet Studio is an interactive, in-browser dev tool for writing Python programs that run in-browser using the Flet framework. Although you can’t download a stand-alone, compiled Flet package from the studio, just the source code, the interactivity alone is worth the effort.
Using LLMs to find bugs in C extensions for Python
Daniel Diniz used Claude Code to find almost 600 bugs across 14 projects, and his workflow suggests a model to follow: Let the LLMs find the bugs, and let the humans write the bug reports and fixes.
