Python is taking on all contenders these days, as more languages elbow into Python’s domain of AI, machine learning, and data science. Those stories and more, in this week’s report.
Is Mojo still a contender for Python’s ML/AI crown? What languages outside of the Python ecosystem are good for data science? What’s the deal with Python dataclasses? And what’s this shiny new distributed-processing framework from the folks who gave us PyTorch? Find the answers to these questions and more, in this week’s report.
Top picks for Python readers on InfoWorld
Revisiting Mojo: A faster Python?
Until recently, it wasn’t possible to run Mojo on your own machine. Now that you can, it’s time for another look. Is Mojo still a contender for Python’s data science crown?
AI and machine learning outside of Python
There’s no question Python is the default choice for machine learning and data science. That doesn’t mean other options are off the table, though. Here’s a look at what you can do with Java, Rust, Go, and .NET.
How to use Python dataclasses
Python classes can be verbose, and even simple ones are often overloaded with boilerplate. Learn how to use dataclasses for more streamlined Python class creation.
PyTorch team unveils framework for programming clusters
Monarch, as it’s called, lets you program entire clusters of machines in parallel with the same directness and power as the APIs used in PyTorch.
More good reads and Python updates elsewhere
Decompression is up to 30% faster in CPython 3.15
A future release of Python uses a markedly faster version of the Zstandard decompresion library, for possible speedups when installing binary wheels and other scenarios.
The future of Python web services looks GIL-free
How might free-threaded Python change a common use case like web services? Giovanni Barillari runs it down, with side-by-side tests and benchmarks, then shares his findings.
Using Python and Rust to build a fast Model Context Protocol server
A walkthrough of using PyO3 to build Rust-enhanced tooling for AI agents, with a convenient Python front end.
PyO3 now supports Python 3.14
Speaking of PyO3, everyone’s favorite Rust-and-Python binding tool now supports Python 3.14, phases out everything pre-Python 3.10, and introduces a new .cast() API for better type conversions.
