Using an LLM to migrate a Python web app to Rust seemed like a fun project, but then hit the bumps. Credit: Agus_Gatam - shutterstock Take a wild ride with us, as we use a large language model to convert a Python app to Rust. Also, could Pandas finally compel you to ditch Excel? And, is Python’s native JIT the Python performance booster we’ve all been waiting for? All this and more, in this week’s report. Top picks for Python readers on InfoWorld What I learned using Claude Sonnet to migrate Python to Rust“Let’s rewrite a Python project in Rust with an LLM,” I said. “It’ll be fun,” I said. It was also bumpy, eye-opening, and, yeah, kinda fun, actually. How to use Pandas for data analysis in PythonStop staring at spreadsheets and take real control of your data tables. Pandas gives you industrial-strength data tools to wrangle, crunch, and visualize your numbers. Get started with Python’s new native JITPython’s native JIT offers faster code with no rewrites, no new libraries, and just a new version of the Python interpreter. Mileage will vary, but if you’re using Python 3.14, you can try it right away. More good reads and Python updates elsewhere PEP 827: Type Manipulation (Proposed)“We propose adding powerful type-level introspection and construction facilities to Python’s type system.” In short: Make it possible to manipulate Python types in ways that are as powerful and expressive as Python itself. It’s an early-stages proposal, but with broad reach. Expect much discussion (and dissension). PEP 747: Annotating Type Forms (Accepted)Another powerful advancement for Python’s type-linting features. Possible use cases include describing in the type system if a given value can be assigned to a specified type or coerced to another type. Why it took three years and two attempts to get Python’s lazy imports featureOr: “How to add a huge new feature to a software ecosystem without fracturing it.” Read the backstory for one of Python’s most powerful new features. The resource-usage implications of removing the GILWill multicore builds of Python make workloads faster? Some, but not all. Will they change how much memory and power your system uses regardless of workload? Very likely. (Warning: This paper uses academic language!) Programming LanguagesSoftware DevelopmentPythonLibraries and FrameworksRubyData ScienceAnalyticsCareers