Python Trending Weekly
ISSUE 122

Python 3.14 is Here – But How Fast Is It Really?

2025-10-11 Season 5

TL;DR

Python 3.14 is out, with free-threading and t-strings stealing the show; PEP 810's lazy imports, the tdom templating engine and the AI-first Air framework sketch the ecosystem's next moves.

Articles 12 Projects 12
🦄

Articles & Tutorials

× 12
01 Python 3.14.0 (final) is here! pythoninsider.blogspot.com

Python 3.14.0 has finally dropped! This release is absolutely packed with game-changing features: official free-threaded Python support, template string literals, deferred annotation evaluation, multi-interpreter support, and Zstandard compression. Plus, there's an experimental JIT compiler, enhanced error messages, and official Android binary releases. This is huge!

02 Python 3.14 Is Here. How Fast Is It? blog.miguelgrinberg.com

The author put Python 3.14 through its paces using recursive Fibonacci calculations and bubble sort algorithms. Compared to Python 3.13, the new version delivers roughly 20% speed improvements in single-threaded scenarios and absolutely shines in multi-threaded environments. While it's still not quite matching PyPy, Node.js, or Rust, Python 3.14 shows serious momentum in performance optimization, bringing significant speed boosts to pure Python code.

03 My Top 5 Favourite Features in Python 3.14 blog.codingconfessions.com

The author breaks down Python 3.14's five standout features: free-threaded Python enabling true parallel computing, concurrent interpreter modules supporting multi-interpreter parallel execution, remote debugging capabilities that can attach directly to running processes, incremental garbage collection reducing latency spikes, and performance optimizations delivering substantial improvements across the board.

04 Introducing Air: A Pioneering AI-First Python Web Framework audrey.feldroy.com

Meet Air, a new web framework that's being developed to blend the best of Django, FastHTML, and Flask. It supports both Python object-to-HTML generation and Jinja templating, comes with built-in HTMX support, and shows incredible promise for AI code generation optimization. This could be a game-changer for AI-driven web development.

05 PEP 810 – Explicit lazy imports peps.python.org

This PEP proposes introducing explicit lazy import syntax for Python 3.15, using lazy import json statements to defer module loading. Modules only get loaded and executed when first used, not at import time, which can dramatically reduce startup times and memory usage. Smart optimization for modern Python apps!

06 Bringing NumPy's type-completeness score to nearly 90% pyrefly.org

Quansight Labs and Meta's Pyrefly team collaborated to boost NumPy's type completeness from a measly 33% to nearly 90%! They doubled coverage to 80% just by fixing a single type annotation bug, then pushed it to 90% by fully typing the MaskedArray class. This massively improves IDE code completion and static type checking experiences.

07 Full text search with Django and SQLite screamingatmyscreen.com

The author demonstrates how to implement full-text search in Django blogs using SQLite's FTS5 extension, avoiding heavyweight solutions like ElasticSearch. By creating virtual tables and database triggers to sync content, plus native SQL queries for search ranking, you get powerful search without the complexity.

08 Why it took 4 years to get a lock files specification snarky.ca

The author reflects on the challenging journey of Python's lock file specification PEP-751, which started in 2019 and went through multiple failed proposals and redesigns. Lock files need to record dependency graphs, support multi-platform compatibility, and handle the NP-complete problem of dependency resolution. Finally, through collaboration with uv, Poetry, PDM, and other tools, they reached consensus on the pylock.toml specification, bringing standardized dependency management to the Python ecosystem.

09 Python Violates PEP 8 inventwithpython.com

The author makes a fascinating observation: whenever someone complains about code violating PEP 8, it's always about snake_case naming conventions, never about any other PEP 8 content. The article explores what code style guides really mean, pointing out that PEP 8 emphasizes consistency over blind adherence – internal project consistency matters more than strict guideline compliance.

10 Why Reactive Programming Hasn't Taken Off in Python (And How Signals Can Change That) bui.app

The author analyzes why reactive programming struggles to gain traction in Python, noting that while RxPY is powerful, it's overly complex for simple state management. The article introduces the reaktiv library and how it uses signal mechanisms to implement transparent reactive programming, making state synchronization as intuitive as spreadsheet formulas.

11 Introducing tdom: HTML templating with t‑strings davepeck.org

Introducing tdom, a library that leverages Python 3.14's new t-strings feature for modern HTML templating. The library supports dynamic attributes, special attribute handling, intuitive content nesting, and JSX-like component syntax, while automatically escaping special characters for security. This makes Python web templating more intuitive and powerful than ever.

12 Koan 1: The Empty Path pythonkoans.substack.com

The first article in the "Python Koans" series dives deep into Python's concept of truthiness, explaining how to control object behavior in boolean contexts by defining __bool__ and __len__ methods. It shows that "truth" in Python isn't just a simple True/False binary concept, but rather how objects express their own meaning. (Shared by @julio-mcdulio)

🐿️

Projects & Resources

× 12
01 AI_NovelGenerator: AI Novel Generator github.com

A versatile novel generator powered by large language models, supporting world-building, character development, intelligent chapter generation, state tracking systems, and semantic search engines. Features a visual workbench that automatically connects context and plot threads. (2.5K stars)

02 pwndbg: Exploit Development and Reverse Engineering with GDB & LLDB Made Easy github.com

A powerful GDB and LLDB plugin designed specifically for low-level software developers, hardware hackers, reverse engineers, and exploit developers. Provides rich utilities and enhanced features that fill gaps in traditional debuggers, making the debugging experience much more user-friendly. (9.4K stars)

03 devpush: Like Vercel, but open source and for all languages github.com

An open-source, self-hostable deployment platform supporting Python, Node.js, PHP, and more. Features Git-based zero-downtime deployments, real-time log monitoring, team collaboration management, custom domains, and SSL certificates. Uses Docker containerization with multi-environment management and branch mapping support. (2K stars)

04 pyscn: An Intelligent Python Code Quality Analyzer github.com

A Python code quality analysis tool that detects dead code using control flow graphs, identifies code clones with APTED + LSH algorithms, and provides coupling metrics and cyclomatic complexity analysis. Built with Go + tree-sitter for blazing fast processing (100K+ lines/second), with HTML and JSON report output support.

05 FuncToWeb: Transform any Python function into a web interface automatically. Zero boilerplate github.com

Converts any Python function into a web interface – just add type hints and call run() to generate a complete web form. Supports all Python built-in types, file uploads, dropdown menus, optional parameters, and more. Features real-time upload progress bars and error handling.

06 Logly: Rust-powered, Loguru-like logging library for Python github.com

A high-performance Python logging library developed in Rust, offering a Loguru-like API experience. The Rust backend ensures memory safety and thread safety, supporting multiple log levels, structured JSON output, intelligent file rotation, multi-target output, and async callbacks. Zero configuration required, with rich filtering options and custom formatting.

07 ApeRAG: Production-ready GraphRAG with multi-modal indexing, AI agents, MCP support, and scalable K8s deployment github.com

Combines Graph RAG, vector search, and full-text search technologies, supporting multi-modal document processing, AI agents, and MCP protocol integration. Provides five index types (vector, full-text, graph, summary, visual). Integrates MinerU document parsing service, supports Kubernetes deployment and enterprise-grade management features including audit logs, model management, and graph visualization.

08 autoswagger: Autoswagger by Intruder - detect API auth weaknesses github.com

Automatically detects unauthenticated APIs by parsing Swagger/OpenAPI documentation. Supports multi-stage discovery mechanisms, parallel endpoint testing, and parameter value brute forcing. Integrates Presidio for PII detection and regex-based sensitive information identification. Can detect personal information leaks, key exposure, and other security issues in APIs. (1.5K stars)

09 computer-use-preview: Computer Use Preview github.com

Google's computer use preview model that controls browsers through natural language commands. Supports Playwright local browser control and Browserbase cloud environments, integrating Gemini API and Vertex AI. Simple commands let AI agents automatically complete complex tasks like web searches, form filling, and page navigation, with mouse position highlighting and visual debugging features. (1K stars)

10 varia: Download manager that supports regular downloads, torrents and videos github.com

A cross-platform download manager supporting Linux and Windows systems, capable of handling regular files, BitTorrent torrents, and video/audio stream downloads. Serves as a graphical frontend for aria2 and yt-dlp, offering multiple installation methods including Flatpak, Snap, and AUR, plus browser extension support. Features auto-updates, post-shutdown download completion, and multi-language support.

11 holm: Web development framework that brings the Next.js developer experience to Python github.com

A web framework built on FastAPI, htmy, and FastHX that brings Next.js-like developer experience to Python. Supports file-system routing, JSX-syntax components, server-side rendering, and HTMX integration. No JavaScript dependencies or build steps required, providing a fully typed Python development environment. Supports async components, JSON and HTML APIs, and automatic page layouts.

12 snakeviz: An in-browser Python profile viewer github.com

A browser-based Python performance profiling data visualization tool for viewing cProfile module output, serving as a graphical alternative to the standard library's pstats module. Offers icicle and sunburst chart visualizations, supports IPython magic commands and Jupyter Notebook integration, and provides intuitive displays of function call relationships and time consumption. (2.5K stars)

ESC
↑↓ select · ↵ open175 issues