"Enabling" C threads in a Python / Wasm environment

The author details the significant hurdles of implementing high-performance C++ multithreading within a Pyodide (Python on WebAssembly) environment, specifically for real-time applications where Python's GIL and FFI overhead are too slow. His solution involves a custom-patched Pyodide built with -pthread and a specialized C++ thread pool that avoids deadlocks by using low-level Wasm buil...

https://yosefk.com/blog/enabling-c-threads-in-a-python-wasm-environment.html

Auto updating knowledge graph from meeting notes

Build a meeting knowledge graph from Google Drive using LLM extraction and graph database, with automatic continuous updates.

https://github.com/cocoindex-io/meeting-notes-knowledge-graph

Tenso

High-performance zero-copy tensor serialization for Fastest Transmission.

https://github.com/Khushiyant/tenso

randomized-svd

randomized-svd is a lightweight, high-performance Python library for computing the Randomized Singular Value Decomposition (rSVD).

https://github.com/massimofedrigo/randomized-svd

TrustGraph

The Context Graph Factory for AI. Build, manage, and deploy AI-optimized Context Graphs.

https://github.com/trustgraph-ai/trustgraph

What async really means for your python web app?

The author challenges the hype around async Python, demonstrating through benchmarks that for most database-driven web apps, synchronous Django with connection pooling often outperforms or matches async alternatives. The study concludes that unless a service is highly distributed or the primary bottleneck, the overhead of async can actually degrade performance, making sync the more effic...

https://hackeryarn.com/post/async-python-benchmarks/

Static Protocols in Python: Behaviour Over Inheritance

Patrick Müller explores how Python's static protocols enable "static duck typing," allowing developers to define interfaces based on behavior (structural typing) rather than inheritance (nominal typing). He argues that this approach promotes loose coupling and greater flexibility in software design, particularly in complex fields like machine learning where diverse libraries and coding s...

https://patrickm.de/static-protocols-in-python/

shareAI-lab / learn-claude-code

How can we build a true AI agent? Like Claude Code.

https://github.com/shareAI-lab/learn-claude-code

LLMRouter

An Open-Source Library for LLM Routing.

https://github.com/ulab-uiuc/LLMRouter

How uv got so fast

uv's speed comes from engineering decisions, not just Rust. Static metadata, dropping legacy formats, and standards that didn't exist five years ago.

https://nesbitt.io/2025/12/26/how-uv-got-so-fast.html

awesome-agentic-patterns

A curated catalogue of awesome agentic AI patterns.

https://github.com/nibzard/awesome-agentic-patterns

Unit testing your code’s performance, part 1: Big-O scaling

Your software’s speed needs some form of testing; big-O scaling is a good starting point.

https://pythonspeed.com/articles/big-o-tests/

Recent optimizations in Python's Reference Counting

CPython 3.14+ introduced LOADFASTBORROW, a bytecode optimization that skips reference count increments when loading local variables in hot loops, using static lifetime analysis to ensure safety. This reduces memory management overhead in performance-critical code like tight loops, mimicking Rust-style borrowing while preserving Python's reference counting model, with further JIT optimi...

https://rushter.com/blog/python-refcount/