- Python Weekly
- Posts
- Python Weekly (Issue 766 October 8 2026)
Python Weekly (Issue 766 October 8 2026)
Rahul Chaudhary
October 08, 2026
In partnership with
Welcome to issue 766 of Python Weekly. We have a packed issue this week. Enjoy it!
The Ultimate Claude Code Guide to ship like Anthropic engineers
AI will write 90% of code by the end of 2026, and only 10% of developers will stay relevant.
The engineers in that 10% aren't smarter. They just know how to use AI.
We put together the exact playbook Anthropic engineers use to make sure you're in the top 10%.
Sign up for The Code and get access to:
The Ultimate Claude Code Guide 2026 — 50+ tips and tricks to code 5x faster
The Code newsletter — learn the latest AI tools, tips, and skills to code faster with AI in 5 minutes a day
Articles, Tutorials and Talks
Learn how Datadog made Python profiling async-aware while reducing profiler overhead and preserving context across asyncio tasks.
Python 3.15 introduces practical enhancements such as lazy imports, a built-in immutable dictionary, and streamlined unpacking in comprehensions. These updates, alongside a new sampling profiler and default UTF-8 encoding, aim to improve developer productivity and application performance.
The author outlines an end-to-end ML project evaluating whether TypeSafe’s Jev, accessed through OpenRouter, and smaller local models can effectively filter political fundraising emails from a personal inbox. He ultimately builds a 23.4 MB quantized ONNX SetFit model using mdbr-leaf-mt and confidence thresholding, creating a fast, CPU-friendly local classifier while exposing the remaining failure modes on harder boundary cases.
CPython developers are exploring a gradual introduction of Rust, starting with an optional Rust implementation of zlib in Python 3.16 before considering broader use in areas like I/O, JSON, parsing, and memoryview. The proposal aims to improve safety and performance without rewriting CPython, while addressing concerns around platform support, dependencies, build complexity, and whether core developers are comfortable adopting Rust.
Python 3.15 continues the interpreter’s steady performance gains, with Miguel Grinberg’s benchmarks showing noticeable improvements over earlier releases but not a dramatic across-the-board leap. The results also highlight how performance varies by workload, with some benchmarks benefiting much more than others from recent CPython optimizations.
A Visual Guide to Bag-of-Words, RNNs, CNNs, Transformers, Decision Models with Jev-like APIs, and Calibration.
Most teams add a second database for analytics. Then manage sync, lag, and drift forever. TimescaleDB extends Postgres instead. Hypertables, 95% compression, aggregates. No pipeline.
This tutorial demonstrates how to build and train a compact 25-million parameter multimodal language model from scratch using modern techniques like mixture-of-experts and sparse attention. It further explores the AI researcher's role by guiding viewers through pre-training, reinforcement learning experiments, and the critical importance of designing effective data-driven environments.
Django can serve the apple-app-site-association and assetlinks.json files needed for Apple Universal Links and Android App Links, letting installed apps securely claim website URLs and support password autofill and passkeys. The post walks through the required JSON, Django views, tests, caching, and production verification with Apple’s CDN and Google’s Digital Asset Links tools.
NumPy is being ported toward Python’s Limited API so it could eventually ship abi3 wheels that work across multiple Python versions, reducing the large per-version wheel matrix and easing support for new Python releases. The same work is also pushing NumPy toward subinterpreter support by replacing static types and global state with heap types and per-module state, though multiarray remains the hardest part and abi3 wheels are not ready yet.
Vercel’s Python AI SDK adds support for Jev, a fast universal classifier that answers structured choice, scoring, and probability questions without task-specific training. The post tests Jev on classifying Python vs. English and generating Python ASTs, showing that it works well for narrow decisions but still struggles when pushed into general code generation.
Glazer replaced race-prone parallel startup with readiness signals between Python and Docker containers, allowing components to start concurrently without assuming that running means ready. That design cut cold starts from 8–10 seconds to about 4.5 seconds, and caching the Tor data directory brought startup down to a consistent 1.9–2.1 seconds.
The post explains how the VEX shell AI agent reduced its cold start time from about 4 seconds to ~2 seconds by bypassing heavy initialization for quick CLI commands and deferring expensive imports until they are needed. It used tools like tuna to identify import bottlenecks, moved type-only imports behind TYPE_CHECKING, and lazily imported heavy dependencies inside function scopes.
Pendulum's + operator inspects the call stack to alternate between standard-library wall-clock arithmetic and its own elapsed-time DST logic based on the caller's function name. This fragile hack resolves a conflict between drop-in datetime compatibility and DST-aware arithmetic, but makes addition roughly 600× slower and breaks on different call stacks such as PyPy or dateutil.
The article speeds up binary search by batching independent searches with NumPy’s vectorized operations, then reformulates the algorithm to use only O(1) extra memory. After porting it to C++, the approach achieves up to a 25× speedup over NumPy 2.4’s implementation.
This tutorial demonstrates how to build a custom AI agentic harness in pure Python by progressively integrating LLM communication, tool calling, and Model Context Protocol (MCP) support. It covers the full development loop from scratch, enabling persistent memory and model routing to create functional, autonomous agents.
The article develops a mathematical model for more efficient Indic touchscreen keyboards, starting with Tamil and mapping script structure into Boolean logic and compact input states. It uses Python with SymPy to derive and simplify the logic, then implements the resulting input engine in TypeScript.
CPython allocates memory through layers such as pymalloc, arenas, pools, and blocks, while reference counting and garbage collection manage object lifetimes. The article also explains why freeing Python objects does not always immediately return memory to the operating system.
ScyllaDB is building a new Python driver on top of its Rust driver using PyO3, with most serialization, deserialization, request execution, and concurrency handled in Rust for better performance and maintainability. Benchmarks show especially large gains under concurrent workloads, while the project still needs production hardening, broader feature coverage, and a compatibility layer for applications using the legacy Python driver API.
Microsoft is expanding local AI development on Windows with experimental GGUF support in Windows ML via llama.cpp, new native inference APIs, and an OpenAI-compatible endpoint for running models locally. The broader stack also adds native PyTorch and Triton support on Windows Arm64, giving developers a path to train and optimize models with familiar tools and deploy them through Windows ML across CPUs, GPUs, and NPUs.
Interesting Projects, Tools, and Libraries
An open-source AI accelerator, developed by AI: RTL, ISA, simulator, compiler and profiler in one repo. Runs Qwen3, LFM2.5 and Qwen3.5 on a Kintex-7 PCIe card.
A Python-to-C++ compiler. Typed Python to native binaries with no Python runtime, no GC, no GIL.
Keep Python on your laptop. Run PyTorch operations and hold tensors on a remote GPU, including from a Mac with no CUDA installation.
A Python research framework and terminal-based IDE (RuleFlow Studio) for modeling, evolving, and analyzing discrete complex systems like ECA and SSS.
A boilerplate for CMS websites developed in Wagtail to save you tens of hours of work.
A Python Library for Pixel Art Conversion.
Clausal Prolog is a Python implementation of Prolog for neurosymbolic AI, combining symbolic reasoning with Python and neural models in the same process without a separate runtime.
Learn AI agents from first principles to production. Zero mandatory dependencies, pure standard Python 3.11+, and no magic frameworks.
Pyronaut is a high-performance polyglot application platform that brings Micronaut features like HTTP routing, dependency injection, configuration, validation, and Java library access to Python via GraalPy.
Agent tooling for generative LEGO models building, built with Astra and Opus 5.5, powered by Jev
New Releases
Polars 2.0 introduces initial out-of-core spill-to-disk support, major performance improvements, first-class SQL, a new Map dtype, and stricter dtype handling. With the SQL and performance work, Polars reports leading DataFusion and DuckDB on TPC-H and TPC-DS benchmarks.
Upcoming Events and Webinars
There will be following talks
The Beautiful Improbability of the Real
Pip install rave
There will be a talk, Writing a Large Language Model From Scratch Using Python: The Beginning Steps.
There will be following talks
The Future of Sovereign AI
Digital Twins and Smart Cities
Cybersecurity for Sovereign AI
Deploying LLMs Locally with Gemma
Our Other Newsletters |
Programmer Weekly - A free weekly newsletter for programmers.
Founder Weekly - A free weekly newsletter for entrepreneurs featuring best curated content, must read articles, how to guides, tips and tricks, resources, events and more.