Python Weekly (Issue 762 September 10 2026)

Author

Rahul Chaudhary
September 10, 2026

Welcome to issue 762 of Python Weekly. Let's get straight to the links this week.

100 CLI tools that separate top engineers from everyone else

AI can write your code. It can't run your stack. Engineers who stay valuable know their terminal cold. The Code built the 100 CLI Tools, 2026 Edition, to showcase the stack top teams run on. Sign up and get the toolkit free, plus a 5-minute daily newsletter to stay 6 months ahead.


Articles, Tutorials and Talks

The author details a weekend project to build a stripped-down, whitespace-sensitive Python interpreter in just 1,024 bytes of C without external libraries or macros. Despite the size constraint, it supports variables, arithmetic, control flow, functions, recursion, indentation, and printing, enough to run FizzBuzz.

On solving the Jane Street Reverse Engineering Challenge

Jane Street’s ASIC reverse-engineering challenge turns a raw chip-layout file into a month-long puzzle about reconstructing how the circuit works. The author uses Python, graph algorithms, Verilog, simulation, and Z3 to recover the design, solve the challenge, and even uncover a real hardware bug.

Part of the Nifty Django features series: Django’s Q() objects allow you to some fancy filtering!

Since 3.11 attribute access has not been a dict lookup, and reading __dict__ once takes the specialized path away for good

The post explains how NumPy’s universal functions (ufuncs) work in C and how their internals were upgraded to support multi-output reductions, enabling the long-requested np.minmax function. It also covers ongoing work on segmented reductions for ragged data and converting existing functions into generalized ufuncs (gufuncs) for subarray processing.

How to Stop Third-Party APIs From Ruining Your Code

The video explains the Facade pattern using a Stripe payment integration to show how it can isolate external APIs from application code. It covers reducing coupling, improving testability, and creating cleaner boundaries that are easier to maintain and evolve.

A technical deep dive into how Cohere’s approach to megakernels delivers 1.58x faster LLM serving on H100 devices.

This video demonstrates how to turn SQLite into a vector database for AI applications using the sqlite-vec and sqlite-ai-vector extensions. It covers setting up both packages in Python with OpenAI embeddings, implementing similarity search and metadata filtering, and benchmarking their performance, memory efficiency, and quantization capabilities.

Structural quality in the age of AI-written Python. What the standard toolchain checks, what it does not, and what a real agent-written repository looks like when you measure it.

TPU Inference Externalization Full Steam Ahead

The article examines Google’s push to make TPUs a first-class platform for external LLM inference, including TorchTPU, vLLM, SGLang, and a rapidly maturing open software stack. Benchmarks TPUv7 Ironwood against NVIDIA B200/B300 and dives into the kernel, memory, networking, and serving optimizations behind its performance-per-dollar gains.

Plugin Architecture Across Service Boundaries: An API Contract with Pydantic & FastAPI - Pt. 1

This articles shows how to build a service-level plugin architecture with Pydantic and FastAPI. I will be using an API contract package that plugin providers implement and our platform validates the contract during registration.


Interesting Projects, Tools, and Libraries

Deploy, use, and share web apps on a server you control. Your apps, data, and infrastructure stay yours.

Autonomous research system for measurable, computer-executable research.

SGLang is a high-performance serving framework for large language models and multimodal models.

Fully typed frontend framework for Python with server events and Alpine.js, inspired by Vue and Livewire.

An out-of-tree GPU backend that lets you write a constrained subset of Numba-style Python kernels and run them on the GPU of Apple-silicon Macs through Metal.

Centralized network visibility and continuous asset discovery. Monitor devices, detect change, and stay aware across distributed networks.

An Initial Step towards a Vision-Language Foundation Model for Autonomous Driving.

A minimal coding agent harness that fits in the prompt.

Delegated authority for AI agents. A Pigeon Pass says what an agent may do.

NPC-Forge is a framework for building conversational agents that run on the CPU without relying on machine learning or LLMs.


Upcoming Events and Webinars

There will be following talks

  • Building TowleVision: A Local-First AI Media Production System in Python

  • Aging in Pandas: Building a Biological Age Clock from Public Health Data

There will be following talks

  • MicroPython: Overview and Examples

  • Kubernetes for Home Cooks

There will be following talks

  • Think Like a Cat: How to Stand Out in the Age of AI Applications

  • lxml for DITA Automation

  • Supporting resiliance in the workforce and beyond

There will be a talk, MCP Toolbox for Databases: Connecting AI Agents to Enterprise Data.

There will be a talk, Models for SARS-CoV-2 health policies: Social Distancing amidst Vaccination and Virus Variants.


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.