System Architecture

Weeknotes 377 - AI is becoming an exoskeleton for thinking—but exoskeletons only work if there's still someone inside steering. Apps replacing agents require rethinking. And more from the news.

Introducing local_websocket

Why JSON-LD's purely declarative semantics and monotonic schema expansion make it reliable infrastructure for composable systems.

An interpreted language for Petri net token models with guards, invariants, and data state semantics.

A Practitioner's Guide to Modeling with pflow

The blog's models, concepts, and toolchain have been organized into a book-length guide at book.pflow.xyz.

Iran is slowly emerging from the most severe communications blackout in its history and one of the longest in the world. Triggered as part of January’s government crackdown against citizen protests nationwide, the regime implemented an internet shutdown that transcends the standard definition of internet censorship. This was not merely blocking social media or foreign websites; it was a total communications shutdown. Unlike previous Iranian internet shutdowns where Iran’s domestic intranet—the National Information Network (NIN)—remained functional to keep the banking and administrative sectors running, the 2026 blackout ...

My first post made the case for what a semantic layer can bring to the modern enterprise: a single source of truth accessible to everyone who needs it—BI teams

The rewrite engine is an absolutely brilliant invention by John Brant and Don Roberts, introduced with the Refactoring Browser (see  “A Refa...

This article explores the latest batch of optimizations we did to set-theoretic types and their representation

Exploring why GNU Emacs embeds a Lisp interpreter in C -- from TECO marcos to Greenspun's Tenth Rule, with architecture comparisons to Neovim and VSCode

The most powerful AI software development platform with the industry-leading context engine.

An interactive exploration of quadtrees. Start with brute-force spatial search, build recursive space partitioning step by step, then use it for point lookups, range queries, nearest-neighbor search, collision detection, and image compression.

Scaling agentic AI is forcing a shift from external policy to runtime architecture.

An RSS reader that doesn't count. What happens when you stop treating your feeds like an inbox and start treating them like a river.

A blog about software craft and quality

economical theory that increased efficiency in use of a resource also tends to increase consumption of that resource

In this post, I will show you how I integrate Brotli level 11 compression directly into my 11ty build process to squeeze every possible byte out of my blog’s HTML.

How to guide and encourage AI throughout your organisation within practical guardrails that encourage safe, secure and lawful use of AI. Part 1

Last year, I announced that Canonical had begun supporting the development of upki, a project that will bring browser-grade Public Key Infrastructure (PKI) to Linux. Since then, development has been moving at pace thanks to the tireless work of Dirkjan and Joe. In this post, I’ll explore the progress we’ve made, how you can try an early version, and where we’re going next. Architecture & Progress As a reminder, upki’s primary goal is to provide a reliable, privacy-preserving, and efficient cer...

Why your company probably chose the wrong coordination model

Delving deeper, after selecting technology from the myriad available, you’ll inevitably confront the need to manage intermediate levels. This is particularly true when handling unstructured data, which necessitates transformation into a structured format. Orchestrators and cloud-computing frameworks play a crucial role in this process, ensuring efficient data manipulation across different systems and formats. In the following chapters, I’ll elucidate their role in completing the full architectural picture.

From compression tricks to multi-agent teamwork, here's what makes them tick.

StrongDM's field notes on non-interactive agentic development: specs + scenarios, validation harnesses, feedback loops, and the supporting components.

Inside the StrongDM Software Factory: a new approach to building software with AI using agent-driven execution, scenario-based validation, and digital twin systems—where validation replaces code review.