In Part 1 and Part 2, I reported some peculiar behavior for quantized Qwen3.5-2B models. In Part 3 I moved to Qwen3.5-35B-A3B, and in Part…

This post documents a big refactor in Ormolu that resolved all comments-related issues.

The Bear notes app recently added a CLI for interacting with your notes. Let's use it to build a script for creating backups of my notes! Prerequisites Here is the stack I am using: Bear (duh) Fish…

If you write code, you want your API to be clean, concise, and beautiful. Less explicit, less verbose. Agents may be the opposite. Give them a full explicit API, and don't leave anything hidden.

How I run a coding LLM locally on a single RTX 3090 Ti — the model journey from Claude-distilled to a dense 27B coder, why MoE beat dense for speed, why I stayed on llama.cpp over vLLM on Ampere, and…

. [What Replaces Git When Agents Write the Code](what-replaces-git-when-agents-write-the-code-cover...

Lately I have found myself dealing with more and more terminal sessions on my Mac. These were not just sessions where I had specific areas of work open. These were long running processes. Local…

I'm working on extracting NeoDoc from my static site generator to pitch the format as an option for the Atmosphere of AT Proto. Part of that process is making an explicit spec for HTML output. The…

I’ve been using Opus 5 as my main coding for the last few days 1, and I noticed something odd: the code comments stopped using em-dashes; everything is double hyphens (--) now. Instead of going off a…

FeedWrite stores a queue of text + image pairs and posts them one at a time in sequence, at a interval you choose.

In the previous article, we looked at std::indirect — C++26’s answer to the “I need a heap-allocated member with value semantics” problem. indirect<T> always stores exactly a T. But what if you…

I’m releasing simultaneous major updates to My Calendar and My Calendar Pro today! The My Calendar release is particularly exciting, adding a couple of commonly requested features. My Calendar…

Last edited on August 18th, 2026. Since the end of 2025 that automatically generating useful code can be unlocked by giving large language models (LLMs) the tools and resources to work in loops,…

Based on the requirements of two of my WIPs, I needed robust character and conversation systems. Normally these would be game-specific since the story dictates how dialogue is presented and executed.…

This post is part 3 (and probably the final part) of a multi-part series:part 0 (an intro and brief description of KCL)part 1: unitspart 2: program memorypart 3: look and feelThis post isn't about a…

After reading What software do you use daily in 2026? on Lobster, I learned about Chafa. Described as: The premier UX of the 21st century just got a little better: With chafa, you can now view very,…

I've often been asked what makes a junior developer into a senior developer. Often it's a junior developer asking me the question. And often my answer is system design skills.The best…

Программный код, на котором работал миссия Аполлона-11 в GitHub На человеческий язык этот код переводится как “Это один маленький шаг для человека, но гигантский скачок для всего человечества”

This post describes the semantics of the lambda calculus.

Thanks for visiting my blog! This post is an introduction of sorts, explaining what I'm doing, why I'm doing it and where I'm starting from. So without further ado, what am I doing? I'm going to…

mac·ro·cosm noun \ˈma-krə-ˌkä-zəm\ 1: the great world : universe2: a complex that is a large-scale reproduction of one of its constituents in less vague terms: kosm is an abstract…

Shifting focus towards Android After having found that Google Play didn't offer as much in the way of actual sound synthesizers (rather than sample-based apps!) as were available to iOS users, I took…

The Apollo 11 mission original source code on GitHub This code was behind "That's one small step for man, one giant leap for mankind"

Can a five-year-old MacBook challenge Claude Opus 4.6 at code review? Google Gemma 4 takes on a real 1,322-line merge request. Discover the surprising results and two small settings that turned it…

This is part of a pair of articles on the impacts of agent driven development - you can read thoughts on backlog here. Excitement continues around coding assistants and we’re really now at the…

TL;DR Link to heading Library lending apps hand out .acsm files, which normally need Adobe Digital Editions to turn into a readable epub. ADE doesn’t run on Apple Silicon macOS. I built…

Agentic coding tools have become an extremely useful runtime beyond implementing code. Cursor and Claude Code proved that an LLM in a coding-optimized harness can not only generate useful code…

I've just released biff.run , a very light clj -based task runner; and biff.tasks , a bunch of default CLI tasks that can be used in tools.deps projects. The tasks are mostly the same as what I've…

I’ve been building a website theme. It’s called Popular, after the song Pop!ular by Darren Hayes. A theme whose whole job is helping small communities put on a good show deserved a name…

This is a test post to look at how easy it is to use Markdown: int main ( void ) { return 0 ; }

I demonstrate how to provision a new Ubuntu server on DigitalOcean and deploy a hello-world Clojure web app to it with biff.tasks. biff.tasks

In the interest of honesty — and mild self‑dragging — I should add a post‑script to yesterday’s rant. Despite all the swearing, the dramatic rollback, and the heartfelt threats to abandon Zola…

In a previous post, “.NET 11 Preview - Using Zstandard Decompression With A HttpClient”, we looked at how to configure a HttpClient to transparently decompress traffic compressed using the Zstandard…

If you don't know my app On This Day Rewind yet, give it a try — it shows you the photos you took on this exact day in past years. As much as I would like to write a more detailed post for the new…

Every team needs at least three people. The code person, the data person, and the manager. As a solo dev, I’m all three. And up until now, I had completely forgotten/ignored to be the manager. There…

This is a follow-on to my previous post about what it looks like when a developer tests. Same project, next … More Tests Live on the Edges →

Every program has infinite bugs. You can always invent a new expectation the code fails, and most of those failures will never cost anyone a dollar. Agent reviewers do not find them all, and they do…

I know that LLM or Large Language Models are designed to mimic what people write, but still from time to time it surprises me and make me chuckle. I am working on small tool that runs agents in…

So, your runbook repo is now shared. Three of you commit to it, the playbooks get used in real change windows, and last Tuesday somebody pushed broken YAML straight to main without anyone noticing.…

Every project has way more things to do than resources allotted. That is a given for pretty much any human endeavor. Figuring out what to work on is the interesting part. When it comes to software…

How to give coding agents GitHub access without long-lived PATs in chat: short-lived App tokens, a deterministic credential broker, personal vs bot identity, and fail-closed landing.

If your .NET pods are each using 1–2 GB of RAM, you may be leaving a lot of pod density — and money — on the table. Here are the first things I would look at. 1. Set Realistic Kubernetes Memory…

Git Links is an unbelievably simple format for permalinks in the code. Imagine a file in your repo refers to mydoc.md:32. You specified the line because files can be large. But an edit to mydoc.md…

I can now hand much of the mechanical work to an AI coding tool, then spend my time fixing the interesting parts. This has been my experience with vibe coding too. On the surface, my coding agents…

This is Part 2 of The Machine in the Lab, a series about using an autonomous research program built from large language models (LLMs) to build SetScope, a live Goose song guesser. In the first…

In the latest installment of why we can’t have nice things, I noticed that my markdown viewer of choice, md-tui didn’t support mouse scroll wheel handling… or did it? Running md-tui inside of Tmux,…

Visual Studio Code was a reluctant adoption ages ago when I first tried it. Microsoft makes terrible software but VS Code was an exception. It seemed like they were warming up to opensource. Since…

When the Agent Ports Too Faithfully Vibe coding with an agent is great at producing working code quickly. It is also great at producing the wrong kind of fidelity: matching yesterday’s prototype…