Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
Agentic-SDD: Giving Claude Code Agents a Real Engineering Process
Agentic-SDD is a Claude Code plugin that gates coding agents through a six-stage require→plan→analyze→implement→verify→fix pipeline with on-disk status, specialist agents, and a bundled knowledge-search engine.
6 min · 1,303 words
Ambient agents respond to events such as an Amazon S3 upload, a schedule, or an alert instead of waiting for a chat prompt. This post walks through building framework-agnostic ambient agents on Amazon Bedrock AgentCore using Amazon SQS, AWS Lambda, and Amazon DynamoDB, with a single ask_human tool and a Jobs page for human-in-the-loop review.
24 min · 5,504 words
What Makes LLM Tokenization Slow?Exploring the performance of byte-pair encoding by optimizing a GPT-2 tokenizer.
Andrew Healey dissects GPT-2’s reference BPE tokenizer, measures what makes tokenization slow, and shows concrete optimizations on the hot path of LLM products.
11 min · 2,547 words
Functional ultrasound imaging from scratch
Functional ultrasound imaging (fUSI) is a brain imaging modality that is just starting to be demonstrated in human studies. I’m sure many will be familiar with structural ultrasound imaging, which forms static images using ultrasound, as in a sonogram1. Functional ultrasound instead captures movies of brain tissue very rapidly, tracking subtle changes caused by the blood flow that follows neural activity.
11 min · 2,451 words
A Spring AI tutorial on hardening agents for production: guardrails, evaluation loops, observability, tool authorization, and human-in-the-loop approval beyond a basic MCP-connected demo.
17 min · 3,953 words
Engineer-Led, AI-Assisted: A Practical Workflow for Building Software
Nathan Pickard describes an engineer-led, AI-assisted software workflow covering planning, development, testing, code review, and context management—arguing AI depends on engineering judgment rather than replacing it.
20 min · 4,529 words
Reverse Engineering NovaLogic's Comanche Terrain Maps
A detailed reverse-engineering walkthrough of Comanche: Maximum Overkill MS-DOS terrain/map files—how NovaLogic packed heightmaps and related data, and how to reconstruct the landscapes.
16 min · 3,693 words
How to Build a Reliable AI Assistant with the Claude API
A freeCodeCamp tutorial building ShopHelper with the Claude API: conversation history, tools, multi-block responses, workflow patterns, and evaluating whether prompt changes actually help.
9 min · 2,132 words
Using any C++ library in Godot
Using any C++ library in Godot Godot has become one of the most popular game engines of the last few years. It is free, open source under the MIT license, and small enough to download and start using in minutes.
10 min · 2,353 words
Build an LLM Tokenizer and Attention from Scratch in TypeScript
SitePoint tutorial that implements BPE tokenization, cosine similarity vector search, and scaled dot-product attention in TypeScript—exposing LLM primitives as ordinary readable code.
25 min · 5,827 words
Automating eval design and hillclimbing with Claude
Lance Martin (claude.dev) explains principles for production-like evals with held-out sets, then shows how the claude-api skill’s build-eval and hillclimb commands automate design and overfitting-aware improvement—including cost and performance case studies.
2 min · 462 words
Phyllotaxis: An audio-reactive LED display
A build write-up of an audio-reactive LED sculpture that maps Fibonacci phyllotaxis onto a physical light array—math, fabrication, and real-time reactive firmware.
7 min · 1,644 words
Standard deviation or standard error? What each one measures, and which one belongs in your paper
Selçuk Korkmaz clarifies SD vs SE with interactive figures: use SD to describe sample spread, and report estimates with confidence intervals rather than bare standard errors.
16 min · 3,717 words
Visualising data with a hexagonal heatmap in Swift Charts
Explore alternatives to rectangular heatmap cells for visualising spatial data along coastlines and irregular boundaries.
9 min · 2,015 words
Replacing the old battery on rechargeable bike lights
Julia Evans walks through repairing decade-old rechargeable bike lights by replacing the battery at a makerspace—opening the case, identifying the cell, soldering a replacement, and what she learned as a non-electronics expert.
4 min · 1,011 words
Postgres AT TIME ZONE 'UTC' does NOT do what you think it does
A practical Postgres footgun: AT TIME ZONE 'UTC' converts timestamptz to timestamp without time zone—so month math and equality break unless you apply AT TIME ZONE twice.
2 min · 538 words
Teaching GPU programming in p5.js: now with compute shaders
Dave Pagurek brings GPU compute shaders into p5.js for teaching—why graphics education needs modern GPU abstractions, and how students can write compute shaders in a friendly creative-coding environment.
17 min · 3,974 words
Turn GLM-5.3-Flash into a Jev-like System One model
Johannes Hötter walks through turning GLM-5.3-Flash into a fast Jev-like “System One” decision model—typed options with probabilities in a single forward pass, matching Jev’s accuracy and speed.
11 min · 2,426 words
Modern Object Pascal Introduction for Programmers
A practical introduction to Modern Object Pascal for programmers coming from other languages: units, classes, generics, memory management, exceptions, and more—maintained by Michalis Kamburelis and the Castle Game Engine project.
75 min · 17,153 words
Teaching a World Model to Play Pokémon
Training a JEPA-style LeWorldModel on Pokémon Red screenshots and button presses, then using CEM latent planning to select a starter—why latent collapse, SIGReg, and rollout fine-tuning mattered, and 52/100 plans succeeding after tuning.
2 min · 574 words