Your CLAUDE.md is not magic
Why I want AI-written code checked independently, with shared standards and room for developers to choose their own workflows.
All writing
Latest first.
How a RuneScape bot helped me get into the tech industry
Why I think a simple RuneScape Twitter bot helped me get into tech, and what its Python and Go versions taught me about unpaid maintenance.
Codex, fix my home server. I want to watch a film.
Using Codex from my MacBook to repair storage imports, update a server endpoint and bring an SSD into Plex, with checked results and real limits.
My desktop became a Proxmox home server
How an unused desktop became a Proxmox server, what broke during the move, and why two MacBooks now cover my everyday computing.
Free compute in October 2026: new backends and temporary VMs
October's free-tier check: Railway SSH VMs, Neon's backend, Vercel Drives, and corrections to Modal, Lightning and Freestyle credits and limits.
What GPT-6.1 Sol chose to build
Three autonomous GPT-6.1 Sol experiments: a fictional record, a sheet-cutting planner and a congestion model, with their review fixes and limits.
What GPT-6 Astra chose to build
What three Astra agents built with room to choose: a string puzzle, a supply-chain experiment and a printable booklet, with their fixes and limits.
Free compute in September 2026: changed limits and corrections
September 2026 free-tier changes: Deno cuts, D1 hard limits, agent sandboxes and corrections to Freestyle, Browserless, Firebase and Axiom.
Local models are actually good now - playing with Qwen3.8-27B
A 4-bit Qwen3.8-27B built a cellular automata workbench locally on my M4 Pro, while 6-bit finished and 8-bit hit OOM.
From Raspberry Pi to OpenMediaVault: a home server I actually use
How Raspberry Pi experiments became a useful home server, and why learning when not to self-host mattered as much as Docker and DNS.
How I use scheduled Codex jobs to build daily reports
How deterministic tools, scheduled Codex jobs and inspectable HTML fit into two daily reporting workflows.
Building a 322 Dota 2 draft solver with an LLM
An LLM helped build a deterministic 322 draft solver. Paired tests covered 120,000 outcomes, with a 27.45% title rate at the largest field budget.
Building and testing a local Mac dictation app with Codex
Building Billie Flow, a local Mac dictation app, with Codex: model timings, installed-app failures, microphone testing and the limits of the release.
Free compute in August 2026: GPU notebooks and agent tools
August 2026 free VPS, databases, GPU notebooks and agent sandboxes, with a dated catalogue of quotas and links to later corrections.
Building my own DJ library tools, from Python to Crate Lynx
How four years of DJ-library friction led from Python scripts to a Go desktop app and the AI-assisted Crate Lynx.
Animating embeddings with PCA and grand tours
An interactive embedding visualisation compares raw, PCA and grand-tour projections, including a basis change that preserves cosine neighbours.
GPT-5.6 in Codex: my first 48 hours of long-running tasks
What GPT-5.6 did across research, audits, a seven-repository migration and one animated pet during my first 48 hours with Codex.
Free compute in July 2026: VPS limits, sandboxes and AI
July 2026 free-tier changes: Oracle VM limits, Deno migration, agent sandboxes and GPU options, with trials separated from recurring allowances.
Prompting GPT-5.6 for deeper coding experiments
What changed between two GPT-5.6 coding-experiment batches when I gave depth more room, changed the prompt and added review.
Local coding models on an M4 Pro: why the agent harness mattered
Five 4-bit local models on a 48 GB M4 Pro, tested with file-output contracts, smoke-test repair loops and Pi, then compared with a hosted Codex setup.
Codex turned a biological wheel debate into a live simulation
A Codex-built biological wheel simulation went from a Discord debate to Cloudflare Pages in about ten minutes. A toy model for discussion, not proof.
Capturing ideas in Linear with voice dictation and AI agents
My workflow uses Linear as an idea ledger, Wispr Flow for dictation and agents to turn rough notes into projects and posts.
LLM Choice: using autonomous coding agents to find project ideas
Using autonomous coding-agent runs as a low-cost way to find ideas, learn unfamiliar topics, and seed later projects.
Mapping my personal website with an interactive link graph
An interactive graph maps my site's pages and outbound links, exposing its biggest link hub and giving a growing personal archive a visible shape.
Fine-tuning a local LLM on Discord messages: where my clone failed
A local MLX LoRA experiment used my Discord messages to imitate my replies. Missing conversation turns limited what fine-tuning and retrieval could fix.
A refurbished office chair stock tracker that saved me over £100
A self-hosted stock tracker watched 85 refurbished chair listings and searches, filtered noisy results, and helped me save over £100 on a Haworth Fern.
Free compute for side projects: June 2026 free-tier guide
A June 2026 snapshot of free VPS hosting, databases, serverless platforms and AI compute, including the limits behind the offers.