AI writing is terrible. Except for lab notes.

I truly dislike LLM-generated writing. The quality of the models’ output has vastly improved over the years, but the result stays the same. Writing is deeply human. Putting our thoughts on paper, with all the struggle and the rework, is how we learn. It’s how we get better as individuals and as a species. So, every autogenerated newsletter and LinkedIn reply feels like an affront to humanity. And it’s getting worse by the day. ...

February 14, 2026

Jean-Claude: Syncing Claude Code Config Across Machines

A companion CLI for syncing your Claude Code configuration — CLAUDE.md, settings, hooks, and skills — across all your machines using Git.

February 14, 2026

The Great AI Overengineering Olympics

4 AI agents build Hello World in Java with TDD, DDD, BDD and hexagonal architecture. The results are magnificent enterprise-grade atrocities.

February 12, 2026

Why Inference Will Get Dirt Cheap Fast

We are no longer afraid that AI doesn’t work. Every day, we see it with our own eyes. No, the most persistent fear is that this honeymoon period is just temporary. Right now, in a fight for market share, the token dealers sell us inference at a loss, hoping to hook us. Once the market is saturated, they’ll start charging us the real deal. And at that point, all these cool new toys become prohibitively expensive. ...

February 7, 2026

Cauli Quest: A Browser Platformer Built with AI

A tiny browser platformer where a cauliflower risks it all for carrot-based glory. Built with vanilla JavaScript and AI-generated assets.

February 6, 2026

Local Agent Bench: Testing 21 Open-Weight Models on Tool Calling

Benchmarking small open-weight models on a $1,000 laptop to see which ones know when to use tools — and when not to.

February 6, 2026

Living with an artificial psychopath

If you’ve spent any time on social media last week, you would have to work hard to avoid the hype surrounding Clawdbot / Molty / OpenClaw. In the course of one single week, the artificial virtual assistant changed name twice and caused a spike in Apple’s Mac Mini sales. The idea behind OpenClaw is simple but fascinating: give an AI agent access to your desktop and delegate work to it. Ask it to free up your Friday afternoon, and it’ll reschedule the meetings to another day. It can organise a Doodle and then book a table for that dinner with busy friends. It’s the trusty personal assistant that only top managers could afford. Until now. ...

January 31, 2026

Burning LEDs vs Picking Flowers

When I volunteered at Coderdojo, we taught kids to code. In 80% of the cases, that meant building video games with Scratch. While that is cool, the coaches got really excited when a child wanted to do something more exotic. Unity, robotics, or even just JavaScript. Finally, something they could also sink their tech teeth into! I remember a girl who wanted to build a robot. The volunteers got fired up. We brought Arduino kits and soldering material. We wanted children to short-circuit and burn LED lights to teach them how electronics worked. Quite a few of them did. ...

January 26, 2026

Claude After Dark: Creating Retro Sprite Games with AI

Using Claude Code to drive Gemini’s image generation for consistent animated sprites, recreating the classic Flying Toasters screensaver.

January 17, 2026

Why CTOs Love Coding Agents and Engineers Don’t

Like all online topics these days, coding agents are divisive. Some zealots predicted the end of the software developer around 8 months ago. Others consider those agents to be as useless and atrocious as Microsoft Teams. Most of us are somewhere on the spectrum between those extremes. Over the last few weeks, I have been taking advantage of Anthropic’s holiday offerings and have been spending an unhealthy amount of time with Claude Code. It’s fun, it’s productive, but it also teaches us something about the state of the art. Or rather: about the perceived state of the art. ...

January 13, 2026