Ai

Someone recently asked me how many tokens I used after my Qwen 3.8 27B experiment. I had no idea, nor did I know how to verify it. This post explains how to count tokens, how prompt caching works, and tools to help analyze sessions.

Qwen3.8 27B just got released, a week after Meta's Muse Glimmer 30B. Alibaba advertises benchmark scores on par with Opus 4.6. This post compares Qwen3.8 27B with Muse Glimmer 30B, Sonnet 5 and Opus 5.

I compared Meta's new Muse Glimmer 30B against Gemma4 26B-A4B and Qwen3.6 35B-A3B on a simple agentic search task, running locally on an M3 Max. Muse Glimmer tries harder than any of them, and that is exactly what makes it unusable.

I have been using Claude Code personally and professionally for about six months. This post explains how I set it up, including plugins, tools, MCPs and CLI.

Producing app store screenshots for OnTrack took me around 18 hours in GIMP. I attempted to automate it and decided to build a new web app for my needs: simple, free and multilingual. Building a new tool is also a good excuse to experiment with AI tools, so this post covers how my usage of Claude Code evolved.