How to Build an NYC Taxi Data Dashboard with Claude Dashboards, DuckDB and d3
Step-by-step guide: turn the 3.7M-row NYC TLC yellow taxi Parquet file for January 2026 into an interactive Claude dashboard using a very simple and english-like prompt
Anthropic 5.5 models – DOs and DONTs
Short answer for the Anthropic 5.5 models: let Opus 5.5 plan, let Sonnet 5.5 do the hands-on work, keep Medium as your default effort, and skip Max / xHigh unless
Opus 5 is dead, Long Live to Opus 5.5
After months and months of desperation from users – wait, no: Claude Opus 5 was launched just on July 24 – which means that not even two months (!) have
The Navier-Stokes problem – another view on scientific funding
A bit late to the Navier-Stokes “party”, but as the absolute drama was unravelling, I was thinking about it in another perspective: what does this amount of compute mean and how can
LLMs for scientific research
In one of my last posts, I tried to calculate an estimate of how many years of funding for a small research group was equivalent to what OpenAI spent to
Claude Code – what is an Artifact?
In the last post (Is your ClaudeCode context actually optimised?), I have mentioned the Artifact tool from Claude Code and how it was bloating the System tools. I stand by
Is your Claude Code context actually optimised?
3 months using Claude. One of the first things I did was researching: – What is the base context when I just send one simple “Hey Claude”? – Why to
AI Generated Content Watermark Explained
Have 30 minutes? Then watch this video with a fantastic and clear explanation of the Watermarking in AI Generated content from Computerphile Not only explains what watermarking is, but also
Controlling one agent is easy (Part 2/2)
But what about 5? And 20? And 100? 🤔 In the first part of this mini-series, I gave some tips on how to not burn through the entire 5-hour limit