The year is coming to an end, and looking back I had a lot of fun building quirky, fun, and insightful visualizations. Agentic coding agents have been hugely empowering for me. In 2024 it was “just” autocomplete on steroids. Now it feels more like conducting, directing the agents to go where I please. Often they’re impressive; sometimes they fail at the simplest tasks.

What worked particularly well here was a test-driven (TDD) approach. I asked Claude Code to focus only on the algorithm for selecting the next Tetromino (the Tetris blocks), starting with tests. After that, things went smoothly, though it still took many follow-up prompts (around 100) to reach the final result.

Just letting you know, I think there might be something wrong with the code.

First time I ran it, just off the basic link, it eventually finished with a time of 36:42 (ran at 15:16)

Then I tried turning up the speed and it seemed to make numbers at first but erased its own progress with blank blocks, ending in gibberish.

Maybe another 100 prompts? haha.

I liked it, mainly at 10x speed. The way it built a 3 was totally cheating! A brick just stopped falling without anything below!

Fun idea Watching the digits assemble feels oddly satisfying, especially at normal speed