I’ve been messing around with ChatGPT, Claude, CodePilot, and Cursor to build generative art pieces and interactive visual effects using JavaScript, Canvas, and SVGs (GitHub here). A few observations:

• I was able to create things far beyond my technical reach at incredible speed.
• The process forced me to clarify exactly what I wanted to build.
• I learned a lot through iteration — from terminology (polyominoes) to technicalities (canvas objects are more performative than SVGs).
• The best part? Ambiguous instructions sometimes led to “happy accidents,” sparking new ideas in ways that felt as unpredictable and exciting as working in the physical world.

‍

‍

Polyominoes

I loved learning about polyonimoes, the siblings of dominoes.

Following Canvas Arrows

Responsive Dot Grid