Using LLMs to build out the nice-to-haves that I’ve always wanted but never had time for is one of their great use cases. Visualizations are a perfect use case for this because they don’t have to be perfectly architected, maintainable code. Getting to the correct visual output is good enough, and LLMs excel at iterating something until it looks right.

I've been using LLMs to create visualizations for math papers I come across. Prompting "Create a visualization for each segment of this article in the style of a 3 brown 1 blue video using manim." has yielded impressive results.

It helps me digest the content faster and allows me to read more articles than I otherwise would.

This is v cool.

When I did my microcontroller class with lecturer hand drawing an 8-bit computer, the registers, memory, instructions on the white board, it was v cool to understand how things worked under the hood.

Wondered if someone could make more simulations for what was being taught. Teaching is about deciphering a thing into it's components and seeing how they interact. Vibe coded simulations are a great tool for that.