How I AIHow to use Cursor for interactive prototypes, sound design, and data visualization
At a glance
WHAT IT’S REALLY ABOUT
Designers use Cursor to prototype aesthetics, sound, and data visuals
- Elizabeth Lin shows how designers—not just engineers—can use Cursor as a creative partner to explore web aesthetics, rapidly generate styled pages, and iterate by re-prompting or refining specific issues one at a time.
- She demonstrates a workflow for asking Cursor to suggest design movements, applying a chosen style (e.g., brutalist + Y2K), and using checkpoints to avoid getting stuck in unhelpful generations while saving screenshots of good outcomes.
- Lin also builds a functional, sound-enabled digital piano prototype (hard to do in Figma), using simple “Cursor Rules” to scaffold new prototypes quickly and then interrogating the code/libraries just enough to expand possibilities.
- Finally, she refactors an “ugly” finance dashboard into a cleaner data visualization by removing drop shadows, referencing well-designed products and Tufte, and using targeted prompts—then points to real-data prototyping (e.g., Notion-backed apps) as a major advantage of code-based prototyping.
IDEAS WORTH REMEMBERING
5 ideasStart by asking the model what styles it can implement.
Instead of arriving with a fixed aesthetic, Lin prompts Cursor to list design movements and describe their traits, which gives designers usable language (e.g., “neon, glitch, dystopian” for cyberpunk) to steer direction.
Treat generations as high-level drafts, then design-critique and iterate.
Lin prefers starting broad (“redesign in brutalist + Y2K”) and then applying a designer’s judgment to keep surprising wins (like hover effects or typing animations) while trimming excess.
Use Restore Checkpoint aggressively to avoid bad rabbit holes.
When results drift, she rolls back and tries again; because identical prompts can yield different outcomes, restarting is often faster than incremental patching from a flawed base.
Keep prompts short and tackle issues one or two at a time.
She finds “laundry list” prompts lead the model to forget items; targeted steps (remove drop shadows → fix background fill → simplify palette) produce more reliable progress.
Improve “taste” by feeding strong references the model already knows.
Lin shortcuts explanation by referencing products (Robinhood/Cash App/Stripe), practitioners (Edward Tufte), or aspirational standards (“a top designer at Apple would approve”), which nudges layout, grids, and restraint.
WORDS WORTH SAVING
5 quotesWorking with Cursor has really taught me that tools like Cursor can actually be extremely creative.
— Elizabeth Lin
What design aesthetics and movements are you comfortable implementing? List the styles and describe them to me.
— Elizabeth Lin
I always restore the checkpoint when I don't like something.
— Elizabeth Lin
If I give it, like, a laundry list of items, it'll, like, forget to do the last three.
— Elizabeth Lin
The biggest key is to broaden your sources of inspiration... like a K-pop music video... and see what it takes from it.
— Elizabeth Lin
High quality AI-generated summary created from speaker-labeled transcript.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome