HN user

subairui

23 karma

- Website: https://bairui.dev/ - GitHub: https://github.com/pearmini - Instagram: https://www.instagram.com/subairui24/

Posts5
Comments24
View on HN

Hi HN,

I'm the author!

This article grew out of my master's thesis presentation at NYU ITP. While building creative coding tools over the past few years, I started wondering why we still write code when AI can increasingly generate it for us.

My conclusion isn't that coding is becoming less important. It’s that coding is valuable because toolmaking is one of the fastest ways to learn. The article expands on that idea through examples from projects I've built, including Recho.

Happy to answer questions or discuss.

Thanks, joey9prints!

Yes, it's a bit generic right now, and there's definitely room for improvement. I just checked out Paul Richard's work—it looks amazing and highly relevant.

Also, by default, I don't store either the original image or the generated result. The generated image is only saved if you explicitly click "Archive," and you can delete it whenever you'd like.

Thanks, Twosdai! My initial hesitation about adding a database wasn't really about technical complexity, but more about keeping the product and design simple. A few questions came to mind:

- Do we need accounts? If so, would that add too much friction? If not, could we identify users anonymously through a browser-generated ID?

- As more trees are generated, will the visualization become difficult to manage? Perhaps we only display a subset of them?

- How should we handle inappropriate or offensive inputs? I imagine we could use a moderation library or filter.

My main goal is to keep the experience lightweight and focused on the generation itself. That said, if a database would make participation easier and lead to a better overall experience, I'm definitely open to it.

What's your take? I'd love to hear what's on your mind.