piro0919

AI is a means, not an end. What should matter is whether a project is open source, not what tools were used to build it. I build my Mac apps with Claude Code and release them under the MIT license, and I don't see how that makes them any less open source.

Around 2010–2012 I made videos for Vocaloid songs, with the art coming from separate illustrators. Many of them wouldn't send their illustrations split into layers, which made animating them a pain. And even though it was a music video and I asked for more drawings, some just wouldn't listen and sent a single piece with way too much effort poured into it.

Around 2010–2012, I made videos for Vocaloid producers' songs in Japan. It was almost entirely unpaid. What I got out of it was skills.

I'm building a fan site for a Japanese baseball team on my own. Besides news and videos, it pulls player stats from more than ten sites, from the league's official site to Wikipedia. Writing that many scrapers by hand would have been impossible for me, so I wouldn't have even tried without Claude Code.

For me it's case by case. When I notice a rule in CLAUDE.md has gone stale, I just tell Claude Code to delete it.

I built a music player by talking to Claude Code in Japanese, and I can barely remember a time when words failed and I had to fall back to code. For example, I asked it to "keep the sidebar's shape, but add margin on the top, left and bottom so it looks like it's floating a bit — you know what I mean?" and it got it right on the first try.

I run some of my own tasks with Claude Code on a schedule. When a task involves looking into websites, I'd want it to check with me first if it might put load on someone else's site.

I built a web app with Claude Code without reading the code it wrote. I only asked for things like "add outer margins to the menu and header, with a transparent blurred background," and it came out the way I imagined.

Around 2010–2012, I made videos for Vocaloid songs. Back then, whether the voice sounded robotic or not never even crossed my mind. If it sounded just like a human, that would be fine too.

I'm building a Vocaloid music player with Claude Code. It worked on day one, but the colors were still the pale purple from another app of mine I had started from. I changed them to Miku's teal. That was the part I had to decide myself.

I've rarely wanted Android apps on my Mac. Where I've wanted them is on my iPhone: an app that forwards incoming SMS to Gmail, and a free VPN.

I don't use tmux. I run Claude Code in Cursor windows, and each window title shows the folder name, so that's how I tell them apart. It rarely gets confusing, and an app I built for myself made it a lot easier.

I don't read the code Claude Code writes at all. And yes, it does cause problems. Recently, in a project that was supposed to use shadcn, it quietly stopped using it partway through.

Even with Claude Code writing the code, I still decide which features go in, and how the screens look and what they say. Building with a specific user in mind is something I can't leave to AI yet.

Honestly, I don't think about it much when I ship things, but at the very least the landing page should explain clearly what it is.

I use Claude Code for code review, and honestly it catches more than human reviewers have in my experience. It's been a big help.

When I introduce my apps, I say I made them with Claude Code. I'd say both: Claude wrote the code, but the ideas and the decisions were mine.

Code from Claude Code that looks like it works but is actually broken happens to me all the time. I usually find out when I debug it locally, or when Sentry sends me an error alert.

I'm Japanese and can't write English, so I think in Japanese and have AI translate it. If that makes my comments read like AI, I guess there's nothing I can do about it.

I half agree.

I build apps with Claude Code, and I test them by having the AI click through the UI with Playwright. That's enough to check that things work as specified and nothing is broken.

But I think the purpose is different when an AI tries it and when a person tries it. To put it in extreme terms, AI is for UI and people are for UX. People find UI problems too, though: in my music player, songs got blocked from playing in Safari on a real iPad, and I only found it by using it myself. The things found in this article, like the jokes that didn't land or not knowing what the tool even does, are on the side only people can find.

(I wrote this in Japanese and used AI to translate it.)

I disagree. I think the result, the quality, is what matters, not how it was made.

Around 2010 I made music videos for a Vocaloid producer. Back then, people sometimes said Hatsune Miku songs were worth less because a machine was singing. I didn't think so, and in the end the songs were accepted because they were good.

Of course, Vocaloid songs were still written by people, which is different from AI making the whole song. Even so, I think what matters to listeners is the quality.

(I wrote this in Japanese and used AI to translate it.)