I think that if you build a solid foundation of your project and can articulate somewhat well what it is you want it to do, then you can expect a pretty good result. I typically limit my prompt to a specific file, often specify the lines and outline some of the logic and add references to other files where necessary. Then, Claude gets just enough context to to do what I want it to do.

Another trick I learnt is you can ask Claude to ask you comprehensive questions for clarification. Usually, it will then offer you a choice of 3 options per question that it might have and you can steer it towards the right implementation.

For professional work, I like to offload some annoying bug fixes to Claude and let it figure it out. Then, perusing the changes to make sure nothing silly is being added to the codebase. Sometimes it works pretty well. Other times, for complicated things I need to step in and manually patch. Overall, I'm a lot less stressed about meeting deadlines and being productive at work. On the other hand, I'm more stressed about losing my employment due to AI hype and its effectiveness.

For my side projects, I do like to offload the tedious steps like setup, scaffolding or updating tasks to Claude. Things like weird build or compile errors that I usually would have to spend hours Googling to figure out I can get sorted in a matter of minutes. Other than that, I still like to write my own code as I enjoy doing it.

Overall, I like it as a tool to assist in my work. What I dislike is how much peddling is being done to shove AI into everything.

I am also confused as to how this is relevant to HN.

Apparently more than 100 registerd users with voting rights thought it interesting and few to none felt it flag worthy.

So, much the same bar was cleared as every other article that makes the "interesting to HN community" grade.

Rack that up to more Trekkie-adjacent and metal-heads than you might have expected.

This may sound stupid, but I wonder if using GPUI for a web browser could have some performance benefits...

I don't get the impression they care that much about performance. Besides, it would limit the number of platforms it could run on if it requires a recent GPU.

Using Rust means they’ve limited the number of platforms already… so I doubt they care.

How so? What platforms were excluded by the use of Rust?

Looks cool and congrats on the $1k MRR! Is the app built with electron?

Thanks!

Yep, it’s built with Electron. Performance has been a big focus from day one, and it’s been really performant in all of my testing so far. The goal was a proper desktop-first experience with local performance and direct database access, rather than trying to force it into a web app. Although I do have plans to offer a self-hosted version as well.

I really like custom elements, I wish they were more popular.

I suppose drawing rectangles as GeoJSON on a map is the easy part. How would you build the parking spots allocation within a selected area?

How can you tell?

The "Implemented Features" section in the readme has every heading denoted with an emoji, that's something I've seen pretty much only from LLMs

I actually did that :(

Considering the comments, how much of it was vibe coded? Since you said you did use AI help?

not much, mainly used it to describe how certain iOS 26 implementations work compared to previous iOS versions.

Good if you didn't just vibe code, at least for maintainability. Your app looks nice btw, I hope you'll keep working on it!

Consider my hypothesis disproven!

I'm an iOS developer, usually it will take me over 500 commits before my app gets working. The app's initial commit seems complete already.

Did use an AI coding assistant yes as this is the first app I've built but I also didn't publish it on Github until the app was near 100% working first

This is great, thanks for sharing! I've been thinking about improving error handling in my liveview app and this might be a nice way to start.

Unrelated, but v8.dev website is incredibly fast! Thought it would be content preloading with link hovering but no. Refreshing

Looking at the site it seems to be a (static?) HTML and shared "main.css" and "main.js" files. Both files can be cached by the browser, so it only needs to download a few KB of compressed HTML for each page. I don't think we would notice much of a difference in the navigation from one page to another if they used content preloading

It's how we used to make websites before SPA, and it's refreshing to see that it still makes a noticeable difference even on today's powerful CPUs and high speed networks.

It's also very simple and free of ads or any other extraneous clutter, a bit like hacker news, which is also fast.

There's probably a lesson in there somewhere.

Speaking of websites, does anyone know when will this land in Node? Node 24 has v8 13.6, and this is 13.8... I mean, this seems like a too big of a performance upgrade to just put it in the next release, especially since Node 24 will be the next LTS version.

There is a PR open since June. It's currently blocked by a MacOS CI issue.

https://github.com/nodejs/node/pull/58491