But that would apply to any app that deals with files like this one does.
This one is open source and we can run some code analysis on it, compile locally, etc. I am not well versed in security checks but I guess you get the idea.
This situation reminds me of the Visual Basic apps in the nineties. It was possible to write good quality software with VB and some people did. Nonetheless the majority was dross and the fact that an app was written in VB was a strong indicator of poor quality.
Agents are really good at writing unit tests, but only if you clarify exactly WHAT should be tested. Otherwise they generate slop which passes, but don't catch any bugs nor regression.
Don't know why this comment was flagged, this is exactly what is going on here. AI is famous for these purple gradients and the website reeks "vibe-coded", this is 100% how Claude makes websites, including the already mentioned gradients, emojis, style, etc.
And of course every commit is Co-Authored by Claude Code with excessive commit descriptions also created by Claude. Is this really something we want to see on Hacker News? I wouldn't trust such an application.
Nothing against AI coding but letting AI take the wheel 100% of the time and not even mention it (like he coded it himself) is very dishonest.
The CLAUDE.md file is right there, so they are probably using agentic coding.
But why does it matter? Does the app not work? I don't have a Mac, can't check.
Counter question: How do you know it works?
A file manager better be rock solid, I don't want a bug to delete any files or do other shenanigans.
That is a valid question.
But that would apply to any app that deals with files like this one does.
This one is open source and we can run some code analysis on it, compile locally, etc. I am not well versed in security checks but I guess you get the idea.
You can see that it is[0].
[0] https://github.com/rajivm1991/DroidDock/commits/main/
Especially the 'design' which couldn't be less Mac-native if it tried
Ugh. This term can't die fast enough.
What term would you suggest instead? Agent-assisted slop?
Not bad; anything that actually mentions "AI" would be better than "vibe." That doesn't tell you anything.
Of course it is.
As usual, it is low quality and has zero tests.
This situation reminds me of the Visual Basic apps in the nineties. It was possible to write good quality software with VB and some people did. Nonetheless the majority was dross and the fact that an app was written in VB was a strong indicator of poor quality.
agents should agentically create high quality unit tests
Agents are really good at writing unit tests, but only if you clarify exactly WHAT should be tested. Otherwise they generate slop which passes, but don't catch any bugs nor regression.
Why do you say that?
Because it is true. Claude did almost all of it.
It appears very "vibe coded", the application makes use of the stereotypical AI gradient in the top bar of the application.
The README.md uses a ton of Emojis in the feature Setting.
100% AI.
Don't know why this comment was flagged, this is exactly what is going on here. AI is famous for these purple gradients and the website reeks "vibe-coded", this is 100% how Claude makes websites, including the already mentioned gradients, emojis, style, etc.
And of course every commit is Co-Authored by Claude Code with excessive commit descriptions also created by Claude. Is this really something we want to see on Hacker News? I wouldn't trust such an application.
Nothing against AI coding but letting AI take the wheel 100% of the time and not even mention it (like he coded it himself) is very dishonest.