logoalt Hacker News

mr_mitm • today at 8:13 AM • 4 replies • view on HN

Same here.

Creating a polished, usable app is just so much work, and so much of it isn't fun at all (to me). There are a few key parts that are fun, but building an intuitive UI, logging, error handling, documentation, packaging, versioning, containerization, etc. is so tedious.

I'm bewildered when I read posts by the naysayers, because I'm sitting here building polished apps in a fraction of the time, and they work. At least much better than what I was able to build over a couple of weekends. They provide real value to me. And I'm still having fun building them.

I now vibe coded three apps, two of them web apps, in Rust, and I couldn't write a "Hello World" in Rust if you held a gun to my head. They look beautiful, are snappy, and it being Rust gives me a lot of confidence in its correctness (feel free to disagree here).

Of course I wouldn't vibe code in a serious production project, but I'd still use an AI agent, except I'd make sure I understand every line it puts out.

Replies

latexr • today at 10:02 AM

> Creating a polished, usable app is just so much work, and so much of it isn't fun at all (to me).

Then don’t do it. No one is forcing you. Are you also going to complain that building airplanes and ensuring food safety are too much work and not fun for you? Not everything needs to be or should be dumbed down to appeal to lowest common denominator.

Alternatively, go work at a company where you’re part of a team and other people do what you do not enjoy.

> I'm sitting here building polished apps in a fraction of the time

No, no you are not, guaranteed. “Polishing” means caring about every detail to make it perfect. If you’re letting the LLM make most of it, by definition it’s not polished.

➕ show 1 reply
exceptione • today at 9:08 AM

I can understand you don't want to spend effort for throwaway code.

  >  in a serious production project, but I'd still use an AI agent, except I'd make sure I understand every line it puts out.
That isn't going to cut it. You need to understand the problem domain, have a deep design taste to weigh current and future demands, form a conceptually coherent solution, formalize it to code, then feed back from the beginning. There is no prompt giving your AI those capabilities. You end up with mediocre solutions if you settle for understanding every line it spits out. To be fair, many programmers don't have those capabilities either, so it also a question of quality expectations.

I believe you can use LLMs as advanced search and as a generator for boilerplate. People liking it easy are also being easy with quality attributes, so anyone should be self aware where they are on that spectrum.

risyachka • today at 8:33 AM

>> so much of it isn't fun at all

thats why it was valuable.

All things worth doing are hard.

➕ show 1 reply
inferiorhuman • today at 8:38 AM

  Of course I wouldn't vibe code in a serious production project, but I'd
  still use an AI agent, except I'd make sure I understand every line it
  puts out.
So you value your ability to churn out insignificant dreck over the ability of others to use the internet? Because that's the choice you're making. All of the sites that churn your browser for a few seconds because they're trying to block AI DDoS bots, that's worth your convenience on meaningless projects? The increased blast radius of Cloudflare outages, that's a cost with foisting on to the rest of the internet for your convenience?

Thanks.

➕ show 1 reply