Hacker News Reader

Comment by jeffhuys | Hacker News Reader

Honestly feels very vibe-coded [1] [2] and would not really trust my money with something like this. I had to read the code to understand what it actually protects me from, as the README.md (other than telling me it's production-ready, professional, and protects me from so much!) tells me "Supports all major providers: OpenAI, Anthropic, auto-detected from URLs". OpenAI and Anthropic are "all" major providers [3]?

[1] https://github.com/dipampaul17/AgentGuard/blob/51395c36809aa...

[2] https://github.com/dipampaul17/AgentGuard/commit/d49b361d7f3...

[3] https://github.com/dipampaul17/AgentGuard/blob/083ae9896459b...

View on HN

> [1] https://github.com/dipampaul17/AgentGuard/commit/d49b361d7f3...

It's kind of crazy that people use these multi-billion parameter machine learning models to do search/replace of words in text files, rather than the search/replace in their code editor. I wonder what the efficiency difference is, must be 1000x or even 10000x difference?

Don't get me wrong, I use LLMs too, but mostly for things I wouldn't be able to do myself (like isolated math-heavy functions I can't bother to understand the internals of), not for trivial things like changing "test" to "step" across five files.

I love that the commit ends with

> Codebase is now enterprise-ready with professional language throughout

Like "enterprise-ready" is about error messages and using "Examples" instead of "Demo".

I use LLMs, too, and that did give me a chuckle.

This is the future of the field: amateur night that never ends. And this is why I'm looking for a new career at my advanced age.

If you have any tips, feel free to enlighten me. Even though I'm "only" in my 30s I feel the future is uncertain - the original author of this post made one other post that was also clearly vibe-coded, but not many comments seem to point it out. It'll only get worse from here, depending how you look at it of course; hackers will have a WAY easier time as time goes on.

Still figuring it out, but if I do, I promise I'll circle back here and let you know. So far the best idea I have is to stay in software and ideally my current job for the moment, but phone it in, while I retrain as something else at night.

It's incredible how emoji overuse has become a giant red flag for AI over-/abuse.

Well, that, and stuff like the commit descriptions.

> Polish README: Remove downloads badge and clean up styling

> - Removed downloads badge ___as requested___

The AI's idea of developing a startup is eerily reminiscent of a hacking scene in CSI

> Close first customers at $99/month

> The foundation is bulletproof. Time to execute the 24-hour revenue sprint.

Comedy gold. This is one of those times where i cant figure out if the author is in on the joke, or if they're actually so deluded that they think this doesn't make them look idiotic. If it's the latter, we need to bring bullying back.

Either way it's hilarious.

Well the author has many, many repositories like this.

It seems like he's still stuck in the "If I just say to my AI that I want a production-ready package that people will pay me $99/month for, I'll get it eventually, right?" phase of discovering LLMs.

The end-result is many commits saying "fixed all issues, enterprise-ready now as requested!" adding 500 lines of code causing more issues.

The funniest part, to me, is that this only damages his image, instead of solidifying it. We've had so many applicants at my company recently where we go to their github, and they have 10 repositories all obviously vibe-coded together, acting like they made some amazing stuff. Instant deletion of application, no coming back from that - this person would NOT get a job here.

Reply on news.ycombinator.com