Re:Why the push for Agentic when models can barely follow a ...
by jrflowers on Tuesday October 14, 2025 @04:48AM
(#45577713)
I personally always love the “I wrote an entire codebase with claud” posts where the response to “Can we see it?” is either the original poster disappearing into the mist until the next AI thread or “no I am under an NDA. My AI-generated code is so incredible and precious that my high-paying job would be at risk for disclosing it”
[ Parent
| on HN ]
Re:Why the push for Agentic when models can barely follow a ...
by isodev on Tuesday October 14, 2025 @06:23AM
(#45578285)
NDA on AI generated code is funny since model outputs are technically someone else’s code. It’s amazing how we’re infusing all kinds of systems with potential license violations
Re:Why the push for Agentic when models can barely follow a ...
by fragmede on Tuesday October 14, 2025 @06:49AM
(#45578448)
If anyone actually believed those requests to see code were sincere, or if they at least generated interesting discussion, people might actually respond. But the couple of times I've linked to a blog post someone wrote about their vibe-coding experience in the comments, someone invariably responds with an uninteresting shallow dismissal shitting all over the work. It didn't generate any interesting discussion, so I stopped bothering.
https://mitchellh.com/writing/non-trivial-vibing went round here recently, so clearly LLMs are working in some cases.
Re:Why the push for Agentic when models can barely follow a ...
by skydhash on Tuesday October 14, 2025 @07:57AM
(#45578958)
And I think, in this blog post, the author stated that he does heavy editing of what’s generated. So I don’t know how much time is saved actually. You can get the same kind of inspiration from docs, books, or some SO answer.
Re:Why the push for Agentic when models can barely follow a ...
by habinero on Tuesday October 14, 2025 @07:22AM
(#45578671)
Haters gonna hate, but the haters aren't always wrong. If you just want people to agree with you, that's not a discussion.
Re:Why the push for Agentic when models can barely follow a ...
by hkt on Tuesday October 14, 2025 @06:00AM
(#45578148)
Honestly I've generated some big ISH codebases with AI and have said so and then backed off when asked.. because a) I still want to try to establish more confidence in the codebase and b) my employment contract gleefully states everything I write belongs to my employer. Both of those things make me nervous.
That said, I have no doubt there are also bots setting out to generate FOMO
Re:Why the push for Agentic when models can barely follow a ...
by ponector on Tuesday October 14, 2025 @08:30AM
(#45579256)
Everything you wrote belongs to them. But it's not you, it's Claude is the author.
Re:Why the push for Agentic when models can barely follow a ...
by ponector on Tuesday October 14, 2025 @08:27AM
(#45579233)
And they are usually 10x more productive as well!
Re:Why the push for Agentic when models can barely follow a ...
by benibela on Tuesday October 14, 2025 @09:18AM
(#45579671)
Someone posted these single file examples: https://github.com/joaopauloschuler/ai-coding-examples/tree/...