Hacker Newsnew | past | comments | ask | show | jobs | submit | nubg's commentslogin

Dear Openrouter blog authors, could you please stop writing your blogposts with LLMs?

The content of your posts is really insightful and interesting, but it's feel like junk quality because of the way LLMs write blogposts.

What was your prompt?


A lot of it was finger written -- curious which part sounded like LLM to you?

> > Here's something most developers overlook: if an LLM has a 2% JSON defect rate, and Response Healing drops that to 1%, you haven't just made a 1% improvement. You've cut your defects, bugs, and support tickets in half.

This sounds AI written.


Meaning parts were LLM written? With no disclosure?

"With no disclosure?"

Why do you have an expectation that a company will disclose to you when they use AI for their copywriting? Do you want them to disclose the software they used to draft and publish? If a manager reviewed the blog post before it went live?


Using words written by other people without disclosure has always been frowned upon. It's called plagiarism.

Plagiarism is bad for a lot of reasons, all of which also apply to the undisclosed use of generative AI.


Why not just publish the prompt? I can then take an LLM of my taste to reformat it the way I want.

Basically, I'm asking for open source blogging!


Next up: blog healing

This is extremely interesting, great idea. Really both thumbs up. Looking for more ideas/lifehack approaches to learning via LLMs.

This tell us more about your (lacking) skills at using AI than the state of AI tools themselves.

You're probably using the free lobotomized versions of LLMs, and shooting a one-off short ambiguous prompt and wondering why it didn't turn out the way you imagined.

Meanwhile people spend hundreds of dollars on pro LLM access and learn to use tool calling, deep research, agents and context engineering.


No, I was using top tier models (paid) and even had an extensive eval test suite against multiple models.

Understood.

I tried this app and immediately got caught. The app window was visible in plain sight to the interviewer. The interviewer asked me what I was doing. I tried to pretend I didn't know what I was talking about, but obviously that only made things worse because they were seeing everything. I got banned from ever interviewing at that company and they said they would report me. I'm not sure where they reported me to, if it's some company information sharong platform or the authorities, but needless to say I am scared and hate myself. I was just nervous about that interview and just wanted to make sure I passed it as I do have the skills they were gonna ask about in the interview. I fucked up.

Could the author please post the prompt this article was generated with?

Very interesting. Is it correct for me to imagine it as some kind of "LoRA" thats continuously adapted as the model goes through its day?

If so, could there perhaps be a step where the LoRA is merged back into the main model?

That would be like sleeping :-)


I don't think that's a great analogy.

LoRAs tend to be adapters bolted onto to systems by people other than the system designers, and they are low rank factorizations.

There is nothing low rank or adapter here.


Kind-of. You could theoretically use LoRA for this, in fact, but it probably wouldn't have enough capacity to make it a proper substitute of the attention mechanism. Instead a full MLP is trained as input chunks get processed.

This seems pretty big news?

Perhaps the start of instant messaging becoming more like email?


Start of facebook's malicious compliance.

Is there something new here? Because it seems the same as the DMA compliance they had to do in early 2024?

See e.g.: WhatsApp Messaging Interoperability - https://news.ycombinator.com/item?id=39633936 - Mar 2024 (146 comments)


It was discussed couple days ago as well

"WhatsApp will become interoperable with other messaging apps in Europe" - https://news.ycombinator.com/item?id=46107109 - 01-dec-2025 - 88 comments


No, lol, WhatsApp remains in full control of user IDs. Maybe if they make WA federated, but we both know this is not gonna happen.


this is pointless, stop wasting time.

just kidding of course, don't get discouraged by some random poster on the internet

Great documentation and dev-friendly feature set. Looks promising.


Their ignorance is your opportunity.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: