Hacker News new | ask | show | jobs
by albertgoeswoof 4 days ago
Please stop using AI to write for you, it ruins what is otherwise a fascinating story, and on reflection I struggle to trust it.

If you used AI to generate the blog post, did you use AI to generate the screenshots and story?

8 comments

Agreed. The post looks great. The story is great but the AI style in this case does distract.

I'm not against using AI for writing at all but you want to be careful that the output doesn't contain too much of this noise over signal type of wording that repeats and wants to just sell you something.

Unfortunate to see AI police as top comment on a good amusing post!
I am curious, what exactly triggers your AI senses in this post?
"This wasn't some dev environment. This wasn't test data."

It's not X, it's Y. And repetitions of three.

Thats a literary style called anaphora. Some people learned this in school, so they can use it to emphasize something. IMHO this is not a strong sign of AI, in fact I think this text has no strong AI indicators.
Something having name is not a sign of not-ai. It when it is in places for no reason other then blow up the text length. As if school kid was trying to make the text longer to fit the minimal amount of words.
Thank you for speaking up -- I feel the same way.
Agree, i don't think it was ai written, or atleast written very well with AI.

I am usually able to pick up ai writing quickly but didnt feel it in this case

Same, and I'm growing tired of the witch hunt comments derailing the topic.

For example, why is this thread the top thread in this post?

You're right to call that out
Also, “ What I didn't expect was what happened next.”

The Unicode arrows is also something Claude is using really often: “Camera -> RTMP ingest -> MediaKind -> broadcast partners -> your TV.”

And the table at the end is such a Claude thing.

The general style, a series of short sentences that feel like they are building up a punchline is what tells me it’s Claude, but the whole thing does stink of LLM generation

The table don't even add information, all but one cell share the same value for "When".
Ah la la, this is something I actually write sometimes. And em-dashes.

This hunt for AI is sometimes counterproductive

I found myself writing exactly like this for a while after reading pages and pages of this special construct before my bs detector understood what it is...
What do you think why AI uses that pattern so often?

Could it be in the human generated training data?

Seems like a logical fallacy. Just because AI uses that pattern doesn’t mean that pattern means AI

Gemini loves talking about "The Nuclear Option" too
You've hit a classic "gotcha".
It's honestly unbelievable how people continue to paste stuff like this. Do they not know that credibility is lost instantly?
I love how in one post everyone piles on someone for using the word "retarded", meanwhile in another post someone with autism gets piled on for using AI for accessibility.
I remember a frontpage post from like 2 days ago:

"If you want human attention show human effort" or something in that direction. I think this fits here just right.

I think in this case the human effort was put into the actual discovery, honestly I don't mind if AI helped him write the blog post if the result is enjoyable and not sloppy
I don't mind it here at all, in fact I didn't even notice it's AI before reading this comment. It's clearly not a one-shot AI slop but a well thought out and edited by a human post.

Not everyone who has something interesting to say is a good writer, and I think it's great if AI can help them tell their stories.

Yeah I used Claude as a writing assistant for the initial draft. I'm autistic and long-form writing isn't my strong suit, getting a 4000 word blog post to flow well is genuinely hard for me. But I do edit it pretty heavily after, the voice and the jokes and the structure are mine, the AI just helps me get a baseline down so I'm not staring at a blank page. The research, the screenshots, the disclosure, that's all me. I've been doing this stuff for years.
My opinion is that this is a great story.

But the haters are going to hate.

If you had not used AI to fix your post, I bet the top post will be complaining about your grammar.

Some people will always find something negative. Simple as that.

While I appreciate the positivity, but I've honestly grown to appreciate grammar mistakes. Just like it's not X, it's Y indicated AI, those indicated that the effort behind was human.

Honestly, no need to write 4000 words if the story can be told in 400. The story is what matters, not word count or "flow".

I think the criticism is constructive. It's really not about hating. I'd wager many of the people who convey this criticism do use AI to aid their writing as well.

It's just that this one in particular lacks one more edit pass removing some of the AI noise on branding-speak and needless repetition (AI tends to list things and beat the point).

> If you had not used AI to fix your post, I bet the top post will be complaining about your grammar.

I'm positive that a post complaining about the grammar would've been (rightfully) downvoted to oblivion on this site.

I understand that it feels helpful but the post ends up repeating the same insight over and over. Reads very sloppy, while you wanted the opposite.
Don't take my criticism of AI writing as criticism of your work. This is stellar stuff! What I'm trying to say is I'd really like to hear it in your words.

I'm glad to hear the voice is yours, and I apologize for assume it was the AI's.

Great post. It was an amusing read, and quite the discovery! Good work, and great job documenting it.
> I'm autistic and long-form writing isn't my strong suit, getting a 4000 word blog post to flow well is genuinely hard for me.

I find the way I interact with the world is exceptionally different from the descriptions of everybody else. Some of the symptoms of such manifest as difficulty communicating... with most people.

There are a subset of people who I have not only no problem, but seemingly a drastically increased information exchange rate.

Do with that observation what you will, but I don't write for the lowest common denominator, the preferred style of AI, because I don't write for the people who cant be bothered to understand me. I'm writing for people like me.

n.b. Maybe you are writing for the lowest common denominator. In which case, say that: "Yeah I know it sounds like AI but it's supposed to be advertising, not a technical white paper or PoC"

> the AI just helps me get a baseline down so I'm not staring at a blank page.

If this was true, the top comment wouldn't be a complaint about how the voice of the article sounds "inauthenticlly human" or like an LLM. It's having a stronger influence on your writing than you're giving it credit for. It's on you to decide if or how much you care, but ideally you wouldn't be lying to either yourself or your readers.

Understandable, but open disclosure would help. You of all people must know how hard it is to hide things from being discovered, and this is something that will be discovered without any doubt. Just explain it and most reasonable people will understand. Some won't but at least they will be honestly warned.
I got 100% Human on Pangram, so either they did the work to have their AI service pass this test, or...they actually wrote it.
These comments don't help much. AI is here, not everybody can write well, AI is gonna be used.
I'd wish we come to a day where people would just post the prompt. Then I can decide what story to generate from it.
I'm still planning to add a "AI-edited version" toggle to my blog. Not that it would do anything, because people wouldn't click it anyway.
Look I don't like to see a wall of AI slop as much as the next person (see: https://news.ycombinator.com/item?id=48551462), but "just post the prompt" is also too dismissive. AI had access to information that we don't have and all you see here is probably a compilation of multiple prompts, edits and various sources (like author's notes) for context.

We can adjust our expectations for people to take some time to make the output theirs.

OTOH, and this is me arguing against myself, maybe this is not too different than the million web sites we saw using the unmodified default bootstrap theme.

I guess my opinions as well as the response of the community are still evolving.

- It's called "writing in bullet points"

- Normies frown upon it

They don't! Before AI, people complain about long emails and what not. The literally preferred to read short ones.
The problem is that people who are bad writers have trouble understanding that AI writes worse than they do
you clearly have never read a 1000 word text written by me (/s, but only partially)
Honestly I would prefer to read a long text from a human that is badly written than a LLM version. It’s fine to not write well
> AI is here, not everybody can write well, AI is gonna be used.

I don't know about you, but I'd love to read a fascinating story written by a relatively poor writer. But if they can't be bothered to write, I assume the story can't be that good.

But this isn't a story, literature, or a fancy piece of art; it's merely a technical blog post that discloses a security vulnerability. Here, the writing serves only as a vehicle to convey a message. Once you've received it, its purpose has been fulfilled. I would agree with you if the writing were an important part of the message, but here it is not. Not everybody can write well, and this guy clearly had something to tell, and that is what matters.
> But this isn't a story, literature, or a fancy piece of art; it's merely a technical blog post that discloses a security vulnerability. Here, the writing serves only as a vehicle to convey a message. Once you've received it, its purpose has been fulfilled.

I disagree wholeheartedly. I'm not a machine. I'm a thinking, feeling, human being.

As a mathematician, I can certainly appreciate precise formulations. They have their place. But this is not that place.

> Not everybody can write well, and this guy clearly had something to tell, and that is what matters.

I'm sadenned that he wouldn't tell it in his way. I'd much rather read his own (poorly written?) words.