Hacker News Reader

Comment by deadbabe | Hacker News Reader

Why are we pushing these pure ChatGPT authored blog posts?

Genuine question: how are you so sure this is ChatGPT authored? And do you mean the whole thing is written by ChatGPT, unedited by a human, or do you mean the author used ChatGPT to edit their article?

There is a typical pattern that LLMs follow when writing

things like:

  "# What Actually Helped

  Here’s a story."

is exactly what I see anytime I ask ChatGPT to generate an article. I guess when you get used to trying to use it to generate content, you become good at identify it.

I'm still not seeing it in this particular article. I really want to know if I'm missing something or it the parent commenter is just being overly dismissive.

Do you use ChatGPT? I’ve read a lot of ChatGPT output… this has all the qualities of typical GPT responses: Short punchy sentences as filler, neatly divided sections of similar length with pointless headings, use of “—“, italics, the length of the article is about the length of a typical response, every word is what you would expect to read, there’s no surprises or unexpected sayings, the article follows this style of “and then… and then… and then…”, it doesn’t foreshadow or weave multiple parallel threads of thought or make callbacks. It’s AI garbage.

Because here people just wanna discuss titles, and the title is the whole content in this case - the prompt.

Someone should build a SAAS to deterministically generate a blog post and title from a URI so we can automate this. I wonder if they'd change the site rules to deal with it.

Bad writing is perennial

As much as I'd like it, it would be inconsequential for YCombinator's portal to ban AI-generated posts.

Reply on news.ycombinator.com