HN2new | past | comments | ask | show | jobs | submit | Kwpolska's commentslogin

You realise that those sites posted multiple articles per day ten years ago, long before LLMs were invented?

Yup. Now they do it with a fraction of the staff and use LLMs. What's your point?

The websites and apps probably have a system prompt that tells them to be more cautious with stuff like this, so that AIs look more credible to the general public. APIs might not.

The story is credited to Benj Edwards and Kyle Orland. I've filtered out Edwards from my RSS reader a long time ago, his writing is terrible and extremely AI-enthusiastic. No surprise he's behind an AI-generated story.

Is he even a real person I wonder

He was murdered on a Condé Nast corporate retreat and they have been using an AI in his likeness to write articles ever since!

Would make for a good book, company hires famous writer, trains an ai on them, tortures them to sign over their likeness rights and then murders them. Keeps up appearances of life via video gen, voice gen and writing gen.


Going on my to watch list

> his writing is terrible and extremely AI-enthusiastic

I disagree, his writings are generally quite good. For example, in a recent article [1] on a hostile Gemini distillation attempt, he gives a significant amount of background, including the relevant historical precedent of Alpaca, which almost any other journalist wouldn't even know about.

1: https://arstechnica.com/ai/2026/02/attackers-prompted-gemini...


For what it's worth, both the article you're linking to and the one this story is about are immediately flagged by AI text checkers as LLM-generated. These tools are not perfect, but they're right more often than they're wrong.

>These tools are not perfect, but they're right more often than they're wrong.

Based on what in particular? The only time I have used them is to have a laugh.


Based on experience, including a good number of experiments I've done with known-LLM output and contemporary, known-human text. Try them for real and be surprised. Some of the good, state-of-the-art tools include originality.ai and Pangram.

A lot of people on HN have preconceived notions here based on stories they read about someone being unfairly accused of plagiarism or people deliberately triggering failure modes in these programs, and that's basically like dismissing the potential of LLMs because you read they suggested putting glue on a pizza once.


I had fun with AI detectors in particular for images, even the best one (Hive in my opinion) was failing miserably with my tests, maybe the one trained on text are better but I find it hard to trust them, in particular if someone know how to fiddle with them.

I just tested originality.ai and it claimed 100% probability that the editors note on the Ars retraction [1] was itself AI generated. For the Gemini article on Benji Edwards it was "only" 56%.

I think your tools need a lot more evidence to be considered reliable.

1: https://arstechnica.com/staff/2026/02/editors-note-retractio...


> immediately flagged by AI text checkers as LLM-generate

Proof? Which one? I would like to test a few other articles with your checker to test its accuracy.


hey! im not op but ive used originality.ai before and it saved my ass. its super sensitive, but also super accurate

I tested it, I think it's super inaccurate.

Also filtered out the following slop generators from my RSS feed, which significantly enhanced my reading experience:

Jonathan M. Gitlin

Ashley Belanger

Jon Brodkin

I wonder how soon I will be forced to whitelist only a handful of seasoned authors.


> I wonder how soon I will be forced to whitelist only a handful of seasoned authors.

Twenty years ago?


Of course videos disappear for copyright, ToS violations, or when the uploaders remove them. They do not disappear just because nobody watched them.

There’s a whole activity around discovering random 15 year old videos with almost no views. It’s usually some random home video

A friend of mine worked two years in YouTube as a content admin.

Basically being given videos to watch all day, especially coming from the middle east (this was ISIS time so any video from the area had someone watching it as soon as uploaded).

Needless to say there's endless gold no view videos according to him.

It's also interesting that it was no open secret that already in 2018 they were all told that they were essentially training machines to do their job.


I was interested in the same thing and built a search for it

https://ytstalker.mov


They also disappear when the government of Pakistan tells Google to erase them: https://lee-phillips.org/youtube/

Because we would miss simonw’s self-promotion blog posts.

jealous much?

What if you have a wildcard cert for *.example.com?

Much better. But you still leave traces from dns queries.

Subfinder has a lot of sources to find subdomains, not only certs: https://github.com/projectdiscovery/subfinder


I worked at a company where the security team disliked wildcard certificates because it exposed us to the risk of someone, somehow, hosting something malicious on a subdomain.

LibreOffice has had a ribbon for a while.

I know, but it was not this direct ms office clone.

This is help for the classic LibreOffice app, not the new web thing (which appears to have no settings?)

I installed it from Windows Store, opened a blank text document, and the styles box appears to contain white text on a white background.

I opened a blank spreadsheet, typed in something, tried to create a pivot table, and it only expanded the selection without showing the dialog box.

I restarted it and those bugs were fixed, but the Pivot Table UI is still the ugly non-interactive one found in LibreOffice (which Excel got rid of 26+ years ago).

Uninstalled.


Yeah, unfortunately this seems to combine the UI and performance issues of LibreOffice with new issues from the new front end.

It also has a basic mistake in text editor UX: the caret blinks independently of caret movement. This means that the caret is invisible half of the time while trying to navigate text. Most text editors avoid this by restarting the blink cycle to force the caret visible on each movement.


> It also has a basic mistake in text editor UX: the caret blinks independently of caret movement. This means that the caret is invisible half of the time while trying to navigate text. Most text editors avoid this by restarting the blink cycle to force the caret visible on each movement.

It doesn't do that on my computer. LibreOffice 7.0.4.2 shipped with GNU/Linux Debian.


Standard LibreOffice Writer doesn't have the issue for me either, only the version shipped in Collabora Office for Desktop.

Sorry, I missed that it wasn't about plain LibreOffice.

File a bug!

My debit card has specific limits, far less than all my funds. There also exist pre-paid cards, ideal for things like this.

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

Search: