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

There’s no mention of any vibe coding in the post. Believe it or not, there are people who are still able to program by themselves.

The lists in the post look like they're LLM-formatted, em-dashes etc. It's fine, it seems like a fun project to vibe code. Not sure about raising money on Patreon for it, but

> The lists in the post look like they're LLM-formatted, em-dashes etc.

No, not “etc”. What else looks LLM-formatted about them? Because em-dashes are not enough to claim LLM.

Look, I get that you don’t care about proper typographic characters. You don’t have to, that’s fine. But many of us humans do.

https://www.gally.net/miscellaneous/hn-em-dash-user-leaderbo...

And going from “LLM-formatted lists” (without any certainty) to vibe-coded project is a huge leap.


They are very even. They are uniformly bolded. They're long and comprehensive. Most humans would have more variation in length unless they were working to a template or a style guide. A long catalogue of tools is also just way more detail than most people would put into an announcement post... but an LLM doesn't get tired and will barf all that out if you don't stop it.

A more robust and perhaps more vivid indicator are sentences like this: "This isn’t a proof of concept or a weekend project. It’s a real authoring environment."

"It's not X, it's Y" is an LLM tell. "It's a real Z" is another. Together? I'm going to conclude it's LLM generated to like a 90% certainty.

And as the sibling notes, the icons look like LLM SVG output. They're more mangled than even a rushed human would do.

Again, it's fine. If I had more time I'd love to try to vibecode a Flash clone.


This isn't just some em-dashes, it's 126 of them in a single announcement post.

The number is irrelevant. They are being used as separators on every list item, of course there are many. They could all be hyphens or arrows or colons and it would be exactly the same.

Look, I’m not saying this wasn’t written with an LLM. What I’m saying is that you don’t know that judging by the em-dashes alone.


The number is relevant when the comment with the most em-dashes on HN (prior to 2023) had only 31. It's also not the only signal here.

There are no colons or semicolons, anyone who really is a stickler for punctuation would not write exclusively with em-dashes.

And not a single colon! Well, outside the smiley in the opening paragraph.

Just read the first two paragraphs and there's a full stylistic whiplash. I'd bet $100 that the first paragraph was human-written and the rest were almost fully Claude.

It's kind of creepy, like a human rolling their eyes back partway through a paragraph and suddenly speaking like Microsoft Bob.


Nor any semicolons. Also full of nonsense. SWF fundamentally runs action script. The idea that you are going to have a AS3 to C# transpiler and a C# scripting engine AND export SWFs is incongruous. Again, "C# code executed when the playhead enters a keyframe." Now I'd love the oppurinity to learn, but ctrl+f c# here [1] shows 0 results. Though I'm sure I'm just misunderstanding the meaning here ( ≖_≖).

[1] https://open-flash.github.io/mirrors/swf-spec-19.pdf


only thing I can imagine is that they are building it with Unity, hence the C#. You can compile other languages to SWF bytecode, that's what Haxe is.

I'm not the parent comment but if you look at the images they posted you can see that the icons of the tools seems very likely to have been generated by a LLM (the SVGs).

> the icons of the tools seems very likely to have been generated by a LLM

They just look like bad rushed placeholder icons. And why does that immediately scream vibe-coded?


Humans tend to use free icon collections or search the web for suitable images, rather than hand-writing SVG markup.

Why even consider hand-written SVG when drawing tools like Inkscape exist?

Sure, there are occasions to write markup from scratch, but vanishingly few in my experience.


They have the same same inconsistencies and incoherencies that I have encounter with LLMs creating SVG. Very different from what a human would have created (and I don't think you have to train your eyes to see it).

if you truly can't clock that this is AI written, you probably need to desperately develop this skill as soon as possible, because everything outside of the first sentence is so very, very clearly slop.

It’s fake money. It’s not a real betting platform, so you can bet whatever.

Depends on the type of human you want to attract.

Believe it or not, humans can be quite imaginative and creative and had been coming up with these types of hacks way before current LLMs.

> I only hope the models get good enough to not be so samey in the future.

Why would you hope to be more easily fooled?


Not GP but I'm personally hoping that if I'm inevitably doomed to be exposed to this horseshit every day that it becomes tolerable to read. For world-shaking language-based superintelligences, they can't write to save their very expensive lives.

> I'm personally hoping that if I'm inevitably doomed to be exposed to this horseshit every day that it becomes tolerable to read.

Thank you for replying, but that doesn’t answer the question. Why would you want to make made up bullshit output more tolerable to read? Being intolerable to read is a feature, it’s a useful signal to know a piece of text may not have had human review, and that you should spend your time reading something else.

I use that same strategy with website consent banners. If a website is so invasive that they go out of their way to make rejection hard (which, by the way, is against the law), I know it’s a company not worth supporting.


> Well the main thing he is known for is "it's all gonna crash" and that's a fact that this page admits he's wrong about.

Have there been specific claims about when it’s going to crash? I find it hard to believe he claimed it was all going to crash by early 2026. Maybe I’m wrong, I haven’t read all of his posts. But neither did the author, they admit in the repo this is all LLM, nothing was verified by humans.


> All verdicts are LLM-scored, not human-verified.

In other words, could be all slop. Or maybe it’s not. Maybe it’s mixed. No one knows.


Fair critique. The methodology doc covers this: both pipelines agree on the high-confidence clusters (security vulnerabilities, bubble predictions) even though they disagree on edge cases. The repo is public specifically so people can spot-check. If you find a claim where the scoring is wrong, I'd genuinely like to know.

> If you find a claim where the scoring is wrong, I'd genuinely like to know.

So you’re asking me to do the work you should have done in the first place? If you didn’t put any effort into it, why should I waste my time checking your non-work and correcting it to your credit?

If you had actually put in the effort then sure, I’d be amenable to helping making this the best it can be. But you didn’t, so what’s the point? Why should anyone spend their time fixing other people’s slop?


I am curious whether claims are scored more accurately by LLMs when reviewed and edited by LLMs prior to posting the claim.

> Apple Silicon systems also require Rosetta2

Then this software is on a clock. Apple has announced the end of Rosetta.

https://9to5mac.com/2026/02/16/macos-26-4-will-notify-users-...


Safari’s extension model could be really good by now, had they not stopped putting effort into it. You are able to define which extensions have access to which websites, and if that applies always or only in non-Private¹ mode. You can also easily allow an extension access for one day on one website.

But there are couple of things I find subpar:

You can’t import/export a list of website permissions. For a couple of extensions I’d like to say “you have access to every website, except this narrow list” and be able to edit that list and share it between extensions.

On iOS, the only way to explicitly deny website access in an extension’s permissions is to first allow it, then change the configuration to deny. This is bonkers. As per the example above, to allow an extension access to everything except a narrow list of websites is to first allow access to all of them.

Finally, these permissions do not sync between macOS and iOS, which increases the maintenance burden.

¹ Private being the equivalent to incognito.


Contents of the blog are themselves written by LLM.

https://github.com/coollabsio/llmhorrors.com/blob/main/CLAUD...

The whole website seems to be focused on promoting the author and their projects more than sharing the information. Just link to the original.

https://www.reddit.com/r/googlecloud/comments/1reqtvi/82000_...

Posted to HN twice recently.

https://news.ycombinator.com/item?id=47231708

https://news.ycombinator.com/item?id=47184182


What do you expect from a website named llmhorrors.com?

I would expect it to not be written by an LLM. Molly White didn’t run Web3 is Going Great on the blockchain.

https://www.web3isgoinggreat.com/


And looking at her main website https://www.citationneeded.news/ there is a tip jar but it doesn't accept crypto. I was expecting her to take at least the major coins like Ada, Eth and BTC, but she's consistent with her views.

False equivalence, Tesla also does not run their website from a Model S.

The joke is, LLM Horrors is anti-LLM, Web3 is Going Just Great is anti Web3. The equivalent for Tesla would be Tesla putting a ICE inside their model 2 if they didn't believe in EVs.

Another plea for @dang to integrate pangram into all story and comment submissions

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

Search: