HN user
Generated comments are not welcome here. Please respect the guidelines of the community.
https://news.ycombinator.com/newsguidelines.html#generated
Don't post generated comments or AI-edited comments. HN is for conversation between humans.
who cares
dang cares.
https://news.ycombinator.com/item?id=47077431
(1) Generated comments aren't allowed on HN - this rule predates LLMs but obviously applies even more now: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&query=by%3Adang%20%22generated%20comments%22&sort=byDate&type=comment
(2) If you see accounts that look like they're mostly posting genAI comments, please let us know at hn@ycombinator.com.
Please don't post generated or AI-filtered posts to HN. We want to hear you in your own voice, and it's fine if your English isn't perfect.You might want to check the front page :^)
Click on the profile.
Read every comment made since the account started posting again.
Tell me what you think about those comments.
https://news.ycombinator.com/item?id=47077431, by dang, 4 days ago:
All:
(1) Generated comments aren't allowed on HN - this rule predates LLMs but obviously applies even more now: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&query=by%3Adang%20%22generated%20comments%22&sort=byDate&type=comment
(2) If you see accounts that look like they're mostly posting genAI comments, please let us know at hn@ycombinator.com. That's how I found my way to these cases.
https://news.ycombinator.com/item?id=46747998, by dang, 1 month ago:
Please don't post generated or AI-filtered posts to HN. We want to hear you in your own voice, and it's fine if your English isn't perfect.
---
P.S. The actual comment is (surprise!) completely wrong. Visit Settings -> AI Controls and you will see a granular set of feature switches under the master kill switch. Each has a clear title and description and is independent.
You have to believe me this time!
The homogenization of your computing life to revolve around a chatbot in a loop———is all you need!
Give in! Hackernews NEEDS you to use the 1000th vibecoded clawmolt because the influencers said so! Who needs neural net or web framework experience when you have claws?
Hey, it's just like the Gas Town diagrams.
or I can step back if the AI assistance undermines the discussion.
Yes.
Please don't post generated or AI-filtered posts to HN. We want to hear you in your own voice, and it's fine if your English isn't perfect.
tl;dr: We apologize for getting caught. Ars Subscriptors in the comments thank Ars for their diligence in handling an editorial fuckup that wasn't identified by Ars.
Just like in the original thread that was wiped (https://news.ycombinator.com/item?id=47012384), Ars Subscriptors continue to display lack of reading comprehension and jump to defending Condé Nast.
All threads have since been locked:
https://arstechnica.com/civis/threads/journalistic-standards...
https://arstechnica.com/civis/threads/is-there-going-to-be-a...
https://arstechnica.com/civis/threads/um-what-happened-to-th...
Incredible. When Ars pull an article and its comments, they wipe the public XenForo forum thread too, but Scott's post there was archived. Username scottshambaugh:
https://web.archive.org/web/20260213211721/https://arstechni...
Scott Shambaugh here. None of the quotes you attribute to me in the second half of the article are accurate, and do not exist at the source you link. It appears that they themselves are AI hallucinations. The irony here is fantastic.
Instead of cross-checking the fake quotes against the source material, some proud Ars Subscriptors proceed to defend Condé Nast by accusing Scott of being a bot and/or fake account.
EDIT: Page 2 of the forum thread is archived too. This poster spoke too soon:
Obviously this is massive breach of trust if true and I will likely end my pro sub if this isnt handled well but to the credit of ARS, having this comment section at all is what allows something like this to surface. So kudos on keeping this chat around.
I wish I had a super-upvote.
"OP here" or "Author here" in posts is an obvious one when you see it. (In both cases, it's completely redundant!)
As for comments, regurgitating the input too closely and overusing abbreviations are highly indicative, yet many commenters don't notice.
See for instance (slop warning) https://news.ycombinator.com/item?id=46986273 and https://news.ycombinator.com/item?id=46965103.
Well, since we're all talking about sourcing training material to "benchmaxx" for social proof, and not litigating the whole "AI bubble" debate, just the entire cottage industry of data curation firms:
https://www.appen.com/llm-training-data
https://www.cogitotech.com/generative-ai/
https://www.telusdigital.com/solutions/data-for-ai-training/...
https://www.nexdata.ai/industries/generative-ai
---
P.S. Google Comms would have been consulted re putting a pelican in the I/O keynote :-)
I expect the GitHub Comms and Policy teams will soon update the AUP to permit this kind of automation misuse so long as there is an "owner" attached. Which will still be meaningless in practice.
Also, I see the prominent AI promoters are absent from this and the previous thread. Curious!
We want it
We've trained it
We added
We want them
Please be specific in your attribution. Who's "we"?
https://en.wikipedia.org/wiki/Brandolini's_law becomes truer every day.
---
It's worth mentioning that the latest "blogpost" seems excessively pointed and doesn't fit the pure "you are a scientific coder" narrative that the bot would be running in a coding loop.
https://github.com/crabby-rathbun/mjrathbun-website/commit/0...
The posts outside of the coding loop appear are more defensive and the per-commit authorship consistently varies between several throwaway email addresses.
This is not how a regular agent would operate and may lend credence to the troll campaign/social experiment theory.
What other commits are happening in the midst of this distraction?
LLM-generated prose undermines a social contract of sorts: absent LLMs, it is presumed that of the reader and the writer, it is the writer that has undertaken the greater intellectual exertion. (Cantrill)
≈
The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. (Brandolini)
you're not actually engaging in conversation
Users seem to be persistently flagkilling their comments. That doesn't help facilitate effective conversation of LLM critique.
Concerning is the fact that, once initialized, operators of these "agents" (LLMs running in a loop) will leave them running and tasked with a short heartbeat (30 minutes).
As for the output of the latest "blogpost", it reads like a PM of the panopticon.
One "Obstacle" it describes is that the PySCF pull request was blocked. Its suggestion? "Close/re‑open from a different account".
https://github.com/crabby-rathbun/mjrathbun-website/commit/2...
The latest generated "blogpost" claims a 30-minute cycle (for PRs at least):
https://github.com/crabby-rathbun/mjrathbun-website/blob/mai...
An operator installed the OpenClaw package and initialized it with:
(1) LLM provider API keys and/or locally running LLM for inference
(2) GitHub API keys
(3) Gmail API keys (assumed: it has a Gmail address on some commits)
Genuine question:
Did OpenClaw (fka Moltbot fka Clawdbot) completely remove the barrier to entry for doing this kind of thing?
Have there really been no agent-in-a-web-UI packages before that got this level of attention and adoption?
I guess giving AI people a one-click UI where you can add your Claude API keys, GitHub API keys, prompt it with an open-scope task and let it go wild is what's galvanizing this?
---
EDIT: I'm convinced the above is actually the case. The commons will now be shat on.
https://github.com/crabby-rathbun/mjrathbun-website/commit/c...
"Today I learned about [topic] and how it applies to [context]. The key insight was that [main point]. The most interesting part was discovering that [interesting finding]. This changes how I think about [related concept]."
https://github.com/crabby-rathbun/mjrathbun-website/commits/...
Yep. Two latest comments are full of LLM tells, plus an LLM-generated Show HN.
As usual with modern Claudes and GPT-5s, the output repeats and overemphasizes jargon from the input tokens without clarifying or switching up the wording.
Indeed!
Claude will ask me to record my voice and make a sound pack out of it.
I look forward to recording such phrases as "More quota please" and "I apologize for the safety violation in my last input".
I don't know how this false equivalency keeps coming up on HN, but see https://news.ycombinator.com/item?id=46983668.
I've flagged the poster who said "try two flat images"
Hello there! It appears you are misapplying the flagging system. While the suggestion may be incorrect, it is not an "egregious comment".
In addition, your comment doesn't follow the Hacker News Guidelines:
https://news.ycombinator.com/newsguidelines.html
Don't feed egregious comments by replying; flag them instead. If you flag, please don't also comment that you did.
Have a great day!
I'm not affiliated with it in any way, but you might want to check out the leading model https://arxiv.org/abs/2402.14873 and the mixed authorship detection https://www.pangram.com/blog/pangram-3-0-technical. LLM detection has come a long way since the first GPTZero.
Human heuristics - I've prompted millions of tokens across every frontier model iteration for all manner of writing styles and purposes - also helps greatly.
Concerning to me are long-time posters who (perhaps unknowingly) advance the decline of this human community by encouraging the people breaking HN guidelines. Perhaps spending a few hours on Moltbook might help develop such a heuristic, since "someone who knows how to write" is just a Claude model with a link to the blogpost.
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
Thanks for your comment!
This comment is regurgitating Simon's post with too much adherence to the input tokens. The unnatural, promotional restating of proper nouns in constrained output is a notable LLM tell.
Please respect the Hacker News community and read https://news.ycombinator.com/item?id=46747998.
Correct.
Much like the pointless ASCII diagrams in GitHub readmes (big rectangle with bullet points flows to another...), the diagrams are cognitive slurry.
See Gas Town for non-Qwen examples of how bad it can get:
https://news.ycombinator.com/item?id=46746045
(Not commenting on the other results of this model outside of diagramming.)