You did this with an AI and you do not understand what you're doing here • Hacker News Board

Re: You did this with an AI and you do not understand what you're doing here

#91

Post by a235 » Mon, Sep 22, 2025, 10:32 AM UTC

Maintainer or curl gave recently a talk on AI slop in security reports, showing this and other examples:

https://youtu.be/6n2eDcRjSsk?si=p5ay52dOhJcgQtxo -- AI slop attacks on the curl project - Daniel Stenberg. Keynote at the FrOSCon 2025 conference, August 16, in Bonn Germany by Daniel Stenberg.

Plus, linked above, his blogpost on the same subject https://daniel.haxx.se/blog/2025/08/18/ai-slop-attacks-on-th...

Re: You did this with an AI and you do not understand what you're doing here

#92

Post by belter » Mon, Sep 22, 2025, 10:33 AM UTC

dansmith1919 wrote: post #10

Crazy how he doubled down by just pasting badger's answer into Chat and submitting the (hilariously obvious AI) reply: > Thanks for the quick review. You’re right — my attached PoC does not exercise libcurl and therefore does not demonstrate a cURL bug. I retract the cookie overflow claim and apologize for the noise. Please close this report as invalid. If helpful, I can follow up separately with a minimal C reproduc…

Crazy on how the current 400 Billion AI bubble is based on this being feasible...

Re: You did this with an AI and you do not understand what you're doing here

#93

Post by Ralfp » Mon, Sep 22, 2025, 10:33 AM UTC

listic wrote: post #83

dansmith1919 wrote: post #12

Earlier quoted context omitted.

At some point they told ChatGPT to put emoji's everywhere which is also a dead giveaway on the original report that it's AI. They're the new em dash.

What was it with em dash?

People usually don't type embdash, just use regular dash (minus sign) they have already on the keyboard. ChatGPT uses emdash instead.

Re: You did this with an AI and you do not understand what you're doing here

#94

Post by zelphirkalt » Mon, Sep 22, 2025, 10:33 AM UTC

rsynnott wrote: post #9

This must be _absolutely exhausting_.

Yeah, I guess if I was him, I would just close issues silently and ban the person who created them, if possible. I don't think I could be as nice as he is.

Re: You did this with an AI and you do not understand what you're doing here

#95

Post by henrebotha » Mon, Sep 22, 2025, 10:34 AM UTC

zaphodias wrote: post #80

rapidaneurism wrote: post #69

Earlier quoted context omitted.

I wonder if there was a human in the loop to begin with. I hope the future of CVS is not agents opening accounts and posting 'bugs'

I don't think there are humans involved. I've now seen countless PRs to some repos I maintain that claim to be fixing non-existent bugs, or just fixing typos. One that I got recently didn't even correctly balanced the parenthesis in the code, ugh. I call this technique: "sprAI and prAI".

Hey don't hate on us humans who genuinely do open random PRs to random projects to fix typos. https://github.com/pulls?q=is%3Apr+author%3Ahenrebotha+archi...

Re: You did this with an AI and you do not understand what you're doing here

#96

Post by ioteg » Mon, Sep 22, 2025, 10:34 AM UTC

[flagged]

Re: You did this with an AI and you do not understand what you're doing here

#97

Post by listic » Mon, Sep 22, 2025, 10:34 AM UTC

BoredPositron wrote: post #46

dansmith1919 wrote: post #10

Crazy how he doubled down by just pasting badger's answer into Chat and submitting the (hilariously obvious AI) reply: > Thanks for the quick review. You’re right — my attached PoC does not exercise libcurl and therefore does not demonstrate a cURL bug. I retract the cookie overflow claim and apologize for the noise. Please close this report as invalid. If helpful, I can follow up separately with a minimal C reproduc…

It's an n8n bot without user input. If you Google the username you'll find a GitHub full of agent stuff.

Who was likely to start it and for what purpose?

Re: You did this with an AI and you do not understand what you're doing here

#98

Post by unmole » Mon, Sep 22, 2025, 10:35 AM UTC

Havoc wrote: post #21

dansmith1919 wrote: post #10

Crazy how he doubled down by just pasting badger's answer into Chat and submitting the (hilariously obvious AI) reply: > Thanks for the quick review. You’re right — my attached PoC does not exercise libcurl and therefore does not demonstrate a cURL bug. I retract the cookie overflow claim and apologize for the noise. Please close this report as invalid. If helpful, I can follow up separately with a minimal C reproduc…

Makes me wonder whether the submitter even speaks english

The username sounds Turkish. Make what you will of it.

Re: You did this with an AI and you do not understand what you're doing here

#99

Post by flumpcakes » Mon, Sep 22, 2025, 10:35 AM UTC

Has anyone seen a good use of AI in the wild? Every example I see is honestly depressing, such as this.

Re: You did this with an AI and you do not understand what you're doing here

#100

Post by henrebotha » Mon, Sep 22, 2025, 10:36 AM UTC

pjmlp wrote: post #37

jsheard wrote: post #26

Earlier quoted context omitted.

I think the JS/Node scene was the pioneer in spamming emojis absolutely everywhere, well before AI. Maybe that's where the models picked it up from.

I really hate all those CLI applications and terminal configurations that look like circus came to town.

I don't love emojis for this purely because they're graphically inconsistent; I can't style them with my terminal font or colour scheme. But I'm a huge fan of using various (single-width) unicode chars with colour to make terminal output a lot easier to parse, visually. Colour and iconography are extremely useful.