Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

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

hackerone.com

Post reply on HN
Print view View original

371–380 of 571 posts

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

#371

Post by alexpotato »

Sharlin wrote: post #131
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…

Unfortunately that seems to be the norm now – people literally reduce themselves to a copy-paste mechanism.

I watched someone do this during an interview.

They were literally copy and pasting back and forth the LLM. In front of the interviewers! (myself and another co-worker)

https://news.ycombinator.com/item?id=44985254

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

#372

Post by rchaud »

kspacewalk2 wrote: post #234
jsheard wrote: post #85

Earlier quoted context omitted.

"Actually Indians" was coined to refer to "AI" products which turn out to be outsourced human labor in disguise. Builder.ai was the most infamous example.

It's amazing to me that the human-labour-in-disguise thing was first reported in 2019, but the company only went bankrupt in 2025.

The PowerPoints that sold investors on the company were written and discussed by humans.

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

#373

Post by jdefr89 »

flumpcakes wrote: post #99

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

It is best used for yack shaving in my opinion. Anything other than that and I feel like I cannot trust its output.

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

#375

Post by dwaltrip »

golemotron wrote: post #213
theoreticalmal wrote: post #193

Earlier quoted context omitted.

I have never seen an AI meeting summary that was useful or sufficient in explaining what happened in the meeting. I have no idea what people use them for other than as a status signal

Why do people want to signal their low status?

It’s an attempt to be “cutting edge”

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

#376

Post by shadowgovt »

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…

Quite a few people using AI are using it not only to do analysis, but to do translation for them as well; many people leaping onto this technology don't have English as a fluent language, so they can't evaluate the output of the AI for sensibility or "not sounding like AI."

(It's a noise issue, but I find it hard to blame them; not their fault they got born in a part of the world where you don't get autoconfig'd with English and as a result they're on the back-foot for interacting with most of the open source world).

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

#378

Post by dwaltrip »

pelagicAustral wrote: post #170
dalmo3 wrote: post #146

Earlier quoted context omitted.

The user is now expressing sarcasm.

You're absolutely right, based on the tenor of the previous message exchange, it is likely that brap is indeed sarcastically responding to gryfft. Do you want me to explain the mechanics of this interaction?

puke

gasps for air

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

#379

Post by shadowgovt »

cornholio wrote: post #136
zaphodias wrote: post #80

Earlier quoted context omitted.

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".

We will quickly evolve a social contract that AI are not allowed to directly contact humans and waste their time with input that was not reviewed by other humans, and any transgression should by swiftly penalized. It's essentially spam, automatically generated content that is profitable in large volume because it offsets the real cost to the victims, by wasting their limited attention span. If you wantme to read your…

... and if we can't enforce it with social contract, we'll enforce it with AI on the receiving end.

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

#380

Post by jdefr89 »

baq wrote: post #20
ares623 wrote: post #8

Earlier quoted context omitted.

This is headline driven development. Sooner or later one of these reports will make it and there will be much rejoicing.

s/much rejoicing/pandora's box/ I guess. the thing is, these people aren't necessarily wrong - they're just 1) clueless 2) early. the folks with proper know-how and perhaps tuned models are probably selling zero days found this way as we speak.

Professional Security Researcher here.. I haven't really seen any models reliably find and exploit a 0day. Folks are are at least TRYING to develop such models internally at the MIT lab where I work, but not sure how far along they are coming yet.. If a model is developed that can find a 0day or two (like Big Sleep which I think maybe found some) I won't be surprised but keep in mind fuzzers find thousands of real 0days with far less compute... These capabilities are of course something worth looking into, but too many people are promising 0day oracles already and that simply just isn't where we are right now (or ever? ). Sorry for bad grammar typing quickly from phone here.
Post reply on HN