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

Throwing AI-generated walls of text into conversations

noslopgrenade.com

Post reply on HN
Print view View original

351–360 of 453 posts

Re: Throwing AI-generated walls of text into conversations

#351

Post by imperfect_blue »

handoflixue wrote: post #347
idle_zealot wrote: post #261

Earlier quoted context omitted.

> I choose to look at it like any other cultural communication difference There's a way to tell if this is actually the case: can you find the members of this culture that like getting slopgrenaded? A communication culture needs both speakers and listeners. I see the speakers, I have yet to see the listeners. I could just be missing them though. In the case that the listeners are greatly outnumbered by the speakers w…

Anecdotally, LLMs are apparently good enough at writing to routinely make the front page of Hacker News. Twitter has a plague of bot comments. From all these up-votes, we can infer there is an audience and this is good faith. Partly, I admit, I just tend to be wary of shame as a tool. "Someone said something I dislike" has a long history of being abused. Conversely, even something like your gum example, one can point…

Using LLM as a writing assistant to properly format your ideas, while lazy, is not the same thing as a slop bomb. They're highly rated because the original ideas are good, even if the writing quality is sloppish, and it genuinely contains human input that you can't get simply by asking LLMs a question.

Re: Throwing AI-generated walls of text into conversations

#352

Post by grey-area »

handoflixue wrote: post #347
idle_zealot wrote: post #261

Earlier quoted context omitted.

> I choose to look at it like any other cultural communication difference There's a way to tell if this is actually the case: can you find the members of this culture that like getting slopgrenaded? A communication culture needs both speakers and listeners. I see the speakers, I have yet to see the listeners. I could just be missing them though. In the case that the listeners are greatly outnumbered by the speakers w…

Anecdotally, LLMs are apparently good enough at writing to routinely make the front page of Hacker News. Twitter has a plague of bot comments. From all these up-votes, we can infer there is an audience and this is good faith. Partly, I admit, I just tend to be wary of shame as a tool. "Someone said something I dislike" has a long history of being abused. Conversely, even something like your gum example, one can point…

Many people don’t read the articles before upvoting they tend to just read the headline. So I don’t think this proves much.

Re: Throwing AI-generated walls of text into conversations

#353

Post by krupan »

otherme123 wrote: post #275
SoftTalker wrote: post #134

Earlier quoted context omitted.

I'd prefer just "I don't know." I can ask Claude on my own.

Some years ago, answering "just google it!" was considered rude in some forums. It was asumed that some people didn't have enough googlefu, so googling for them was considered a good help. Also, almost all StackOverflow could be boiled down to a "RTFM" and close the questions. Today there might be people who can't extract enough juice from LLMs, so it is not entirely useless to say "I was able to extract this info fr…

This is not my experience at all. Getting a LMGTFY or an RTFM response really did mean that your question was easily answered by Google or the "fine" manual. It wasn't rude, it was an education, even a confidence boost saying, "you can figure this out, I believe in you!"

Re: Throwing AI-generated walls of text into conversations

#354

Post by HeartStrings »

gitaarik wrote: post #326
HeartStrings wrote: post #156

Earlier quoted context omitted.

We should instaban anyone who uses these patterns on ycombinator. If you talk like that for realsies, well too bad, try being more human and less linkedin next time

But AI is trained on human text. Human writers have already been flagged as AI – by humans who think they're AI, while they were not.

AI has been trained on the common author. If you don’t want to sound like AI, find your unique voice and write less like everyone else.

Re: Throwing AI-generated walls of text into conversations

#355

Post by byzantinegene »

fortyseven wrote: post #244
SoftTalker wrote: post #134

Earlier quoted context omitted.

I'd prefer just "I don't know." I can ask Claude on my own.

That's how I'm finding I feel, too. It's not that I don't use LLMs myself, but I trust myself enough to know how to filter that data that comes back from them and compare it to actual facts. I don't necessarily know the processing of this information from the other person, to trust that they vet the information properly. So instead of it being helpful to me, it becomes incredibly irritating. Especially because I just…

exactly, the people who copy-paste AI responses are typically the ones who don't do due diligence in verifying the responses.

Re: Throwing AI-generated walls of text into conversations

#356

Post by gitaarik »

HeartStrings wrote: post #354
gitaarik wrote: post #326

Earlier quoted context omitted.

But AI is trained on human text. Human writers have already been flagged as AI – by humans who think they're AI, while they were not.

AI has been trained on the common author. If you don’t want to sound like AI, find your unique voice and write less like everyone else.

They haven't only been trained on the common author. They have also been trained to prefer quality and to dismiss hallucinations. So if a human actually does some quality writing, people might think it's AI, like has already happened.

Re: Throwing AI-generated walls of text into conversations

#358

Post by imhoguy »

ungreased0675 wrote: post #233
csbrooks wrote: post #129

Earlier quoted context omitted.

To me, acting in good faith means saying something like "I'm not sure, but Claude says this, which sounds right: [short informative clip from Claude's wall of text]". Don't pretend it's your response, make sure it has info you think is useful, and edit it down.

Even that is bad. It’s similar to “I don’t know, but these are the top three links when I searched Google.” It’s a useless response, verbal chaff. And friendlies don’t use countermeasures.

"Google it: " was the right response back in the time, now it should be "Claude it: ".

Re: Throwing AI-generated walls of text into conversations

#359

Post by imhoguy »

wisty wrote: post #339
signatoremo wrote: post #311

Earlier quoted context omitted.

No. "I don't know" may be interpreted as "I don't care". Adding additional info may be a way to say "I do care". Similar to sometimes we say "I don't know, but I googled it, and I got this". Also, LLM's answers can be good if the prompts are good, so can still be helpful.

Turning up to a potluck with no food says "I don't care". Turning up to a potluck with a bag of garbage you got for free in the dumpster on the way says "I don't care, and I'm so dumb I think you're too dumb to know this is trash".

"I don't know. Do you want me to figure it out?" would be my take, depending on circumstances.

Re: Throwing AI-generated walls of text into conversations

#360

Post by imhoguy »

Terr_ wrote: post #308
filoeleven wrote: post #284

Earlier quoted context omitted.

And the money keeps on rolling upward...

To cynically summarize a disturbing portion of LLM business-models: 1. Find companies which require or reward employees for generating bullshit, fluff, or unreliable signals for actual work that isn't measured well. 2. Offer those employees convenient BS-as-a-Service. 3. Convince the corporation to subsidize the new bullshit-production method as a business expense. I hope this results in the overloaded collapse of th…

4. Wait for AI token price skyrocket.

5. Approach them to buy your convenient anti-BS-as-a-Service optimizer.

Rinse, repeat.

Post reply on HN