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

Don’t use AI to write things that you present as your own work

satisfice.com

Post reply on HN
Print view View original

101–108 of 108 posts

Re: Don’t use AI to write things that you present as your own work

#101

Post by scotty79 »

solid_fuel wrote: post #100
scotty79 wrote: post #98

Earlier quoted context omitted.

> No… No they cannot. Have you tried? Give it your feed. Ask it to pass it through antislop + SLOP_Detector + slop-forensics and whatever else your agent can find for you. Will it filter out every last bit of AI content? No, but I'm sure it will spare you from more than half. Will it leave in every last bit of human produced content? No. But the ones it's going to filter out are gonna read like slop to you anyways. W…

> Have you tried? Give it your feed. Ask it to pass it through [...] I've reworded this a few times but I still can't figure out how to say it gently, so I'm just going to say it. Your lack of knowledge about something - in this case the detection of LLM generated content - is not on the same footing as actually having knowledge about it. In this case I am already familiar enough with the state of the art for content…

> Your lack of knowledge about something - in this case the detection of LLM generated content - is not on the same footing as actually having knowledge about it.

I'm not sure why are you saying that. Do I need to be on equal footing to suggest something to you? I'm obviously not on equal footing as a complete ignorant. I never touched the subject because my brain slop filter serves me perfectly well even though it's not 100% accurate.

> And both achieve the same result: no usable tool.

The goal is not to build revolutionary tool that achieves state-of-the-art in slop filtering. The goal is to make your life a bit better by cobbling few things together at near zero cost.

> It's not worth spending the time or money it would take to slop together something that isn't going to do the job.

What money? Have you maxed out your sub for the week? What time? I do such things while I'm cooking my dinner.

Processing this topic in your head (which you obviously did for quite a while to earn your footing) already costed you more time and at least as much money.

Re: Don’t use AI to write things that you present as your own work

#102

Post by solid_fuel »

scotty79 wrote: post #101
solid_fuel wrote: post #100

Earlier quoted context omitted.

> Have you tried? Give it your feed. Ask it to pass it through [...] I've reworded this a few times but I still can't figure out how to say it gently, so I'm just going to say it. Your lack of knowledge about something - in this case the detection of LLM generated content - is not on the same footing as actually having knowledge about it. In this case I am already familiar enough with the state of the art for content…

> Your lack of knowledge about something - in this case the detection of LLM generated content - is not on the same footing as actually having knowledge about it. I'm not sure why are you saying that. Do I need to be on equal footing to suggest something to you? I'm obviously not on equal footing as a complete ignorant. I never touched the subject because my brain slop filter serves me perfectly well even though it's…

> I'm not sure why are you saying that.

To explain to you that you are wasting your time and that your suggestions on this are not going to move the needle, nor are they particularly valuable.

> I never touched the subject because my brain slop filter serves me perfectly well even though it's not 100% accurate.

The whole idea is to avoid wasting brain power on filtering slop.

> The goal is not to build revolutionary tool. The goal is to make your life a bit better.

And since it isn't possible to build the required tool, wasting time and money trying to get an LLM to do something that it can't will in no way make my life better.

> What money? Have you maxed out your sub for the week?

Are you somehow under the impression that all or even most people pay for LLM subscriptions? I concluded it wasn't worth the cost at all.

> What time? I do such things while I'm cooking my dinner.

You either spend time writing the prompt and checking the result, or you don't. I'm not interested in jiggling an LLM while I'm cooking.

> Processing this topic in your head (which you obviously did for quite a while to earn your footing) already costed you more time and about as much money.

Actually if you have knowledge about the field it doesn't take much time at all to sort things by difficulty and understand what approaches are possible and what are worthless. And I have the advantage of carrying that knowledge forward, which means I don't need to spend time prompting an LLM to learn what I already know.

Re: Don’t use AI to write things that you present as your own work

#103

Post by nh23423fefe »

justanotherjoe wrote: post #77
627467 wrote: post #72

Earlier quoted context omitted.

Remember how the internet works: you request the content. Web content is not a highway billboard you didnt ask for. If you found a "crappy" content consider your referral. We should all consider our sourcing. The world doesn't owe us anything

This kind of thinking assumes quantitative changes never result in qualitative change. Which is demonstrably false. Obviously there are thresholds where phase transitions happen. Take saltwatwer and cell. After some threshold of saltiness is passed, the water flow reverses from into the cell to away from the cell. And the cell frikkin dies.

This kind of argument doesn't bother to demonstrate the parallel in the logic. This bad form of argumentation should be labelled non rigorous. Then I would've have read it at all.

You never bothered explaining why saltwater is ai writing and readers are cells, nor what the parallel of immersion or osmotic pressure would be.

You lazily wanted me to fill in the details and be convinced. Why did you create this bad argument and leave it unlabelled.

Re: Don’t use AI to write things that you present as your own work

#104

Post by mjd »

eloisius wrote: post #36
mjd wrote: post #29

Earlier quoted context omitted.

Your use of the phrase “undeserved politeness” suggests to me that maybe you aren't someone I should look to for advice on how to behave.

That's probably true, but to be clear, I'm not communicating my own values here. I'm trying to understand your rationale. >This recruiter didn't deserve my authentic voice or my personal toil Yet they deserve inauthentic politeness from a chatbot? It just sounds like "fuck you" with more steps.

Not from the chatbot. From me. I'm the one who sent the message.

And it wasn't inauthentic. I wanted to be polite and inoffensive, but I didn't know how.

Re: Don’t use AI to write things that you present as your own work

#105

Post by scotty79 »

solid_fuel wrote: post #102
scotty79 wrote: post #101

Earlier quoted context omitted.

> Your lack of knowledge about something - in this case the detection of LLM generated content - is not on the same footing as actually having knowledge about it. I'm not sure why are you saying that. Do I need to be on equal footing to suggest something to you? I'm obviously not on equal footing as a complete ignorant. I never touched the subject because my brain slop filter serves me perfectly well even though it's…

> I'm not sure why are you saying that. To explain to you that you are wasting your time and that your suggestions on this are not going to move the needle, nor are they particularly valuable. > I never touched the subject because my brain slop filter serves me perfectly well even though it's not 100% accurate. The whole idea is to avoid wasting brain power on filtering slop. > The goal is not to build revolutionary…

> The whole idea is to avoid wasting brain power on filtering slop.

Not really different from wasting your muscle power in a gym. My rejections of trash content, by AI or humans are faster than ever.

> And since it isn't possible to build the required tool, wasting time and money trying to get an LLM to do something that it can't will in no way make my life better.

I guess I have zero chance to make you revisit your well informed opinion by doing anything practical. No matter how trivial and effortless.

> Are you somehow under the impression that all or even most people pay for LLM subscriptions? I concluded it wasn't worth the cost at all.

Ah, that explains a lot. Basically all of it.

> You either spend time writing the prompt and checking the result, or you don't. I'm not interested in jiggling an LLM while I'm cooking.

What do you do when you wait for the food to cook? My brains craves novelty and stimulation pretty much constantly. I used to game and read and watch stuff, but AI agentic coding is as fun and those things, possibly more.

Re: Don’t use AI to write things that you present as your own work

#106

Post by solid_fuel »

scotty79 wrote: post #105
solid_fuel wrote: post #102

Earlier quoted context omitted.

> I'm not sure why are you saying that. To explain to you that you are wasting your time and that your suggestions on this are not going to move the needle, nor are they particularly valuable. > I never touched the subject because my brain slop filter serves me perfectly well even though it's not 100% accurate. The whole idea is to avoid wasting brain power on filtering slop. > The goal is not to build revolutionary…

> The whole idea is to avoid wasting brain power on filtering slop. Not really different from wasting your muscle power in a gym. My rejections of trash content, by AI or humans are faster than ever. > And since it isn't possible to build the required tool, wasting time and money trying to get an LLM to do something that it can't will in no way make my life better. I guess I have zero chance to make you revisit your…

If it’s so important to you personally that someone waste time and money trying to make an LLM slop machine shit out something that they can’t, be my guest and go ahead.

> My brains craves novelty and stimulation pretty much constantly.

Your brain is deep fried in dopamine if you’re seriously splitting your time between cooking and prompting. You should take a walk, drop the subscription for a while, get outside and soak up the sun.

I’m worried that you’re already pretty far down the path to psychosis, going by the rest of your responses.

Re: Don’t use AI to write things that you present as your own work

#107

Post by Seattle3503 »

LearnYouALisp wrote: post #92
Seattle3503 wrote: post #70

Earlier quoted context omitted.

What if I run it through an AI detection model (eg Pangram) first, and if the shows up as AI I skip it entirely. Does that work for you?

However, you will need to cover multiple models as I've found some get no flag on some engines (unless that one does it? I have only used free sites like GPT-Z, ZG, Gmy, Sapling, Copyleaks)

I'm happy with Pangram so far.

Re: Don’t use AI to write things that you present as your own work

#108

Post by scotty79 »

solid_fuel wrote: post #106
scotty79 wrote: post #105

Earlier quoted context omitted.

> The whole idea is to avoid wasting brain power on filtering slop. Not really different from wasting your muscle power in a gym. My rejections of trash content, by AI or humans are faster than ever. > And since it isn't possible to build the required tool, wasting time and money trying to get an LLM to do something that it can't will in no way make my life better. I guess I have zero chance to make you revisit your…

If it’s so important to you personally that someone waste time and money trying to make an LLM slop machine shit out something that they can’t, be my guest and go ahead. > My brains craves novelty and stimulation pretty much constantly. Your brain is deep fried in dopamine if you’re seriously splitting your time between cooking and prompting. You should take a walk, drop the subscription for a while, get outside and…

> If it’s so important to you personally that someone waste time and money trying to make an LLM slop machine shit out something that they can’t, be my guest and go ahead.

Today, on nearly daily basis, I'm taking a stab at the things that were impossible until very recently. I already made useful progress on things that used to be impossible for me personally.

Slop detection wasn't my interest because I don't have a problem with it. But it might maybe soon. I'm currently creating a system for creating cool stories. Don't worry. For my own consumption and pleasure not for any kind of publishing. And I might make anti-slop tools a part of the pipeline to achieve best outcomes for myself.

> Your brain is deep fried in dopamine if you’re seriously

Nah, I've been like that since before personal computer became a thing. ADHD probably.

Also I'm walking outside a lot and my brain races from one cool thhought to another while I'm walking. I barely notice the surroundings because most of the time they are about 20% as interesting as my thoughts.

Post reply on HN