Print Page | Close Window

ChatGPT

Printed From: NYC Midnight : Creative Writing & Screenwriting
Category: GENERAL DISCUSSION
Forum Name: Creative Writing Corner
Forum Description: Discuss NYC Midnight Creative Writing Competitions or Creative Writing in general.
URL: https://forums.nycmidnight.com/forum_posts.asp?TID=51407
Printed Date: 28 Mar 2024 at 8:22am
Software Version: Web Wiz Forums 12.05 - http://www.webwizforums.com


Topic: ChatGPT
Posted By: copperdogma
Subject: ChatGPT
Date Posted: 13 Jan 2023 at 4:05pm
[I originally posted this to https://forums.nycmidnight.com/topic51403_post544477.html" rel="nofollow - Questions & Answers . I realized that's more for asking questions of nycmidnight staff, so I'm reposting it here.]

I'm surprised to find there isn't a single post on here about ChatGPT yet.

Primer

If you're not familiar, it's an AI (large language model) released at the beginning of December. It's a simple chat interface and you can ask it to do pretty much any creative or analysis task you can think of, and it does a surprisingly capable job. It's not brilliant, but it's like having access to an average person with infinite time on their hands to do whatever you ask of them.

In my opinion, its novelty lies in its ability to write consistently coherent and self-consistent long-form text and the fact it has a memory (about 5 pages), so if you ask it to "Write me a short story about a pirate and a unicorn," you can then say "That was pretty good, but add a plot twist." And it will.

You can sign up for free to try it:  https://chat.openai.com/chat" rel="nofollow - https://chat.openai.com/chat

Obvious uses for writing contests:
- Generate ten story ideas for the following prompt: Genre: Romance, Word: Kettle, Action: Breaking a vase
- Generate odd names for a restaurant you might find in San Francisco.
- Here's a story I wrote: [your story pasted here]. Punch up the dialogue and add a joke.
- Write a story of maximum 250 words for the following prompt: Genre: Action, Word: Motorcycle, Action: Breaking a vase
etc.

I feel it's obvious that getting it to write your entire story defeats the purpose of a writing contest. But after that it gets.. fuzzy.

What are people's opinions on using it for writing? I've been debating this with my fellow writers recently.


-=Cam


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2



Replies:
Posted By: Pages
Date Posted: 13 Jan 2023 at 6:09pm
Originally posted by copperdogma copperdogma wrote:

I feel it's obvious that getting it to write your entire story defeats the purpose of a writing contest.

I suppose that depends on your reason for joining the challenge. First prize here is a fair amount, enough that it could be the goal for some people. If you join only to win, using AI to write it could be an option.

Having said that, I think it's cheating. If you use AI to write, you're not honing your skill as a writer. It would be like using a ghostwriter, where you tell them the story or outline and they write it for you. The credit for the story should go to the ghostwriter, IMHO.

As for using an AI for non-challenge/contest writing, well, I think it's lazy. The more you write the better you become as a writer. The more you read the more you understand story structure, characterization, and whatnot. Using an AI shortcuts all of that.

I will concede that AI-written work is a new and evolving art form, one worth critiquing and consuming separate from human-written work. I also think that if an AI wrote something, in full or in part, it should be tagged as such so the reader knows the author didn't write it alone.


-------------
http://tinyurl.com/yayxr6x6" rel="nofollow - SS R1 Premium Ingredient
https://tinyurl.com/2p8npsfa" rel="nofollow - ALL NYCM


Posted By: taaaylor
Date Posted: 13 Jan 2023 at 6:27pm
I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 


-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: BarbaraFL
Date Posted: 13 Jan 2023 at 7:22pm
THIS.

I hate the idea of it and people WILL use it for these contests and it's not ok. It's cheating. 

Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 


-------------
Scplay https://tinyurl.com/m8hjvcyr" rel="nofollow - Regrettable Hobby
SStory https://bit.ly/4bdTAWN" rel="nofollow - Be a Good Girl


Posted By: countercheck
Date Posted: 13 Jan 2023 at 8:12pm
I think these tools are here to stay, only going to get better, and railing against them isn't particularly useful. Are they moral to use in competition? If feel sh*tty if I were beaten by a story entirely written by chatgpt, both because it feels unfair and because it means my writing is no better than glorified autocomplete. But if people want to use it for ideation? Sure. 

-------------
https://forums.nycmidnight.com/topic56943_post611970.html#611970" rel="nofollow - R1 250


Posted By: Deschain
Date Posted: 13 Jan 2023 at 8:37pm
[posted this earlier in the original thread]

i wouldn't use it in a competition, even for the small task of coming up with a restaurant name. but also i know enough about it to know there's no way to enforce a ban on its use, so at this point people should be aware that they're potentially competing against AI-assisted writing. anyway i can't really worry about what other people do. i have my own artistic code i follow, what other people do is outside of my control.

In case anyone is curious, this is what ChatGPT did with the Round 1 MF250 prompts I had (ghost story, applying lipstick, "splash"). I gave it just the basic instructions, further instructions could've produced something better...

Once there was a young woman named Mary who lived in an old mansion. She was known for her beauty, but also for her vanity. She would spend hours in front of the mirror, applying makeup, particularly her favorite red lipstick.

One evening, as she was getting ready for a party, she heard a strange noise coming from the bathroom. She ignored it, thinking it was just the pipes, but the noise persisted. As she was about to apply her lipstick, the noise grew louder. Suddenly, the bathroom door flew open and a ghost appeared before her. It was the ghost of a woman who had died in the mansion years ago. She had been betrayed by her lover and had taken her own life, but her spirit had been trapped in the mansion ever since.

The ghost pointed to Mary's lipstick, and in a haunting voice, said "You will die the same way I did, by your own hand."

Mary, terrified, dropped the lipstick and ran out of the room. She never returned to the mansion and never wore red lipstick again.

From that day on, whenever anyone visited the bathroom, they would hear a splash and the ghostly woman's voice warning "Don't make the same mistake I did."



-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: taaaylor
Date Posted: 13 Jan 2023 at 8:52pm
Originally posted by BarbaraFL BarbaraFL wrote:

Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 
THIS.

I hate the idea of it and people WILL use it for these contests and it's not ok. It's cheating. 

 

Absolutely.

I'd take it a step further: people who use ChatGPT to replace writing aren't writers. Writers write. Imagine cheating your way to that level of banal mediocrity. It's becoming a problem with Kindle Unlimited and a-holes stealing money from actual writers by pumping out nonsense ChatGPT books and posting them, thus stealing from the communal money pool for total pages read. It has very real monetary impacts already.

I will never be nice about this šŸ˜‚šŸ˜‚šŸ˜‚ This is the only gate I keep. Wanna be a writer? WRITE. Otherwise you're just a plagiarist with extra steps.


-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: Suave
Date Posted: 13 Jan 2023 at 8:59pm

I am not liking it. Just gave it a try with my prompts from round one. Got it to generate four stories, which it did, though they were all pretty similar in plot, they were different. I can see people using this and then tweaking the results. This is not a good thing for our contests. I enjoy the creative adventure too much to use something like this.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: jennifer.quail
Date Posted: 13 Jan 2023 at 9:18pm
Yeah, this is not only going to be a nightmare for the contests, I already have teacher friends who've gotten cheating assignments they caught out as being AI generated. It's bad enough AIs are causing copyright/trademark issues via how the art generator ones work (please don't use those either) and costing those creators money, this is now a completely lazy way out for a contest. I would even consider it cheating to use it as an idea generator. Yes, there are 'make up a character name' generators, but they don't literally write prose for you. Maybe it's my undergrad uni's very hard-core policy (plagiarism was a single-sanction offense for a first offense: you're expelled) but they're just straight-up immoral to use for any sort of writing you're trying to pass off as your own, right up there with stealing someone else's work.

-------------
250 R1: https://forums.nycmidnight.com/topic56891_post611502.html#611502" rel="nofollow - What Even Is a Mirliton
SSC R1: https://bit.ly/49hAP39" rel="nofollow - Dissolution


Posted By: Deschain
Date Posted: 13 Jan 2023 at 10:33pm
I'll be honest though, I was annoyed at first when I found out people were using betas, but when I realized how common it was I just had to accept it. Everyone has different ideas of competitive fairness. If I have to compete against writers bolstered by betas then that's what I'll do. If I have to compete against writers bolstered by ChatGPT then that's what I'll do. If someone asks ChatGPT to help with something they could otherwise ask a beta for, is it really any different? Personally I think not, but I'm sure other people disagree. Anyway, all I can control is my own process.

-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: NMiller
Date Posted: 13 Jan 2023 at 10:40pm
Originally posted by Deschain Deschain wrote:

If someone asks ChatGPT to help with something they could otherwise ask a beta for, is it really any different? Personally I think not, but I'm sure other people disagree. Anyway, all I can control is my own process.


Beta's (spouses, children, friends, forum spreadsheet volunteers) have helped me to clarify my thoughts, but have never written the story for me. They made sure what is in my head is actually coming across to others clearly. They have helped with grammar and supported or vetoed creative attempts. Ultimately it is my choice to take the advice or ignore it. I am not sure how that would be considered cheating?


-------------
http://tinyurl.com/284w2dv7" rel="nofollow - Blowing It SS
http://tinyurl.com/53sywsdx" rel="nofollow - Tirupur 250
All


Posted By: Suave
Date Posted: 13 Jan 2023 at 10:47pm
Originally posted by NMiller NMiller wrote:

Originally posted by Deschain Deschain wrote:

If someone asks ChatGPT to help with something they could otherwise ask a beta for, is it really any different? Personally I think not, but I'm sure other people disagree. Anyway, all I can control is my own process.


Beta's (spouses, children, friends, forum spreadsheet volunteers) have helped me to clarify my thoughts, but have never written the story for me. They made sure what is in my head is actually coming across to others clearly. They have helped with grammar and supported or vetoed creative attempts. Ultimately it is my choice to take the advice or ignore it. I am not sure how that would be considered cheating?

I have to agree with Mr. Miller. A beta reader is nothing like having some else write a story for you and then claim it as your own. Getting advice on your piece, being told you have used a word too many times in close proximity, add a comma, is not cheating.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Deschain
Date Posted: 13 Jan 2023 at 11:01pm
Originally posted by NMiller NMiller wrote:

Beta's (spouses, children, friends, forum spreadsheet volunteers) have helped me to clarify my thoughts, but have never written the story for me. They made sure what is in my head is actually coming across to others clearly. They have helped with grammar and supported or vetoed creative attempts. Ultimately it is my choice to take the advice or ignore it. I am not sure how that would be considered cheating?

I didn't say it was cheating. It isn't against the rules, so it's not cheating. It's just a different process than mine. But I personally don't see much difference between asking ChatGPT to check your grammar versus asking a beta to check your grammar.

Remember, people (like me) joining the contest for the first time probably haven't even heard of betas, so it's a competitive advantage, particularly for the betas who are also part of NYCM and have knowledge of what judges look for. There isn't anything inherently wrong about competitive advantages. Just competing in a handful of contests is a competitive advantage over newcomers. But at the same time, I don't see a difference between a newcomer asking ChatGPT to check their grammar instead of asking a friend.


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: Deschain
Date Posted: 13 Jan 2023 at 11:08pm
Originally posted by Suave Suave wrote:

I have to agree with Mr. Miller. A beta reader is nothing like having some else write a story for you and then claim it as your own. Getting advice on your piece, being told you have used a word too many times in close proximity, add a comma, is not cheating.

Just to be clear, I am specifically talking about asking ChatGPT to do a task you would otherwise ask a beta. I am not suggesting that having ChatGPT write your story is equivalent to getting advice from a beta.

Also I am not saying/insinuating that using a beta is cheating. It's not against the rules, and therefore it's not cheating.


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: Popsicle Crows
Date Posted: 13 Jan 2023 at 11:34pm
Yeah, beta reader and AI are totally different things, the rules on second opinions are irrelevant.
Betas give an opinion on something created from a writer's mind whereas AI programs do the entire writer's job.
Not even vaguely close to the same thing.


Posted By: Popsicle Crows
Date Posted: 13 Jan 2023 at 11:42pm
On a brighter note, the people behind that app are putting trackers similar to watermarks into the work generated by their app to help foil people who would try to pass off AI work as their own. That's good news.


Posted By: Deschain
Date Posted: 14 Jan 2023 at 12:03am
Originally posted by Popsicle Crows Popsicle Crows wrote:

Yeah, beta reader and AI are totally different things, the rules on second opinions are irrelevant.
Betas give an opinion on something created from a writer's mind whereas AI programs do the entire writer's job.
Not even vaguely close to the same thing.

As I said, I am specifically talking about asking ChatGPT to do a task you'd ask a beta to do.

I am not talking about ChatGPT writing someone's story for them.

I am specifically asking if people think there's a difference between:
(1) asking ChatGPT to check the grammar of a story you wrote
(2) asking a beta to check the grammar of a story you wrote




-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: taaaylor
Date Posted: 14 Jan 2023 at 12:10am
Originally posted by Deschain Deschain wrote:

Originally posted by Popsicle Crows Popsicle Crows wrote:

Yeah, beta reader and AI are totally different things, the rules on second opinions are irrelevant.
Betas give an opinion on something created from a writer's mind whereas AI programs do the entire writer's job.
Not even vaguely close to the same thing.

As I said, I am specifically talking about asking ChatGPT to do a task you'd ask a beta to do.

I am not talking about ChatGPT writing someone's story for them.

I am specifically asking if people think there's a difference between:
(1) asking ChatGPT to check the grammar of a story you wrote
(2) asking a beta to check the grammar of a story you wrote


I'm pretty sure anything you upload to ChatGPT, you're giving it permission to disseminate and reproduce your text as often as possible in whatever means the AI deems fit. I don't think it's cheating, as some people do need external programs to adapt to disability (e.g. Grammarly). However, I think it's highly ill-advised to upload your own work there unless you like giving away unlimited rights for the bot to reproduce your work in perpetuity 



-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: Popsicle Crows
Date Posted: 14 Jan 2023 at 12:10am
I was specifically talking about people using it to do the work for them.

To answer your question though, a fellow writer wouldn't make the mistake of sanitizing someone's writing and a beta reader gives an opinion so you don't have to take their advice/opinion. 
The big story about this AI program isn't about using it as a stand in for WORD MS grammar checker. Most common writing software already has something to check spelling and grammar.


Posted By: Popsicle Crows
Date Posted: 14 Jan 2023 at 12:19am
A professional editor wouldn't trust that program either. No nuance. Much like how bias can creep into AI in most other applications, it's only as good as the things that go into it. So that would be everything from near-illiterate sources to obselete language and styles to very boring to absolutely great. The great writing it learns from would be statistically insignificant otherwise everyone would be a published author. 


Posted By: Deschain
Date Posted: 14 Jan 2023 at 12:52am
Originally posted by Popsicle Crows Popsicle Crows wrote:

I was specifically talking about people using it to do the work for them.

To answer your question though, a fellow writer wouldn't make the mistake of sanitizing someone's writing and a beta reader gives an opinion so you don't have to take their advice/opinion. 
The big story about this AI program isn't about using it as a stand in for WORD MS grammar checker. Most common writing software already has something to check spelling and grammar.

My mistake, I thought you were responding to my post since you were comparing betas and ChatGPT. I didn't realize you were trying to start a separate discussion.

And no, that doesn't answer my question, that answers an unwritten question you just posed to yourself. You can write whatever you want, but please don't attribute it to me. I prefer to write my own words. I'm very picky about them.


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: Popsicle Crows
Date Posted: 14 Jan 2023 at 1:36am
@Deschain
You said: "If I have to compete against writers bolstered by betas..."
They aren't 'bolstered', betas give other writers opinions or advice.

You asked "If someone asks ChatGPT to help with something they could otherwise ask a beta for, is it really any different?"

You posed a question about a situation that isn't plausible so it could only be answered by explaining why the two situations can't be compared. Apples and oranges. I answered WHY it was different.
You are perfectly free to hate my answer :-)



Posted By: Deschain
Date Posted: 14 Jan 2023 at 2:32am

Originally posted by Popsicle Crows Popsicle Crows wrote:

@Deschain

You said: "If I have to compete against writers bolstered by betas..."

They aren't 'bolstered', betas give other writers opinions or advice.

Originally posted by NMiller NMiller wrote:

Betas...have helped me to clarify my thoughts...They made sure what is in my head is actually coming across to others clearly. They have helped with grammar and supported or vetoed creative attempts.

bolster - to support or improve something or make it stronger ( https://dictionary.cambridge.org/us/dictionary/english/bolster" rel="nofollow - https://dictionary.cambridge.org/us/dictionary/english/bolster )

 

***

 

Originally posted by Popsicle Crows Popsicle Crows wrote:

You asked "If someone asks ChatGPT to help with something they could otherwise ask a beta for, is it really any different?"

You posed a question about a situation that isn't plausible...

Originally posted by NMiller NMiller wrote:

Betas...have helped with grammar...

Originally posted by Deschain Deschain wrote:

I am specifically asking if people think there's a difference between:
(1) asking ChatGPT to check the grammar of a story you wrote
(2) asking a beta to check the grammar of a story you wrote



-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: Popsicle Crows
Date Posted: 14 Jan 2023 at 2:45am
Maybe a beta would have helped you clarify your question?

Betas don't do the work of improving or making your stories stronger, they suggest and advise and then the writer decides and does the work.

Same answer to your question. An AI wouldn't be used to help with grammar unless someone wanted to take out the nuance. As NMiller noted, betas can advise other writers about grammar etc and are massively helpful. obviously NMiller did the work, not a computer. 

My answer is still the same, yes, a massive difference. One is part of the normal writing process, the other is a new way to take all the personality out of someone's work by having technology make decisions for you. 
Apples and wheelbarrows.


Posted By: Deschain
Date Posted: 14 Jan 2023 at 2:55am
Originally posted by Popsicle Crows Popsicle Crows wrote:

Maybe a beta would have helped you clarify your question?

Are you offering to help clarify my question? I didn't realize it was so confusing. Thank you so much!

Originally posted by Deschain Deschain wrote:

I am specifically asking if people think there's a difference between:
(1) asking ChatGPT to check the grammar of a story you wrote
(2) asking a beta to check the grammar of a story you wrote



(this was fun, let's do it again sometime Wink)


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: Popsicle Crows
Date Posted: 14 Jan 2023 at 3:01am
I wasn't but as you showed in your quote, a beta reader could whereas ChatGPT might make it worse.
I found your question very clear, that's why it was so easy to answer. 

And definitely up for a debate any time. It's fun.
Night, sleep well.



Posted By: BarbaraFL
Date Posted: 14 Jan 2023 at 6:13pm
I just asked it to write a 250-word micro story about my dog (I am not in the competition) just out of curiosity...and it killed off my dog LOL

-------------
Scplay https://tinyurl.com/m8hjvcyr" rel="nofollow - Regrettable Hobby
SStory https://bit.ly/4bdTAWN" rel="nofollow - Be a Good Girl


Posted By: Suave
Date Posted: 14 Jan 2023 at 8:38pm
Originally posted by BarbaraFL BarbaraFL wrote:

I just asked it to write a 250-word micro story about my dog (I am not in the competition) just out of curiosity...and it killed off my dog LOL

Ask it to do it again and see what you get.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: copperdogma
Date Posted: 15 Jan 2023 at 12:56am
Originally posted by taaaylor taaaylor wrote:

Originally posted by BarbaraFL BarbaraFL wrote:

Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 
THIS.

I hate the idea of it and people WILL use it for these contests and it's not ok. It's cheating. 

 

Absolutely.

I'd take it a step further: people who use ChatGPT to replace writing aren't writers. Writers write. Imagine cheating your way to that level of banal mediocrity. It's becoming a problem with Kindle Unlimited and a-holes stealing money from actual writers by pumping out nonsense ChatGPT books and posting them, thus stealing from the communal money pool for total pages read. It has very real monetary impacts already.

I will never be nice about this šŸ˜‚šŸ˜‚šŸ˜‚ This is the only gate I keep. Wanna be a writer? WRITE. Otherwise you're just a plagiarist with extra steps.

I hear you. And your original posts. And I'm not sure I can disagree with you anywhere.

From chatting with my other writer friends, we posed this question: If you read a story and loved it, what would change if you were then told it was written by AI?

The answer we settled on was something like: Reading a story from a human being feels like that person is trying to say something to you. They have intent and purpose with their story, and you have a connection with them because you've read it. A story written by AI loses all of that.

But regardless of our feelings about it, this is the future, and it's here now.

Currently ChatGPT, from my experiments, writes terrible stories;) But this stuff if advancing rapidly, so not being worried yet would be like standing in the rain before an approaching hurricane and saying "it's no big deal -- just a drizzle."

This will be the future. I wanted to ask the question about ChatGPT to get people in a community I love thinking about it and the ramifications for the future.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: BarbaraFL
Date Posted: 15 Jan 2023 at 6:49am
Originally posted by Suave Suave wrote:

Originally posted by BarbaraFL BarbaraFL wrote:

I just asked it to write a 250-word micro story about my dog (I am not in the competition) just out of curiosity...and it killed off my dog LOL

Ask it to do it again and see what you get.

I asked it why it was so mean to kill off my dog in the story and it apologized


-------------
Scplay https://tinyurl.com/m8hjvcyr" rel="nofollow - Regrettable Hobby
SStory https://bit.ly/4bdTAWN" rel="nofollow - Be a Good Girl


Posted By: Suave
Date Posted: 15 Jan 2023 at 6:53am
Originally posted by BarbaraFL BarbaraFL wrote:

Originally posted by Suave Suave wrote:

Originally posted by BarbaraFL BarbaraFL wrote:

I just asked it to write a 250-word micro story about my dog (I am not in the competition) just out of curiosity...and it killed off my dog LOL

Ask it to do it again and see what you get.

I asked it why it was so mean to kill off my dog in the story and it apologized

Haha, that is hilarious!


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Popsicle Crows
Date Posted: 15 Jan 2023 at 9:27am
Originally posted by BarbaraFL BarbaraFL wrote:

Originally posted by Suave Suave wrote:

Originally posted by BarbaraFL BarbaraFL wrote:

I just asked it to write a 250-word micro story about my dog (I am not in the competition) just out of curiosity...and it killed off my dog LOL

Ask it to do it again and see what you get.

I asked it why it was so mean to kill off my dog in the story and it apologized


OMG, it kicked a puppy. The ultimate bad guy move. LOL

Can you ask it if it can self-destruct?
And again, Trey's story '177 Seconds' wins.



Posted By: copperdogma
Date Posted: 15 Jan 2023 at 1:51pm
My wife is a university prof and I've stressed her out by telling her about ChatGTP. She's now worried her students will be using it to cheat.

She did some research and found this ChatGTP detector:  https://openai-openai-detector.hf.space/" rel="nofollow - https://openai-openai-detector.hf.space/

I tried it on a dozen things that I wrote or I had it write, and it detected the generated stories 100% of the time.

This article, though, says it's only about 66% accurate:  https://www.zdnet.com/article/can-ai-detectors-save-us-from-chatgpt-i-tried-3-online-tools-to-find-out/" rel="nofollow - https://www.zdnet.com/article/can-ai-detectors-save-us-from-chatgpt-i-tried-3-online-tools-to-find-out/

Which is still pretty good.

The issue will be that AI and the way these models are trained are something of an arms race. Almost in a literal sense. One popular way of training them is antagonistic training. Basically you have two AIs, one generating content and the other checking that content and rating how close to the goal it is. All an AI company would have to do is use the ChatGPT detector as the antagonist in its training, which would cause it's generating AI to create more and more realistic writing until the detector was consistently rating the generated story as "100% human made."

So the detector will be good for... Oh.. I'd guess a year, max?


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: Popsicle Crows
Date Posted: 15 Jan 2023 at 1:59pm
Originally posted by copperdogma copperdogma wrote:

My wife is a university prof and I've stressed her out by telling her about ChatGTP. She's now worried her students will be using it to cheat.

She did some research and found this ChatGTP detector:  https://openai-openai-detector.hf.space/" rel="nofollow - https://openai-openai-detector.hf.space/

I tried it on a dozen things that I wrote or I had it write, and it detected the generated stories 100% of the time.

This article, though, says it's only about 66% accurate:  https://www.zdnet.com/article/can-ai-detectors-save-us-from-chatgpt-i-tried-3-online-tools-to-find-out/" rel="nofollow - https://www.zdnet.com/article/can-ai-detectors-save-us-from-chatgpt-i-tried-3-online-tools-to-find-out/

Which is still pretty good.

The issue will be that AI and the way these models are trained are something of an arms race. Almost in a literal sense. One popular way of training them is antagonistic training. Basically you have two AIs, one generating content and the other checking that content and rating how close to the goal it is. All an AI company would have to do is use the ChatGPT detector as the antagonist in its training, which would cause it's generating AI to create more and more realistic writing until the detector was consistently rating the generated story as "100% human made."

So the detector will be good for... Oh.. I'd guess a year, max?
It looks like the creators of the program are going to put something similar to a watermark into it so people can't use it for cheating.


Posted By: laleighda
Date Posted: 15 Jan 2023 at 2:02pm
Thanks for posting that tool! Writing is one of the greatest skills a person can have and I try to promote it in my classes (I teach at a state college, criminal justice). I've gotten good at detecting when someone copies and pastes from a website, but am happy to have a tool that'll help detect AI.

Thanks again!


Posted By: copperdogma
Date Posted: 15 Jan 2023 at 2:06pm
Originally posted by Popsicle Crows Popsicle Crows wrote:

Originally posted by copperdogma copperdogma wrote:

My wife is a university prof and I've stressed her out by telling her about ChatGTP. She's now worried her students will be using it to cheat.

She did some research and found this ChatGTP detector:  https://openai-openai-detector.hf.space/" rel="nofollow - https://openai-openai-detector.hf.space/

I tried it on a dozen things that I wrote or I had it write, and it detected the generated stories 100% of the time.

This article, though, says it's only about 66% accurate:  https://www.zdnet.com/article/can-ai-detectors-save-us-from-chatgpt-i-tried-3-online-tools-to-find-out/" rel="nofollow - https://www.zdnet.com/article/can-ai-detectors-save-us-from-chatgpt-i-tried-3-online-tools-to-find-out/

Which is still pretty good.

The issue will be that AI and the way these models are trained are something of an arms race. Almost in a literal sense. One popular way of training them is antagonistic training. Basically you have two AIs, one generating content and the other checking that content and rating how close to the goal it is. All an AI company would have to do is use the ChatGPT detector as the antagonist in its training, which would cause it's generating AI to create more and more realistic writing until the detector was consistently rating the generated story as "100% human made."

So the detector will be good for... Oh.. I'd guess a year, max?
It looks like the creators of the program are going to put something similar to a watermark into it so people can't use it for cheating.

I heard about that! I'm a programmer by trade and I was having trouble imagining how they'd do that. This article has a good explainer:

https://theamericangenius.com/tech-news/chatgpt-watermark/#:~:text=You%20can't%20currently%20use,stage%20of%20ChatGPT%20feature%20rollouts.

TLDR: Cryptogenic watermarking. "The key to making AI content watermarking virtually undetectable is ensuring the distribution of words still has a unique and random appearance like regular AI-generated text."

That's interesting. If they do it consistently enough within the text, anything you don't rewrite fully by hand would still contain that distribution ("watermark"), so the more you leave unchanged the higher it would rate on something detecting generated text.

That actually sounds workable.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: Popsicle Crows
Date Posted: 15 Jan 2023 at 2:12pm
Originally posted by copperdogma copperdogma wrote:


I tried it on a dozen things that I wrote or I had it write, and it detected the generated stories 100% of the tiI heard about that! I'm a programmer by trade and I was having trouble imagining how they'd do that. This article has a good explainer:
https://theamericangenius.com/tech-news/chatgpt-watermark/#:~:text=You%20can" rel="nofollow - https://theamericangenius.com/tech-news/chatgpt-watermark/#:~:text=You%20can 't%20currently%20use,stage%20of%20ChatGPT%20feature%20rollouts.

TLDR: Cryptogenic watermarking. "The key to making AI content watermarking virtually undetectable is ensuring the distribution of words still has a unique and random appearance like regular AI-generated text."

That's interesting. If they do it consistently enough within the text, anything you don't rewrite fully by hand would still contain that distribution ("watermark"), so the more you leave unchanged the higher it would rate on something detecting generated text.

That actually sounds workable.
It does! Fingers crossed that they make it happen. Sounds like they want to be successful creatives instead of the bad guys that pisses off people. Better for them to improve their own program and be the heroes.


Posted By: Deschain
Date Posted: 15 Jan 2023 at 2:40pm
Originally posted by copperdogma copperdogma wrote:

From chatting with my other writer friends, we posed this question: If you read a story and loved it, what would change if you were then told it was written by AI?

It's upsetting whenever you read something that you feel emotionally invested in and then learn you were deceived by the author. I imagine this would feel the same as when you read something labeled as a memoir or autobiography and then learn the story was fabricated.

However, if I know ahead of time that a story was generated by AI, and I enjoy it, then I guess I enjoy it.


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: copperdogma
Date Posted: 15 Jan 2023 at 2:43pm
Originally posted by Deschain Deschain wrote:

Originally posted by copperdogma copperdogma wrote:

From chatting with my other writer friends, we posed this question: If you read a story and loved it, what would change if you were then told it was written by AI?

It's upsetting whenever you read something that you feel emotionally invested in and then learn you were deceived by the author. I imagine this would feel the same as when you read something labeled as a memoir or autobiography and then learn the story was fabricated.

However, if I know ahead of time that a story was generated by AI, and I enjoy it, then I guess I enjoy it.

I know what you mean. I've been applying that same question to other things. What if I knew a video game I loved was mostly created by AI? What about a washing machine? An instruction manual? A piece of furniture? The answer changes in strange and subtle ways with each thing.

I feel like we're going to spend a LOT of time in the (near) future grappling with these questions on an ever-increasing cadence.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: taaaylor
Date Posted: 15 Jan 2023 at 3:32pm
Originally posted by copperdogma copperdogma wrote:

I hear you. And your original posts. And I'm not sure I can disagree with you anywhere.

From chatting with my other writer friends, we posed this question: If you read a story and loved it, what would change if you were then told it was written by AI?

I cannot agree with the fundamental premise that I would ever like something written by an AI. Particularly creative writing. Unlike art, there are NO objective ratios or formulae by which to measure if something is "right" in creative writing. And even then, the art AIs are still trash compared to a competent artist. It's highly telling that ChatGPT can only semi-succesfully reproduce nonfiction text or code. Nonfiction is consistent, if you write like a dry piece of cardboard. It has a formula that can be derived from millions of similar content articles online.

Creative writing is fundamentally incomparable.

Originally posted by copperdogma copperdogma wrote:

Currently ChatGPT, from my experiments, writes terrible stories;) But this stuff if advancing rapidly, so not being worried yet would be like standing in the rain before an approaching hurricane and saying "it's no big deal -- just a drizzle."

This will be the future. I wanted to ask the question about ChatGPT to get people in a community I love thinking about it and the ramifications for the future.


1) Of course I'm worried. Like many visual artists found after Midjourney AI went big, we are relying on the honor system of the people who built the bot and the people who trained it by uploading content. There are artists who never agreed to have their art included in the learning algorithm who have now been included. (I would not have used Midjourney if I knew that; I naively trusted their claim that all images it trained on were in the public domain.) I have 500k+ words available for free on my Reddit writing account. I have no doubt that my work will be stolen for this endeavor, to train robots to try to replace writers like me.

However, I am not worried about being outwritten by a robot. I am worried about how to protect my IP in this era without damaging my readers or my own success. It will be extremely hard to share writing for free on indexable searches in the next 10 years. I've already had my writing stolen over and over by TikTok or Youtube narrators. This is just the next iteration of lazy art thieves. f**k them all. No respect.

2) As long as quantum computing is only a pipedream, ChatGPT will never be as good as a human creative writer who's been writing for more than about five minutes. Full stop. The simple technical limits of the algorithm + the infinite capacity of the human mind make it no-contest.


-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: Deschain
Date Posted: 15 Jan 2023 at 4:04pm
Originally posted by copperdogma copperdogma wrote:

I know what you mean. I've been applying that same question to other things. What if I knew a video game I loved was mostly created by AI? What about a washing machine? An instruction manual? A piece of furniture? The answer changes in strange and subtle ways with each thing.

I feel like we're going to spend a LOT of time in the (near) future grappling with these questions on an ever-increasing cadence.

of course, we might also be looking at emerging artificial consciousness. so at some point this could go from stories written by advanced language models, to stories written by something that believes it's a sentient being. 

personally, I'm not that hung up on what created something, as long as I know what it is. this isn't a new issue, just the new iteration of an old issue. musicians had the same problem with synthesizers. painters had the same fears about cameras.

I'm sure somebody who doesn't understand how debate works might come along here and list differences between AI and cameras and synthesizers and try to argue that these are different scenarios. but the truth is writers are feeling threatened by a new technology, and i think their fears are understandable, but personally it doesn't affect my process and I'm open to the new possibilities it will open up for other people artistically. 


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: copperdogma
Date Posted: 15 Jan 2023 at 4:14pm
Originally posted by Deschain Deschain wrote:

Originally posted by copperdogma copperdogma wrote:

I know what you mean. I've been applying that same question to other things. What if I knew a video game I loved was mostly created by AI? What about a washing machine? An instruction manual? A piece of furniture? The answer changes in strange and subtle ways with each thing.

I feel like we're going to spend a LOT of time in the (near) future grappling with these questions on an ever-increasing cadence.

of course, we might also be looking at emerging artificial consciousness. so at some point this could go from stories written by advanced language models, to stories written by something that believes it's a sentient being. 

personally, I'm not that hung up on what created something, as long as I know what it is. this isn't a new issue, just the new iteration of an old issue. musicians had the same problem with synthesizers. painters had the same fears about cameras.

I'm sure somebody who doesn't understand how debate works might come along here and list differences between AI and cameras and synthesizers and try to argue that these are different scenarios. but the truth is writers are feeling threatened by a new technology, and i think their fears are understandable, but personally it doesn't affect my process and I'm open to the new possibilities it will open up for other people artistically. 

This is a good point. And I like your analogies to past human vs machine advancements.

One thing I like about all of this is it makes us ask WHY we write in the first place. The answer, I think, informs a lot of how we react to it.

If writing is our livelihood then AI poses a fundamental risk to our existence.

If writing is our passion, maybe AI doesn't make any difference as we'd do it anyway.

If we're writing to improve, maybe AI doesn't make any difference.

If we're writing for external validation, AI poses competition.

I'm sure most people overlap somewhere in there and there are 1000 other variations on this.

Personally, I write here because I want to write a book and I want to be more skilled before I attempt it. Also I want to use any accolades from my stories here to try to get a publisher more interested. My end goal with the book I want to write it to help as many people as possible. If AI beats me to it and helps the world in my stead, I'll be sad I've lost a project that means a lot to me, but my goal has been met and the world is (hopefully) a better place because of it.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: Deschain
Date Posted: 15 Jan 2023 at 4:23pm
yeah I'm in the same boat with not aspiring to be a professional writer, but the professional writers are definitely feeling threatened right now and I understand why. in general, people are afraid of technology replacing them, but I think what really ends up happening is the technology opens up new possibilities for people artistically. there are still painters, but now there are movies thanks to cameras. 

but a big part is honesty. and when it comes to that I stand with my fellow artists in saying it really is absolutely wrong to deceive people with it. I don't want to read AI generated news articles, or things like that. 


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: copperdogma
Date Posted: 15 Jan 2023 at 4:33pm
Originally posted by taaaylor taaaylor wrote:

I cannot agree with the fundamental premise that I would ever like something written by an AI. Particularly creative writing. Unlike art, there are NO objective ratios or formulae by which to measure if something is "right" in creative writing. And even then, the art AIs are still trash compared to a competent artist. It's highly telling that ChatGPT can only semi-succesfully reproduce nonfiction text or code. Nonfiction is consistent, if you write like a dry piece of cardboard. It has a formula that can be derived from millions of similar content articles online.

Creative writing is fundamentally incomparable.

Originally posted by copperdogma copperdogma wrote:

Currently ChatGPT, from my experiments, writes terrible stories;) But this stuff if advancing rapidly, so not being worried yet would be like standing in the rain before an approaching hurricane and saying "it's no big deal -- just a drizzle."

This will be the future. I wanted to ask the question about ChatGPT to get people in a community I love thinking about it and the ramifications for the future.


Originally posted by taaaylor taaaylor wrote:

1) Of course I'm worried. Like many visual artists found after Midjourney AI went big, we are relying on the honor system of the people who built the bot and the people who trained it by uploading content. There are artists who never agreed to have their art included in the learning algorithm who have now been included. (I would not have used Midjourney if I knew that; I naively trusted their claim that all images it trained on were in the public domain.) I have 500k+ words available for free on my Reddit writing account. I have no doubt that my work will be stolen for this endeavor, to train robots to try to replace writers like me.

However, I am not worried about being outwritten by a robot. I am worried about how to protect my IP in this era without damaging my readers or my own success. It will be extremely hard to share writing for free on indexable searches in the next 10 years. I've already had my writing stolen over and over by TikTok or Youtube narrators. This is just the next iteration of lazy art thieves. f**k them all. No respect.

2) As long as quantum computing is only a pipedream, ChatGPT will never be as good as a human creative writer who's been writing for more than about five minutes. Full stop. The simple technical limits of the algorithm + the infinite capacity of the human mind make it no-contest.

I'm fascinated by your take on all of this.

Also, as a side note, I've read a few of your stories on here in the past. You're a brilliant writer! Thank you for sharing. You inspire me to be better.

You may be right with your assumption that an AI can never be better than a human at creative writing, but I'm not sure about that myself.

First, I'd say Midjourney can definitely outperform, both in terms of execution AND creativity, a lot of artists at this point. No the best artists. And it still, I think, it's main issue is that is has nothing to say.

Second, and this addresses the Midjourney issue itself, is that AI isn't done evolving. It's just barely begun. Tech tends to follow an S-curve of evolution, and AI is about to go vertical with its capabilities. It might be hubris to assume an AI can never outperform humans in every arena.

I don't see it being very hard to add something to an AI generation pipeline to ensure it has something to SAY with everything it creates. And it would be trivial to add something that ensures novelty. It's already ingested most of the internet. It can't be hard to tell it that when generating something, ensure it's sufficiently different from anything else it's ever seen.

Now you have an AI that has something fascinating to say and it's saying it in a compelling way that's gripping to read. I'd guess we're going to see that a year from now? Maybe two.

One of your main points seems to be the unfairness/illegality of repurposing others' work. I totally agree that plagiarism is a big issue here, and it's grossly unfair to take the work of someone else and use it as your own.

I think it's more tricky when it comes to AIs. As humans, we slowly absorb the collective works of everything we've ever experienced, and we synthesize something new from that. If we're bad at it, it's derivative. If we're good, it's insightful and novel. This is basically how these AIs work. They're trained on a massive body of work, but they synthesize their own new works from that, much as we do.

Yes, you can abuse it. But the AIs are inherently ammoral (although ChatGTP is putting in moral guardrails). It's no different than asking an unscrupulous writer to write a story in the exact style of Stephen King or a paining in the style of Van Gough. It's not original. It's essentially a form of theft. You can ask an AI to do that and it will comply. But if you ask it to create something novel, it will, and it does it much like we do: create something new from its vast experience of what is. Currently it's often bland cardboard and generically derivative from all previous works, but that's going to change quickly.

But maybe I'm not addressing your core points of humans vs. AI.

Personally I hope you're right. I hope AI can never replace what makes us fundamentally human. That there's something at our core that allows us to create at a level an AI can never match. Or maybe, like bespoke furniture, humans who can afford it will reject things not created by other human beings purely on aesthetic/spiritual grounds.







-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: copperdogma
Date Posted: 15 Jan 2023 at 4:41pm
Originally posted by Deschain Deschain wrote:

yeah I'm in the same boat with not aspiring to be a professional writer, but the professional writers are definitely feeling threatened right now and I understand why. in general, people are afraid of technology replacing them, but I think what really ends up happening is the technology opens up new possibilities for people artistically. there are still painters, but now there are movies thanks to cameras. 

but a big part is honesty. and when it comes to that I stand with my fellow artists in saying it really is absolutely wrong to deceive people with it. I don't want to read AI generated news articles, or things like that. 

I think they're right to fear it. I have a few writer friends who are pretty worried, and I know some people who manage writers who are considering getting rid of them or not hiring more.

What we do here as storytellers is, currently, an extraordinarily creative and human version of writing. Stories are meant to inspire, provoke thought, or bring someone to tears with emotion.

But 90% of human writing is.. vanilla. FAQs and CEO profiles and blog posts for SEO. It's not worth the expense to make it better. It's quantity over quality. These peoples' jobs are threatened today.

But that's often how technology works: it slowly eats away at the dull, dirty, dangerous jobs of the world. In the worst case, people lose their livelihoods. In the best case, they get far more rewarding jobs doing something that requires more of their unique humanity to accomplish.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: laleighda
Date Posted: 15 Jan 2023 at 4:54pm
Originally posted by copperdogma copperdogma wrote:

Originally posted by Deschain Deschain wrote:

yeah I'm in the same boat with not aspiring to be a professional writer, but the professional writers are definitely feeling threatened right now and I understand why. in general, people are afraid of technology replacing them, but I think what really ends up happening is the technology opens up new possibilities for people artistically. there are still painters, but now there are movies thanks to cameras. 

but a big part is honesty. and when it comes to that I stand with my fellow artists in saying it really is absolutely wrong to deceive people with it. I don't want to read AI generated news articles, or things like that. 

I think they're right to fear it. I have a few writer friends who are pretty worried, and I know some people who manage writers who are considering getting rid of them or not hiring more.

What we do here as storytellers is, currently, an extraordinarily creative and human version of writing. Stories are meant to inspire, provoke thought, or bring someone to tears with emotion.

But 90% of human writing is.. vanilla. FAQs and CEO profiles and blog posts for SEO. It's not worth the expense to make it better. It's quantity over quality. These peoples' jobs are threatened today.

But that's often how technology works: it slowly eats away at the dull, dirty, dangerous jobs of the world. In the worst case, people lose their livelihoods. In the best case, they get far more rewarding jobs doing something that requires more of their unique humanity to accomplish.

 

Very interesting take. I think all of us should be prepared for the role of automation and AI, including those in the creative field. It is coming, or is already here. It can be scary, but think of all the technical tools we do use - I am 100% reliant on spell check, and slightly less so on thesaurus. I am in a writing group, and someone was scolded for using a rhyme generator, but that is a technology that exists, and the person using it still wrote his poems, but was just helped a little bit by the tool. 

What's the line in the sand? Seems to be different for people.

Side note - @Deschain is your SN in reference to Roland Deschain? 


Posted By: Deschain
Date Posted: 15 Jan 2023 at 8:11pm
Originally posted by laleighda laleighda wrote:


Side note - @Deschain is your SN in reference to Roland Deschain? 

it is! reading the Dark Tower series throughout quarantine is what inspired me to start writing again. I don't actually like horror so I had never read King before, but that series was amazing and it really inspired me to write without imposing so many rules on myself as a storyteller. so the screen name is my reminder to write fearlessly...


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: laleighda
Date Posted: 15 Jan 2023 at 8:33pm
Yesssss! Love seeing other Dark Tower fans. Write fearlessly - I love it! Solid advice that I try and heed as well :)


Posted By: Suave
Date Posted: 15 Jan 2023 at 9:23pm

I have always written because I want to create some of the stuff that
as a kid I read - and I read a lot. I have no doubt that AI will eventually
equal anything that a writer can produce - I say this because I can imagine
how I would tweak the AI as I kept improving the program. I can imagine 
telling my AI I want a story that uses different elements of a few authors
style, not content, pinpointing what I wanted incorporated - damn, I just
came up with a plot for a Sci-Fi. Of course, there is a lot of work to be done
in the ingestion of those authors writings into the AI. Training the AI to recognize:
style, plot development, word choices, thought process connecting all that.
If I can write a 400 page book incorporating all that, then I can make a AI that can
as well - well, cept I am not a programmer, but I can imagine it so it can be
done.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: countercheck
Date Posted: 16 Jan 2023 at 9:12am
I think it's important to recognize that the AI ChatGPT is built on is probably a dead end. It's not an emerging consciousness. It has no awareness of what it writes. It doesn't understand the concept of truth, or logic, and what's more, can't meaningfully interact with those concepts. It's just amped up autocomplete. At best, it's going to become a module in a more capable network of systems.

This is part of what makes it so cool as a utility and so terrifying as a source of misinformation. It doesn't tell truth or lies, it just spews endless streams of bullsh*t on a theme.

If you want to create an endless stream of bullsh*t for generating SEO maximizing websites or flooding forums and twitter, then it's your thing. If you want to generate weird and wonderful combinations of things, it can also do that. 

The worryingly part of it, from a professional writing standpoint, if it has a large enough sample, it can do it in your voice.


-------------
https://forums.nycmidnight.com/topic56943_post611970.html#611970" rel="nofollow - R1 250


Posted By: Deschain
Date Posted: 16 Jan 2023 at 9:50am
i agree with much of what you're saying CC. emerging consciousness, as far as anybody can tell, requires massively parallel hardware, not just software. Chat GPT is just another small step towards that possibility.

but like a lot of new tools, I do think there are potential ways people could involve this in their creative process, or that it could contribute to culture. for instance, there are voices who would probably never be able to write a memoir. ChatGPT (or something like it), with its learning abilities and unlimited patience, could perhaps enable someone with Down Syndrome write a memoir just by asking the person questions. or another possibility, I don't think ChatGPT is on a level that it would be capable of learning a new language just by interacting with somebody of that language, but when AI reaches that point, with so many unique languages and cultures and danger of disappearing completely, and not nearly enough linguists and ethnographers out there to save them all, something like this could be used to capture and preserve entire cultures, folklores, and histories before they disappear.


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: countercheck
Date Posted: 16 Jan 2023 at 10:35am
Originally posted by Deschain Deschain wrote:

i agree with much of what you're saying CC. emerging consciousness, as far as anybody can tell, requires massively parallel hardware, not just software. Chat GPT is just another small step towards that possibility.

Yeah. I'm not even sure it's really a step towards that. It's a neat tool, and I could see it as a linguistics module for another AI - that's basically how they built Clausewitz - one AI to figure out game strategy, and another one to communicate with the other players.

Originally posted by Deschain Deschain wrote:

but like a lot of new tools, I do think there are potential ways people could involve this in their creative process, or that it could contribute to culture. for instance, there are voices who would probably never be able to write a memoir. ChatGPT (or something like it), with its learning abilities and unlimited patience, could perhaps enable someone with Down Syndrome write a memoir just by asking the person questions.

Yes. It absolutely could democratize language generation. I've been thinking pedagogically, it could also help students learn differences in style and tone by converting text from one style or voice into another. Or help people who can't write in an approved, prestigious voice to gain access to privilege that might otherwise be out of reach i.e. letting a young artist used to writing on twitter to produce a reasonable grant application.



Originally posted by Deschain Deschain wrote:

or another possibility, I don't think ChatGPT is on a level that it would be capable of learning a new language just by interacting with somebody of that language, but when AI reaches that point, with so many unique languages and cultures and danger of disappearing completely, and not nearly enough linguists and ethnographers out there to save them all, something like this could be used to capture and preserve entire cultures, folklores, and histories before they disappear.

This is both... plausible and not. Currently, GPT-Chat can't learn from interactions, and it's actually quite dangerous to taint your training set with new data. Phonetic transcription would probably be more useful. That being said:  https://maximumeffort.substack.com/p/i-taught-chatgpt-to-invent-a-language" rel="nofollow - https://maximumeffort.substack.com/p/i-taught-chatgpt-to-invent-a-language is a FASCINATING article. To be clear, it's not actually LEARNING here, I think it's just adding text to a dialogue and rereading it to generate more text each post.


-------------
https://forums.nycmidnight.com/topic56943_post611970.html#611970" rel="nofollow - R1 250


Posted By: Popsicle Crows
Date Posted: 16 Jan 2023 at 12:00pm
It has a use!
The comedians in this show used it to 'write' something in the voice of Trump, Mary Poppins and more lol.
https://www.cbc.ca/radio/becausenews/comedian-ali-hassan-takes-the-news-quiz-1.6713577" rel="nofollow - Comedian Ali Hassan takes the news quiz | CBC Radio
Near the start of the show. 
If I ever used it, I would ask it to rewrite today's news is the voice of Ozzy Osbourne. LOL
Someone also fed it famous works and it made them pretty horrible, The Raven etc.




Posted By: copperdogma
Date Posted: 16 Jan 2023 at 3:15pm
Originally posted by countercheck countercheck wrote:

I think it's important to recognize that the AI ChatGPT is built on is probably a dead end. It's not an emerging consciousness. It has no awareness of what it writes. It doesn't understand the concept of truth, or logic, and what's more, can't meaningfully interact with those concepts. It's just amped up autocomplete. At best, it's going to become a module in a more capable network of systems.

This is part of what makes it so cool as a utility and so terrifying as a source of misinformation. It doesn't tell truth or lies, it just spews endless streams of bullsh*t on a theme.

If you want to create an endless stream of bullsh*t for generating SEO maximizing websites or flooding forums and twitter, then it's your thing. If you want to generate weird and wonderful combinations of things, it can also do that. 

The worryingly part of it, from a professional writing standpoint, if it has a large enough sample, it can do it in your voice.

I'm not convinced it's a dead end. It's not aware of what it writes, but it seems the more parameters it's trained with (ChatGPT = 175 billion) directly correlates to how "smart" it is. There's a Chinese model that's already been trained on 175 TRILLION.

And I'd wager ChatGPT, overall, is about as "smart" as an average person on most tasks, and this is very early days of a very rapid phase of improvement.

Ideally you're right and it can never be truly conscious. Both to preserve that special spark for us humans and to ensure it doesn't decide to go Terminator on us. Ideally it would end up an extraordinarily capable assistant.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: copperdogma
Date Posted: 16 Jan 2023 at 3:22pm
Originally posted by Deschain Deschain wrote:

i agree with much of what you're saying CC. emerging consciousness, as far as anybody can tell, requires massively parallel hardware, not just software. Chat GPT is just another small step towards that possibility.

but like a lot of new tools, I do think there are potential ways people could involve this in their creative process, or that it could contribute to culture. for instance, there are voices who would probably never be able to write a memoir. ChatGPT (or something like it), with its learning abilities and unlimited patience, could perhaps enable someone with Down Syndrome write a memoir just by asking the person questions. or another possibility, I don't think ChatGPT is on a level that it would be capable of learning a new language just by interacting with somebody of that language, but when AI reaches that point, with so many unique languages and cultures and danger of disappearing completely, and not nearly enough linguists and ethnographers out there to save them all, something like this could be used to capture and preserve entire cultures, folklores, and histories before they disappear.

I love your ideas of using it to help someone with Down Syndrome or learning everything about a dying culture. I can't stop thinking about this thing for exactly those sorts of reasons.

Today it's essential as capable as an average person, but one that can do tasks that would take hours, days, or years in an instant. It doesn't get tired or impatient, even with my ridiculous questions on topics I know nothing about. I just read a post about someone who has severe social anxiety using it as a very patient tutor and helping him form questions to real people that he didn't have the confidence to try on his now.

I think it could learn a new language, though. Not in its current form -- it doesn't learn anything from chatting with people in its current form. But I think it could learn new languages just from chatting with someone in a language it doesn't know. It already has an abstracted understanding of language and it's apparently one of the best translators out there already. I bet it could pick up a new language. It learns them currently by just ingesting enough text in a new language, so chatting with people would be exactly that but slower.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: countercheck
Date Posted: 16 Jan 2023 at 3:44pm
Originally posted by copperdogma copperdogma wrote:

Originally posted by countercheck countercheck wrote:

I think it's important to recognize that the AI ChatGPT is built on is probably a dead end. It's not an emerging consciousness. It has no awareness of what it writes. It doesn't understand the concept of truth, or logic, and what's more, can't meaningfully interact with those concepts. It's just amped up autocomplete. At best, it's going to become a module in a more capable network of systems.

This is part of what makes it so cool as a utility and so terrifying as a source of misinformation. It doesn't tell truth or lies, it just spews endless streams of bullsh*t on a theme.

If you want to create an endless stream of bullsh*t for generating SEO maximizing websites or flooding forums and twitter, then it's your thing. If you want to generate weird and wonderful combinations of things, it can also do that. 

The worryingly part of it, from a professional writing standpoint, if it has a large enough sample, it can do it in your voice.

I'm not convinced it's a dead end. It's not aware of what it writes, but it seems the more parameters it's trained with (ChatGPT = 175 billion) directly correlates to how "smart" it is. There's a Chinese model that's already been trained on 175 TRILLION.

And I'd wager ChatGPT, overall, is about as "smart" as an average person on most tasks, and this is very early days of a very rapid phase of improvement.

Ideally you're right and it can never be truly conscious. Both to preserve that special spark for us humans and to ensure it doesn't decide to go Terminator on us. Ideally it would end up an extraordinarily capable assistant.

It's not a matter of size, it's a matter of structure. It's only able to replicate language it's been trained on. This is why it's good at simple arithmetic, but bad at adding big numbers - it doesn't understand math, it just knows that the next character in the sequence "2+2=" is "4". It's not adding, it's completing the phrase. If you tell it "My cow has died. How long before it is alive again?" it'll spout something about the gestational period of cows, not "sorry, cow necromancy is impossible."

It's REALLY REALLY good at completing phrases - this is what lets it write pretty good limericks and passable sonnets. The more parameters you add and the more hidden layers you include in the deep network, the better it'll get at what it already does well - and to be clear, it is ASTONISHINGLY good. It's going to seriously affect a lot of industries.

I wouldn't call it smart exactly. But like I said, once they start linking together multiple systems that are good at different things, then watch out.


-------------
https://forums.nycmidnight.com/topic56943_post611970.html#611970" rel="nofollow - R1 250


Posted By: copperdogma
Date Posted: 16 Jan 2023 at 4:22pm
Originally posted by countercheck countercheck wrote:

Originally posted by copperdogma copperdogma wrote:

I'm not convinced it's a dead end. It's not aware of what it writes, but it seems the more parameters it's trained with (ChatGPT = 175 billion) directly correlates to how "smart" it is. There's a Chinese model that's already been trained on 175 TRILLION.

And I'd wager ChatGPT, overall, is about as "smart" as an average person on most tasks, and this is very early days of a very rapid phase of improvement.

Ideally you're right and it can never be truly conscious. Both to preserve that special spark for us humans and to ensure it doesn't decide to go Terminator on us. Ideally it would end up an extraordinarily capable assistant.

It's not a matter of size, it's a matter of structure. It's only able to replicate language it's been trained on. This is why it's good at simple arithmetic, but bad at adding big numbers - it doesn't understand math, it just knows that the next character in the sequence "2+2=" is "4". It's not adding, it's completing the phrase. If you tell it "My cow has died. How long before it is alive again?" it'll spout something about the gestational period of cows, not "sorry, cow necromancy is impossible."

It's REALLY REALLY good at completing phrases - this is what lets it write pretty good limericks and passable sonnets. The more parameters you add and the more hidden layers you include in the deep network, the better it'll get at what it already does well - and to be clear, it is ASTONISHINGLY good. It's going to seriously affect a lot of industries.

I wouldn't call it smart exactly. But like I said, once they start linking together multiple systems that are good at different things, then watch out.

Mmmm you might be right. I'm a programmer and I have trouble understanding how it works. I know at its it's a neural network, and the output is, like you say, basically a giant autocomplete machine, but there's something else going on at a higher level. The way you can refer back to what it generated and it understands, often deeply, exactly what you mean, is uncanny.

I have trouble reconciling "giant autocomplete machine" with the type of cognition it seems to have.

I've asked it to summarize the moral differences between the Bible and the Quran, generate and rank story ideas according to novelty, and add subtext to a story (it wasn't good at that last one, but it knew what I meant and it tried).

That being said, if it is truly just a giant autocomplete machine, and the dazzling cognition it seems to display is emergent from it's simplistic, autocompletion brain, what would a model an order of magnitude larger be capable of?


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: Beulah
Date Posted: 16 Jan 2023 at 4:24pm
Sounds like it should be a love story, between the AI and the AI Detector.


Posted By: Suave
Date Posted: 16 Jan 2023 at 9:01pm

I think that is how they are now building their walking jumping robots,
at least that is how I would build them. One computer for the walking 
and moving and one for speaking and deciding where to go and how to miss sh*t.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: timmytimtimothy
Date Posted: 16 Jan 2023 at 10:52pm
Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 


This x 1,000,000!

AI 'art' is anti art and anti artists. If you don't want to write your own story or create your own art, that's fine. But don't ruin it for those of us who do. I wish this app and those like it a speedy death. 

-------------
M250 R2: https://rb.gy/ist1ff" rel="nofollow - Everything Be Plunderful (Comedy)

M250 R1: https://rb.gy/j7n66u" rel="nofollow - Le Grand Misunderstanding at Bobbinfeather Estate (HF)


Posted By: Suave
Date Posted: 16 Jan 2023 at 11:00pm
Originally posted by timmytimtimothy timmytimtimothy wrote:

Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 


This x 1,000,000!

AI 'art' is anti art and anti artists. If you don't want to write your own story or create your own art, that's fine. But don't ruin it for those of us who do. I wish this app and those like it a speedy death. 

Sadly, when was the last time anyone with evil intent thought about others.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: copperdogma
Date Posted: 16 Jan 2023 at 11:07pm
Originally posted by Suave Suave wrote:

Originally posted by timmytimtimothy timmytimtimothy wrote:

Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 


This x 1,000,000!

AI 'art' is anti art and anti artists. If you don't want to write your own story or create your own art, that's fine. But don't ruin it for those of us who do. I wish this app and those like it a speedy death. 

Sadly, when was the last time anyone with evil intent thought about others.

I'm interested in talking about it because it's here now, it's already having an impact, and the impact of all AI is going to have a massive impact in the near future. My favourite was of dealing with it is to discuss it and think about it to figure out what life will be like, the pros and cons, and how we can best adapt.

It's not inherently evil. Not that you were saying it was. It's more like a knife. An incredibly useful, versatile tool that can be used for harm. Sometimes unwittingly.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: timmytimtimothy
Date Posted: 16 Jan 2023 at 11:19pm
Originally posted by copperdogma copperdogma wrote:

Originally posted by Suave Suave wrote:

Originally posted by timmytimtimothy timmytimtimothy wrote:

Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 


This x 1,000,000!

AI 'art' is anti art and anti artists. If you don't want to write your own story or create your own art, that's fine. But don't ruin it for those of us who do. I wish this app and those like it a speedy death. 

Sadly, when was the last time anyone with evil intent thought about others.

I'm interested in talking about it because it's here now, it's already having an impact, and the impact of all AI is going to have a massive impact in the near future. My favourite was of dealing with it is to discuss it and think about it to figure out what life will be like, the pros and cons, and how we can best adapt.

It's not inherently evil. Not that you were saying it was. It's more like a knife. An incredibly useful, versatile tool that can be used for harm. Sometimes unwittingly.

I think plagiarising artists' work to form AI's initial learning IS evil, in the immoral, unethical and greed-driven sense. There may be some practical uses for this for people who are missing the point, such as writing boring commercial/corporate copy etc. But then what's the point of writing as an expressive form if we're bombarded with an infinite stream of unconsciously produced words?

I'm a copywriter by profession and a writer of fiction by passion, so needless to say, I absolutely loathe this. 


-------------
M250 R2: https://rb.gy/ist1ff" rel="nofollow - Everything Be Plunderful (Comedy)

M250 R1: https://rb.gy/j7n66u" rel="nofollow - Le Grand Misunderstanding at Bobbinfeather Estate (HF)


Posted By: copperdogma
Date Posted: 16 Jan 2023 at 11:43pm
Originally posted by timmytimtimothy timmytimtimothy wrote:

Originally posted by copperdogma copperdogma wrote:

Originally posted by Suave Suave wrote:

Originally posted by timmytimtimothy timmytimtimothy wrote:

Originally posted by taaaylor taaaylor wrote:

I think it's a trash app and if you care about your fellow writers, you wouldn't use it for any commercial gain or contest of SKILL, full stop. Typing sentences into an AI bot isn't a skill. 

I have no respect for writing AI. Use your brain. Read a book. Take a walk. Do your own research. It's not even good as a search index tool because it will confidently lie to you. 


This x 1,000,000!

AI 'art' is anti art and anti artists. If you don't want to write your own story or create your own art, that's fine. But don't ruin it for those of us who do. I wish this app and those like it a speedy death. 

Sadly, when was the last time anyone with evil intent thought about others.

I'm interested in talking about it because it's here now, it's already having an impact, and the impact of all AI is going to have a massive impact in the near future. My favourite was of dealing with it is to discuss it and think about it to figure out what life will be like, the pros and cons, and how we can best adapt.

It's not inherently evil. Not that you were saying it was. It's more like a knife. An incredibly useful, versatile tool that can be used for harm. Sometimes unwittingly.

I think plagiarising artists' work to form AI's initial learning IS evil, in the immoral, unethical and greed-driven sense. There may be some practical uses for this for people who are missing the point, such as writing boring commercial/corporate copy etc. But then what's the point of writing as an expressive form if we're bombarded with an infinite stream of unconsciously produced words?

I'm a copywriter by profession and a writer of fiction by passion, so needless to say, I absolutely loathe this. 

I hear ya. And I can see why this would be far more personal for you.

I think they're on a quest to build something useful for humanity. I'm sure they also hate when people use what they've created for nefarious ends. I sure would. But like my knife analogy, it's such a broad tool it's not surprising it could be used for.. less than noble pursuits.

And it has infinite benevolent purposes. Sure, you can use it to plagiarize someone's work, but that's immoral as if you had done it yourself by hand. Since it came out I've used it to learn a new programming language, help me debug code, think up tavern names for a roleplaying game, entertained by daughter, analyzed great works of fiction I haven't found the time to read yet and asked it a hundred specific and inane follow up questions to get a better understanding of it, and I've barely scratched the surface.

I agree that the morality of training it on such a wide swath of work, especially without permission, is a concern.

That's what we do as humans. We take in everything we can find. We read, we watch, we look at images, we listen, and from that we create something new, all heavily coloured by what we've experienced in our lives. I bet the people who created ChatGPT feel like that's what they're doing. Training something to learn like we do, by experiencing the world.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: timmytimtimothy
Date Posted: 17 Jan 2023 at 1:35am
Originally posted by copperdogma copperdogma wrote:

I hear ya. And I can see why this would be far more personal for you.

I think they're on a quest to build something useful for humanity. I'm sure they also hate when people use what they've created for nefarious ends. I sure would. But like my knife analogy, it's such a broad tool it's not surprising it could be used for.. less than noble pursuits.

And it has infinite benevolent purposes. Sure, you can use it to plagiarize someone's work, but that's immoral as if you had done it yourself by hand. Since it came out I've used it to learn a new programming language, help me debug code, think up tavern names for a roleplaying game, entertained by daughter, analyzed great works of fiction I haven't found the time to read yet and asked it a hundred specific and inane follow up questions to get a better understanding of it, and I've barely scratched the surface.

I agree that the morality of training it on such a wide swath of work, especially without permission, is a concern.

That's what we do as humans. We take in everything we can find. We read, we watch, we look at images, we listen, and from that we create something new, all heavily coloured by what we've experienced in our lives. I bet the people who created ChatGPT feel like that's what they're doing. Training something to learn like we do, by experiencing the world.


I love your optimism. Alas, I can't help but feel robbing humans of their creative expression is despicable. There are already too many good stories/books in the world for us to read in a lifetime. What becomes of them when there is complete and total saturation of AI-produced work? They get lost in the sea, and so do our voices. Humanity has more to lose than to gain from outsourcing all the things that make us human to AI. 

-------------
M250 R2: https://rb.gy/ist1ff" rel="nofollow - Everything Be Plunderful (Comedy)

M250 R1: https://rb.gy/j7n66u" rel="nofollow - Le Grand Misunderstanding at Bobbinfeather Estate (HF)


Posted By: taaaylor
Date Posted: 17 Jan 2023 at 2:35pm
Saw this on my reddit trawling today. I feel it's relevant: https://www.reddit.com/r/Music/comments/10e9639/this_song_sucks_nick_cave_responds_to_chatgpt/j4q5ge6/" rel="nofollow - https://www.reddit.com/r/Music/comments/10e9639/this_song_sucks_nick_cave_responds_to_chatgpt/j4q5ge6/

And, to be honest, proof of the lack of moral ethic of people using ChatGPT. It has been trained on this musician's lyrics without his consent. It's an insult to art AND it's outright theft. Even massive artists can't protect their IP. What hope do the rest of us have? Please stop using this bot. It doesn't even have the means to detect if copyrighted information is poured into it 


-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: MuffinMom
Date Posted: 17 Jan 2023 at 3:29pm
I did the same thing as you, and I have to say I was very much not impressed with the results. I don't see any of these stories written by ChatGPT winning any contest. I did several prompts. 

Way too much passive voice: "Once there was a young woman..."

Not very impressive plots - the story about a comedy with the action of shampooing hair was just a woman who shampooed her hair with dish soap. Not only was it not original, it also wasn't funny.

I haven't tried asking for a twist. Did you feel like that made it better? It wouldn't improve the passive voice, though.


-------------
https://forums.nycmidnight.com/topic58260_post628690.html#628690" rel="nofollow - 250 Word Story: Now Natasha


Posted By: copperdogma
Date Posted: 17 Jan 2023 at 3:33pm
Originally posted by MuffinMom MuffinMom wrote:

I did the same thing as you, and I have to say I was very much not impressed with the results. I don't see any of these stories written by ChatGPT winning any contest. I did several prompts. 

Way too much passive voice: "Once there was a young woman..."

Not very impressive plots - the story about a comedy with the action of shampooing hair was just a woman who shampooed her hair with dish soap. Not only was it not original, it also wasn't funny.

I haven't tried asking for a twist. Did you feel like that made it better? It wouldn't improve the passive voice, though.

Oh my no. It writes mostly terrible stories;)

If it never got any better, the creative writers of the world wouldn't have to give it a second thought.

It gets more interesting when you realize how fast it's evolved, and it's accelerating.

And it's main purpose isn't to write stories. In fact, I'm not even sure it has a main purpose.

My wife is watching British baking shows and asked why the British always seem to use self-rising flour. I had no idea so we googled it. No real answer in the top 20 links. ChatGTP, because it has a sense of understanding, knew the answer: it's mostly tradition.

I found that interesting.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: taaaylor
Date Posted: 17 Jan 2023 at 3:38pm
Originally posted by copperdogma copperdogma wrote:

Originally posted by MuffinMom MuffinMom wrote:

I did the same thing as you, and I have to say I was very much not impressed with the results. I don't see any of these stories written by ChatGPT winning any contest. I did several prompts. 

Way too much passive voice: "Once there was a young woman..."

Not very impressive plots - the story about a comedy with the action of shampooing hair was just a woman who shampooed her hair with dish soap. Not only was it not original, it also wasn't funny.

I haven't tried asking for a twist. Did you feel like that made it better? It wouldn't improve the passive voice, though.

Oh my no. It writes mostly terrible stories;)

If it never got any better, the creative writers of the world wouldn't have to give it a second thought.

It gets more interesting when you realize how fast it's evolved, and it's accelerating.

And it's main purpose isn't to write stories. In fact, I'm not even sure it has a main purpose.

My wife is watching British baking shows and asked why the British always seem to use self-rising flour. I had no idea so we googled it. No real answer in the top 20 links. ChatGTP, because it has a sense of understanding, knew the answer: it's mostly tradition.

I found that interesting.

This is a really cool* example of ChatGPT being confidently dead wrong. Try using quotes around a word to preserve the spelling so Google doesn't return "self-raising" and "self-rising" as interchangeable spelling variations as they are different types of flour.

It's self-raising flour in the UK, and it's used because it contains a very specific ratio of baking soda to flour: https://topwithcinnamon.com/how-to-make-self-raising-flour/" rel="nofollow - https://topwithcinnamon.com/how-to-make-self-raising-flour/

Quote Self-raising flour (a.k.a. self-rising flour in the US) is a common ingredient in English baking recipes, typically used for scones, pancakes and Victoria sponge cakes. It is simply a pre-mixed combination of raising agents and flour so that you donā€™t need to add baking powder or bicarbonate of soda to your recipe

*by cool, I mean not even remotely cool. Don't trust the robot.



-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: copperdogma
Date Posted: 17 Jan 2023 at 3:49pm
Originally posted by taaaylor taaaylor wrote:

This is a really cool* example of ChatGPT being confidently dead wrong. Try using quotes around a word to preserve the spelling so Google doesn't return "self-raising" and "self-rising" as interchangeable spelling variations as they are different types of flour.

It's self-raising flour in the UK, and it's used because it contains a very specific ratio of baking soda to flour: https://topwithcinnamon.com/how-to-make-self-raising-flour/" rel="nofollow - https://topwithcinnamon.com/how-to-make-self-raising-flour/

Quote Self-raising flour (a.k.a. self-rising flour in the US) is a common ingredient in English baking recipes, typically used for scones, pancakes and Victoria sponge cakes. It is simply a pre-mixed combination of raising agents and flour so that you donā€™t need to add baking powder or bicarbonate of soda to your recipe

*by cool, I mean not even remotely cool. Don't trust the robot.


Oh sorry I must not have explained it very well. Here was the actual interaction:

ME: Why is self-rising flour more popular in britain?

CHATGPT: Self-rising flour is more popular in the United Kingdom, particularly in traditional British cooking, because it was historically used in recipes for quick breads and biscuits. These types of recipes do not require the precise measurements of baking powder and salt that other recipes may need. Additionally, self-rising flour was more readily available in the UK, and was therefore used more often in cooking. The convenience of self-rising flour also made it a popular choice for home cooks.


But you're right -- I'm not sure if that answer is actually correct. That's one of the things I'm fascinated with: How trustworthy is it and in what circumstances? That will drastically inform how widely its used and for what.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: Pages
Date Posted: 17 Jan 2023 at 3:51pm
Originally posted by taaaylor taaaylor wrote:

It has been trained on this musician's lyrics without his consent. It's an insult to art AND it's outright theft. Even massive artists can't protect their IP. What hope do the rest of us have? Please stop using this bot. It doesn't even have the means to detect if copyrighted information is poured into it 

I read that Sudowrite and possibly ChatGPT have been skimming the biggest database of free stories out there: Archive of Our Own, or AO3. This is primarily a fanfiction website. People post their stories for free, so no copyright issues, but it's weird to know that a story you (fine....I) wrote is being used to train an AI on how to write a story.

Originally posted by copperdogma copperdogma wrote:

ChatGTP, because it has a sense of understanding, knew the answer: it's mostly tradition.


ChatGPT is confident - more so than Google. But confidence isn't truth. If you ask ChatGPT a question, it'll give an answer even if the answer is wrong.

Also, once you've had the AI write a story for you, re-prompt it with "write this like a human would write it". You may get better results.




-------------
http://tinyurl.com/yayxr6x6" rel="nofollow - SS R1 Premium Ingredient
https://tinyurl.com/2p8npsfa" rel="nofollow - ALL NYCM


Posted By: copperdogma
Date Posted: 17 Jan 2023 at 3:57pm
Originally posted by Pages Pages wrote:

ChatGPT is confident - more so than Google. But confidence isn't truth. If you ask ChatGPT a question, it'll give an answer even if the answer is wrong.

Also, once you've had the AI write a story for you, re-prompt it with "write this like a human would write it". You may get better results.



I think that trust will be at the crux of it's acceptance and usefulness. Maybe each result could come with a confidence score, and clicking it would explain why it's confident (or not).

Transparency and communication usually build the most trust.



-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: taaaylor
Date Posted: 17 Jan 2023 at 3:59pm
Originally posted by Pages Pages wrote:

Originally posted by taaaylor taaaylor wrote:

It has been trained on this musician's lyrics without his consent. It's an insult to art AND it's outright theft. Even massive artists can't protect their IP. What hope do the rest of us have? Please stop using this bot. It doesn't even have the means to detect if copyrighted information is poured into it 

I read that Sudowrite and possibly ChatGPT have been skimming the biggest database of free stories out there: Archive of Our Own, or AO3. This is primarily a fanfiction website. People post their stories for free, so no copyright issues, but it's weird to know that a story you (fine....I) wrote is being used to train an AI on how to write a story.


This fills me with a total and blind rage. I write online (reddit) and am at significant risk of being stolen from, too. I'm tempted to type in a keyword of my serial that only shows up in my serial, because it will be a dead giveaway if I've been plagiarized but I don't want to give those a-holes my phone number. To me it's especially soulless to steal from fanfic writers, given the potluck nature of the community and the legal gray area of asserting copyright. So f**ked. It rots the entire ecosystem of sharing stories with fellow fans. Thanks for mentioning this, as I'm out of my fanfic days and had no idea.


-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: taaaylor
Date Posted: 17 Jan 2023 at 4:04pm
Originally posted by copperdogma copperdogma wrote:

Originally posted by taaaylor taaaylor wrote:

This is a really cool* example of ChatGPT being confidently dead wrong. Try using quotes around a word to preserve the spelling so Google doesn't return "self-raising" and "self-rising" as interchangeable spelling variations as they are different types of flour.

It's self-raising flour in the UK, and it's used because it contains a very specific ratio of baking soda to flour: https://topwithcinnamon.com/how-to-make-self-raising-flour/" rel="nofollow - https://topwithcinnamon.com/how-to-make-self-raising-flour/

Quote Self-raising flour (a.k.a. self-rising flour in the US) is a common ingredient in English baking recipes, typically used for scones, pancakes and Victoria sponge cakes. It is simply a pre-mixed combination of raising agents and flour so that you donā€™t need to add baking powder or bicarbonate of soda to your recipe

*by cool, I mean not even remotely cool. Don't trust the robot.


Oh sorry I must not have explained it very well. Here was the actual interaction:

ME: Why is self-rising flour more popular in britain?

CHATGPT: Self-rising flour is more popular in the United Kingdom, particularly in traditional British cooking, because it was historically used in recipes for quick breads and biscuits. These types of recipes do not require the precise measurements of baking powder and salt that other recipes may need. Additionally, self-rising flour was more readily available in the UK, and was therefore used more often in cooking. The convenience of self-rising flour also made it a popular choice for home cooks.


But you're right -- I'm not sure if that answer is actually correct. That's one of the things I'm fascinated with: How trustworthy is it and in what circumstances? That will drastically inform how widely its used and for what.

But that's still inaccurate. It's used for more precision, not less. It's not even why it was historically invented. It took me two google searches to discern that. Seriously, you should be more critical of what you read from a bot like that. It's still in training, and if you correct it, it will apologize for its mistake and adapt its statement to the new information you gave it.

Here, now we all can learn about self-raising flour lolol: https://cotswoldflour.com/blogs/news/what-is-self-raising-flour-and-when-should-you-use-it

Here's some advice: test it on something you actually know A LOT about. It's much easier to see the holes in its communication then.


-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: Pages
Date Posted: 17 Jan 2023 at 4:20pm
Originally posted by taaaylor taaaylor wrote:

This fills me with a total and blind rage. I write online (reddit) and am at significant risk of being stolen from, too. I'm tempted to type in a keyword of my serial that only shows up in my serial, because it will be a dead giveaway if I've been plagiarized but I don't want to give those a-holes my phone number. To me it's especially soulless to steal from fanfic writers, given the potluck nature of the community and the legal gray area of asserting copyright. So f**ked. It rots the entire ecosystem of sharing stories with fellow fans. Thanks for mentioning this, as I'm out of my fanfic days and had no idea.

The bot shouldn't plagiarize your stories, but it will use the meat of them for creating a story, dialogue, and/or structure.

As an example, it knows how to add a twist. How does it know that? By understanding what a twist is and how it relates to the bulk of the text presented. It 'reads' the story and understands that clues can be dropped that might indicate a twist later. It's not great at doing this, it's still in training.

By 'meat' I mean: on AO3 there's a lot of stories about the Omegaverse. The AI has scraped all these stories and understands the basic concepts of them (Alphas are in charge, Omegas are subservient, Betas are neutral) and can plug these elements into a story.

To be abundantly clear here, I don't support AI-written stories, but I think it's too late to stop it. Now it's a matter of what we do with the information, how it's integrated into our society, and how painful that integration will be in the short-to-medium term.

I feel like the AI is taking away something intrinsic to humanity (creativity) and making a bid to eliminate humans' need to participate in this creativity. Machines were supposed to do the menial work for us so we'd have more time to create, instead AI is learning to be creative (by scrubbing art and writing sites) so humans can do the menial work (which corporations find cheaper than buying machines. To them, workers are expendable and cheap if there's no healthcare or salary expectations).

Some companies are also looking into how to watermark AI-written pieces. If this gets some footing, then it'll be easier for judges to pick out AI-written work in challenges like NYCM's.


-------------
http://tinyurl.com/yayxr6x6" rel="nofollow - SS R1 Premium Ingredient
https://tinyurl.com/2p8npsfa" rel="nofollow - ALL NYCM


Posted By: taaaylor
Date Posted: 17 Jan 2023 at 4:39pm
Originally posted by Pages Pages wrote:

Originally posted by taaaylor taaaylor wrote:

This fills me with a total and blind rage. I write online (reddit) and am at significant risk of being stolen from, too. I'm tempted to type in a keyword of my serial that only shows up in my serial, because it will be a dead giveaway if I've been plagiarized but I don't want to give those a-holes my phone number. To me it's especially soulless to steal from fanfic writers, given the potluck nature of the community and the legal gray area of asserting copyright. So f**ked. It rots the entire ecosystem of sharing stories with fellow fans. Thanks for mentioning this, as I'm out of my fanfic days and had no idea.

The bot shouldn't plagiarize your stories, but it will use the meat of them for creating a story, dialogue, and/or structure.

As an example, it knows how to add a twist. How does it know that? By understanding what a twist is and how it relates to the bulk of the text presented. It 'reads' the story and understands that clues can be dropped that might indicate a twist later. It's not great at doing this, it's still in training.

By 'meat' I mean: on AO3 there's a lot of stories about the Omegaverse. The AI has scraped all these stories and understands the basic concepts of them (Alphas are in charge, Omegas are subservient, Betas are neutral) and can plug these elements into a story.

To be abundantly clear here, I don't support AI-written stories, but I think it's too late to stop it. Now it's a matter of what we do with the information, how it's integrated into our society, and how painful that integration will be in the short-to-medium term.

I feel like the AI is taking away something intrinsic to humanity (creativity) and making a bid to eliminate humans' need to participate in this creativity. Machines were supposed to do the menial work for us so we'd have more time to create, instead AI is learning to be creative (by scrubbing art and writing sites) so humans can do the menial work (which corporations find cheaper than buying machines. To them, workers are expendable and cheap if there's no healthcare or salary expectations).

Some companies are also looking into how to watermark AI-written pieces. If this gets some footing, then it'll be easier for judges to pick out AI-written work in challenges like NYCM's.

Yes, this is absolutely true and I totally agree. It's important clarification too, so thank you. By plagiarism, I mean copying syntax, imagery, language structure, thematic concepts, etc. as a function of its machine-learning. We will definitely need a different word for this kind of IP-infringement


-------------
https://t.ly/JlEAs" rel="nofollow - SSC R1
https://t.ly/XFj6f" rel="nofollow - 250 Micro R2

SSC 2021 winner


Posted By: Pages
Date Posted: 17 Jan 2023 at 4:57pm
Originally posted by taaaylor taaaylor wrote:

Yes, this is absolutely true and I totally agree. It's important clarification too, so thank you. By plagiarism, I mean copying syntax, imagery, language structure, thematic concepts, etc. as a function of its machine-learning. We will definitely need a different word for this kind of IP-infringement

Yeah, it's likely learning all those things, but they're harder to learn than size-colour-item (big red ball, not red big ball) or whether an adverb is needed (smiled happily vs smiled sadly). The AI now has trillions of parameters, which is astonishing, so I suspect it'll learn these things uncomfortably fast.

One would hope that as it learns it develops it's own style, just like humans do when we read novels when we aren't writing. When we sit down to write it's likely that the author's style will leak into ours, but the more we write the more we develop an individualized narrative style.




-------------
http://tinyurl.com/yayxr6x6" rel="nofollow - SS R1 Premium Ingredient
https://tinyurl.com/2p8npsfa" rel="nofollow - ALL NYCM


Posted By: copperdogma
Date Posted: 17 Jan 2023 at 5:24pm
Originally posted by taaaylor taaaylor wrote:

Saw this on my reddit trawling today. I feel it's relevant: https://www.reddit.com/r/Music/comments/10e9639/this_song_sucks_nick_cave_responds_to_chatgpt/j4q5ge6/" rel="nofollow - https://www.reddit.com/r/Music/comments/10e9639/this_song_sucks_nick_cave_responds_to_chatgpt/j4q5ge6/

And, to be honest, proof of the lack of moral ethic of people using ChatGPT. It has been trained on this musician's lyrics without his consent. It's an insult to art AND it's outright theft. Even massive artists can't protect their IP. What hope do the rest of us have? Please stop using this bot. It doesn't even have the means to detect if copyrighted information is poured into it 

Wow, thanks for sharing that. What a beautiful, heartfelt response he wrote. I'm in awe.

If there's a silver lining to ChatGPT, it's caused a bunch of artists to deeply reflect on what they create, why they create it, and what it means to them and the world.

I've been thinking about this a lot as a (wannabe) writer.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: Suave
Date Posted: 17 Jan 2023 at 8:34pm

Watermarks only work if you copy and paste.
Now if there were a way of inserting certain words in a pattern throughout a story,
sort of like the US dollar has the Eurion Constellation that a photocopy machine will not copy....


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: copperdogma
Date Posted: 17 Jan 2023 at 10:51pm
Originally posted by Suave Suave wrote:


Watermarks only work if you copy and paste.
Now if there were a way of inserting certain words in a pattern throughout a story,
sort of like the US dollar has the Eurion Constellation that a photocopy machine will not copy....

We covered that earlier in the thread, but I'll paste it here:

https://theamericangenius.com/tech-news/chatgpt-watermark/#:~:text=You%20can" rel="nofollow - https://theamericangenius.com/tech-news/chatgpt-watermark/#:~:text=You%20can 't%20currently%20use,stage%20of%20ChatGPT%20feature%20rollouts.

TLDR: Cryptogenic watermarking. "The key to making AI content watermarking virtually undetectable is ensuring the distribution of words still has a unique and random appearance like regular AI-generated text."

That's interesting. If they do it consistently enough within the text, anything you don't rewrite fully by hand would still contain that distribution ("watermark"), so the more you leave unchanged the higher it would rate on something detecting generated text.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2


Posted By: Suave
Date Posted: 17 Jan 2023 at 11:07pm
Originally posted by copperdogma copperdogma wrote:

Originally posted by Suave Suave wrote:


Watermarks only work if you copy and paste.
Now if there were a way of inserting certain words in a pattern throughout a story,
sort of like the US dollar has the Eurion Constellation that a photocopy machine will not copy....

We covered that earlier in the thread, but I'll paste it here:

https://theamericangenius.com/tech-news/chatgpt-watermark/#:~:text=You%20can" rel="nofollow - https://theamericangenius.com/tech-news/chatgpt-watermark/#:~:text=You%20can 't%20currently%20use,stage%20of%20ChatGPT%20feature%20rollouts.

TLDR: Cryptogenic watermarking. "The key to making AI content watermarking virtually undetectable is ensuring the distribution of words still has a unique and random appearance like regular AI-generated text."

That's interesting. If they do it consistently enough within the text, anything you don't rewrite fully by hand would still contain that distribution ("watermark"), so the more you leave unchanged the higher it would rate on something detecting generated text.

Ah... I never followed that link before.
I think that will only catch the idiots though, pretty easy to thwart it.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Zero11
Date Posted: 18 Jan 2023 at 1:07am
Check out this article! 

https://www.npr.org/sections/money/2023/01/17/1149206188/this-22-year-old-is-trying-to-save-us-from-chatgpt-before-it-changes-writing-for" rel="nofollow - https://www.npr.org/sections/money/2023/01/17/1149206188/this-22-year-old-is-trying-to-save-us-from-chatgpt-before-it-changes-writing-for


-------------
https://forums.nycmidnight.com/topic58099_post627682.html#627682" rel="nofollow - 2023 R2 250


Posted By: Popsicle Crows
Date Posted: 18 Jan 2023 at 2:00am
Originally posted by Zero11 Zero11 wrote:

Check out this article! 

https://www.npr.org/sections/money/2023/01/17/1149206188/this-22-year-old-is-trying-to-save-us-from-chatgpt-before-it-changes-writing-for" rel="nofollow - https://www.npr.org/sections/money/2023/01/17/1149206188/this-22-year-old-is-trying-to-save-us-from-chatgpt-before-it-changes-writing-for

That was where I got going down the rabbit hole. Great article and it gives hope.
Another thing that gives me hope is that a large group of people are setting their computers to continually ask for stories, articles, papers etc on why Elon Musk's first name should be Anal instead.

They figure if he started banning everyone on Twitter who offended him within a few days of owning it, he might be so offended by a billion stories making fun of him that he'll want *his program marked so he can trace the people who use it. It would work backwards as well so that would be another way to catch the cheaters.
Waiting for EU to ban it for IP violations. It's already been caught using terms and characters who can only be found in paid (not public) works. They seem to be setting up a case to show that it's stealing work not in the public domain. None of those rights were purchased by the company.

*Like Tesla, not his idea or work. 


Posted By: Deschain
Date Posted: 23 Jan 2023 at 10:32pm
hopefully we can return to better days, when cheaters had to hire ghost-writers from developing nations to graduate college...

https://www.theregister.com/2023/01/23/turnitin_chatgpt_detector/" rel="nofollow - Universities offered software to sniff out ChatGPT-written essays: Teachers hopefully get their hands on the tool within months




-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: Suave
Date Posted: 23 Jan 2023 at 10:34pm

I see Google is going to add this ChatGPT program to their search.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Suave
Date Posted: 23 Jan 2023 at 10:39pm
Originally posted by Deschain Deschain wrote:

hopefully we can return to better days, when cheaters had to hire ghost-writers from developing nations to graduate college...

https://www.theregister.com/2023/01/23/turnitin_chatgpt_detector/" rel="nofollow - Universities offered software to sniff out ChatGPT-written essays: Teachers hopefully get their hands on the tool within months



It is a good start. But I have to think that times strapped students might use something like the AI to write their papers and then just touch them up. Course as time goes on the work might become good enough that the students would not need to touch them up. 

And having teachers use the percentages the new detector might provide without giving the student any chance to respond, well that would not be a good ending - I don't see how the teacher can use this very effectively, could not just say, "Hey, you cheated!" without some sort of solid evidence.


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Deschain
Date Posted: 23 Jan 2023 at 10:53pm
Originally posted by Suave Suave wrote:

Originally posted by Deschain Deschain wrote:

hopefully we can return to better days, when cheaters had to hire ghost-writers from developing nations to graduate college...

https://www.theregister.com/2023/01/23/turnitin_chatgpt_detector/" rel="nofollow - Universities offered software to sniff out ChatGPT-written essays: Teachers hopefully get their hands on the tool within months



It is a good start. But I have to think that times strapped students might use something like the AI to write their papers and then just touch them up. Course as time goes on the work might become good enough that the students would not need to touch them up.

I'm assuming any of the other old plagiarism tricks, like dropping your essay into Google translate and translating it to German and back would probably be enough to cover it up. I don't really know.

for us creative types, this thing seems pretty horrible, but I can tell you from experience there was a great deal of fear about algebraic solvers and things like that. people were scared that students wouldn't understand algebra I guess. I use the concepts I learned in school every single day of my job, but I probably only have to solve an algebraic equation four or five times a year. there's always going to be new tools, there's always going to be apprehension about those tools. people might be completely justified, personally, I think it's a shame if people don't learn how to write for themselves, but there's also a certain inevitability about it. I don't think anyone's wrong for hating chatGPT, but on some level it's like getting mad at the ocean for the tides


-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: Suave
Date Posted: 29 Jan 2023 at 8:29pm


Just found this.

https://www.bangkokpost.com/business/2494387/evaluating-the-impact-of-chatgpt" rel="nofollow - Evaluating the impact of ChatGPT (bangkokpost.com)



-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Popsicle Crows
Date Posted: 29 Jan 2023 at 9:46pm
Originally posted by Suave Suave wrote:



Just found this.

https://www.bangkokpost.com/business/2494387/evaluating-the-impact-of-chatgpt" rel="nofollow - Evaluating the impact of ChatGPT (bangkokpost.com)

That's one of the more concise ones I've seen. I'll be sharing that, thx!
Sound like former ceo Musk is a little peeved that he missed out on the money this will be making, lol. He didn't just ban it from twitter but from using any sources he controls. No twitter harvesting might mean it may pick up a little less neo-nazi stuff from the little scumbag's friends.

Our schoolboard is testing the GPT catcher and it's still at 100% success rate as of this morning. There's also a concerted effort by the happy hackers to mess up some common cheats. 
Sad robot, lol.


Posted By: Suave
Date Posted: 29 Jan 2023 at 10:51pm
Originally posted by Popsicle Crows Popsicle Crows wrote:

Originally posted by Suave Suave wrote:



Just found this.

https://www.bangkokpost.com/business/2494387/evaluating-the-impact-of-chatgpt" rel="nofollow - Evaluating the impact of ChatGPT (bangkokpost.com)

That's one of the more concise ones I've seen. I'll be sharing that, thx!
Sound like former ceo Musk is a little peeved that he missed out on the money this will be making, lol. He didn't just ban it from twitter but from using any sources he controls. No twitter harvesting might mean it may pick up a little less neo-nazi stuff from the little scumbag's friends.

Our schoolboard is testing the GPT catcher and it's still at 100% success rate as of this morning. There's also a concerted effort by the happy hackers to mess up some common cheats. 
Sad robot, lol.

I think the people really who are going to profit from this are the scammers from third world countries. The ones that send you a email with all the funny grammar. They will be able to have their nasty stuff formatted right and maybe get a lot more catches?
 




-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Popsicle Crows
Date Posted: 29 Jan 2023 at 11:58pm
Originally posted by Suave Suave wrote:

Originally posted by Popsicle Crows Popsicle Crows wrote:

Originally posted by Suave Suave wrote:



Just found this.

https://www.bangkokpost.com/business/2494387/evaluating-the-impact-of-chatgpt" rel="nofollow - Evaluating the impact of ChatGPT (bangkokpost.com)

That's one of the more concise ones I've seen. I'll be sharing that, thx!
Sound like former ceo Musk is a little peeved that he missed out on the money this will be making, lol. He didn't just ban it from twitter but from using any sources he controls. No twitter harvesting might mean it may pick up a little less neo-nazi stuff from the little scumbag's friends.

Our schoolboard is testing the GPT catcher and it's still at 100% success rate as of this morning. There's also a concerted effort by the happy hackers to mess up some common cheats. 
Sad robot, lol.

I think the people really who are going to profit from this are the scammers from third world countries. The ones that send you a email with all the funny grammar. They will be able to have their nasty stuff formatted right and maybe get a lot more catches?
 

Yup. Fake legalese documents are probably already rolling off the printers.


Posted By: Monsieur Renard
Date Posted: 30 Jan 2023 at 8:32am
Out of curiosity I fed ChatGPT my prompts for my most recent story after I'd submitted it (I didn't do it beforehand because I didn't want to be influenced by ChatGPT's "ideas"). While GPT is obviously an impressive technology, the story it came up with was pretty mediocre. 

I started feeding it additional ideas to spice the story up a bit, but honestly nothing it output was very emotionally stirring. I wonder what kind of literature it has access to in order to generate this stuff. To me it felt like the writing of someone uncreative who had read about story structure but not really read any stories.

That said, the tech is in its infancy. Who knows what it'll be capable of in five years.


-------------
R1 G86: https://bit.ly/3YlGZe1" rel="nofollow - The Colour of Cream

R2 G86: https://bit.ly/457i1Bi" rel="nofollow - A Normal Junket Thing


Posted By: Deschain
Date Posted: 30 Jan 2023 at 12:48pm
There's a discussion in another thread right now for people whose primary language isn't English. Lots of them are mentioning they have struggles with homophones (i.e. bear/bare, hare/hair).

I would have absolutely no qualms about someone like that pasting their story into ChatGPT and asking it to look for issues with homophones. (Not sure if ChatGPT can do that yet, but it should be close based on other things it can do.)

-------------


https://forums.nycmidnight.com/author-archive_topic43498_post551825.html#551825" rel="nofollow - All Stories


Posted By: bdemoney
Date Posted: 30 Jan 2023 at 4:01pm
Well this chat was certainly enlightening! Being new to the challenges and while I have written stories and even had a few short things published, I honestly hadn't written much more than a handful of poems in the last few years. Got extremely busy with work, husband, kids, friends, life, etc. But now I am realizing I should be practicing and honing my writing skills. 

Imagine my surprise today to stumble upon this thread and see what new technology is out there. It's kind of neat and sad all at the same time. I feel if it is used correctly maybe to generate a few ideas or to correct grammer, that sort of thing it could be helpful. But to think people would just use AI to write their stories for them is very disheartening. I love writing and think people pour their heart and souls into amazing stories. It makes me sad to think that could all end someday and AI will just write the stories for people. I truly do hope they have a way to determine if AI was used to write a story or not. If so they should at least be added as a co-author. If not they should be disqualified. 

I think it is one thing to write a story and have others (beta readers, friends, family) help you make it better. It's a totally different thing to have a computer write your story for you. Just my two sense on the matter. Hope people write their own stories. You all have them inside of you. Just look. 


Posted By: LisaBelle
Date Posted: 30 Jan 2023 at 5:15pm
This has been interesting for me too. I'm the editor of a business-focused magazine and our Chief Brand Officer asked me to add an article about this tech into an upcoming issue literally right before this topic popped up here as far as how it relates to content creation. 

Funny something I didn't know anything about has sprung up in all parts of my writing life simultaneously.


-------------
https://forums.nycmidnight.com/r2-g21-salvation-in-the-rearview-a-a_topic58234.html" rel="nofollow - 250 R2
https://forums.nycmidnight.com/r1-g92-a-spark-of-defiance-hf_topic57037.html" rel="nofollow - 250 R1


Posted By: Popsicle Crows
Date Posted: 30 Jan 2023 at 7:27pm
Originally posted by Monsieur Renard Monsieur Renard wrote:

Out of curiosity I fed ChatGPT my prompts for my most recent story after I'd submitted it (I didn't do it beforehand because I didn't want to be influenced by ChatGPT's "ideas"). While GPT is obviously an impressive technology, the story it came up with was pretty mediocre. 

I started feeding it additional ideas to spice the story up a bit, but honestly nothing it output was very emotionally stirring. I wonder what kind of literature it has access to in order to generate this stuff. To me it felt like the writing of someone uncreative who had read about story structure but not really read any stories.

That said, the tech is in its infancy. Who knows what it'll be capable of in five years.
The literature it has access to is almost everything but since they dumped musk as a co-founder/ceo, it can't use Twitter or a few of the other more private platforms.

So basically, it's using everyone's published or public work without paying for the rights.
That means it's also using all the conspiracy theories and misinformation out there.

Imagine how it could destroy classics by 'fixing' the grammar...lol.
There's so many great sources and free courses out there that so far the only use it seems to have is to help cheaters and people who want to bilk others via internet scams.


Posted By: MagsWrites
Date Posted: 31 Jan 2023 at 1:39am
Originally posted by LisaBelle LisaBelle wrote:

This has been interesting for me too. I'm the editor of a business-focused magazine and our Chief Brand Officer asked me to add an article about this tech into an upcoming issue literally right before this topic popped up here as far as how it relates to content creation. 

Funny something I didn't know anything about has sprung up in all parts of my writing life simultaneously.

Iā€™m a social media content creator for a friendā€™s website. Iā€™m also a book blogger and have my own associated socials and create my own social content for my blog as well. Iā€™ve found GPT useful for my basic ā€œHey itā€™s a Monday! Letā€™s side quest and see what happens!ā€ sort of posts. Essentially the ones that eat up hours and are mindless. I do find that I still have to tweak them to keep them in my voice though. So itā€™s about 1/2 of a time saver. 

I donā€™t write for my friendā€™s website, however the friend who owns and runs it is currently in love with the GPT to help bump up the word count for older articles that we are refreshing and moving back forward on the site. There are two writers on staff who arenā€™t thrilled with this idea, nor am I (due to the voice issue) but weā€™ve all decided ā€œnot our circus, not our monkeyā€ the site is something we all do for fun and away from our real world jobs. If it makes the site owner happy, weā€™ll do what they ask, tweak what we can do itā€™s still our product and move on. Iā€™ll run analytics on those articles and posts every quarter to see how they trend. 

As for using the GPT for true content on my blog - Iā€™m not really a fan. I have a unique voice in my writing whether it be in story craft or reviews. My blog is review based. Iā€™d spend too much time adjusting and wordsmithing the GPT for it to be worthwhile to me. 

I do see the advantage to it for bulk content creation in articles though. Particularly for your bridge paragraphs or opening paragraphs. As itā€™s AI, and it does currently write at a medium writing level, the SEO rate of articles and stories written with the ā€œhelpā€ of this tool will bump blogs, websites, and stories published on searchable platforms up the results on Google and the other engines.


-------------
https://forums.nycmidnight.com/topic46360_post487916.html#487916" rel="nofollow - SSC2022:1HM "Don't Try This at Home"


Posted By: Suave
Date Posted: 31 Jan 2023 at 1:54am


I think sites should have to declare if they use AI generated content. Course I would not go to the ste anymore....


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: Suave
Date Posted: 02 Feb 2023 at 11:27pm
Just found this!

https://www.geekwire.com/2023/microsoft-word-gets-chatgpt-integration-with-new-ghostwriter-third-party-add-in/" rel="nofollow - ChatGPT integrated into Microsoft Word with 'Ghostwriter' add-in (geekwire.com)


-------------
https://shorturl.at/oFHX0" rel="nofollow - SP Butterflies and Bumblebees
https://shorturl.at/sxQU6" rel="nofollow - SS Drama/ A Touch of Humanity


Posted By: copperdogma
Date Posted: 03 Feb 2023 at 12:43am
Originally posted by Suave Suave wrote:

Just found this!

https://www.geekwire.com/2023/microsoft-word-gets-chatgpt-integration-with-new-ghostwriter-third-party-add-in/" rel="nofollow - ChatGPT integrated into Microsoft Word with 'Ghostwriter' add-in (geekwire.com)

That's pretty cool.

A week or two ago Microsoft invested $10 billion in ChatGPT (they now own 49%) and announced they're integrating it with a ton of their products, especially Word. So soon it will have native integration as well.


-------------
https://tinyurl.com/bdhsuwxw" rel="nofollow - R2 FFC
https://tinyurl.com/ywbsdsuk" rel="nofollow - R2 SC3
https://tinyurl.com/2s479zz3" rel="nofollow - R2 SC2



Print Page | Close Window

Forum Software by Web Wiz Forums® version 12.05 - http://www.webwizforums.com
Copyright ©2001-2022 Web Wiz Ltd. - https://www.webwiz.net