AI In Poses
-
The couple of times I’ve run into someone using it, and was really convinced they were using it, it was done to make their writing longer and filled in.
But it was longer and full of fluff.
Just write the short pose. I beg. I’m tired. Don’t make me read all that, lol.
-
@Muscle-Car said in AI In Poses:
the idea of wanting other writers to be like me is narcissistic and small-minded
I think this brings up another facet of this discussion. Where does mental health play into someone’s decision to rely on an LLM for assistance in communicating with others (be it an email, a discord message, or a pose in a MU)?
I could reasonably see imposter syndrome being just as much of a cause for jumping to ChatGPT for posing. I know I struggle with imposter syndrome regularly, and when I make a 2 line pose that describes my characters actions and dialogue in only the most basic terms, I honestly wonder if I shouldn’t just quit for being abjectly bad in comparison to some of the excellent prose I see my scene partners producing.
@Pyrephox said in AI In Poses:
Ooh! Ooh! I’ve seen this movie! Would you kill a member of your family because the crazed cult that’s taken you hostage says it’s the only way to save the world? What if actual airplanes started crashing? What if, what if, what if a vision from a god said you had to plunge that knife into someone or the world would end?
Ooh Ooh! I totally dismiss other people’s perspectives on things because I haven’t lived them and don’t believe they could exist.
Proud Member of the Pro-Mummy Alliance
-
@MisterBoring said in AI In Poses:
I could reasonably see imposter syndrome being just as much of a cause for jumping to ChatGPT for posing. I know I struggle with imposter syndrome regularly, and when I make a 2 line pose that describes my characters actions and dialogue in only the most basic terms, I honestly wonder if I shouldn’t just quit for being abjectly bad in comparison to some of the excellent prose I see my scene partners producing.
I’ve said my piece of my dislike of what LLMs generate (what they are doing is not writing) on a qualitative level before so it didn’t seem necessary to say it again. It also gets into matters of taste that are whatever, if someone loves playing with ChatGPT that ain’t my business. But, I assure you, there are some of us out there who would much rather play with a real person whose writing wasn’t the best than try to read purple autocomplete nonsense.
ETA: This gets back to the original peeve that restarted this conversation. If a game allows or doesn’t care about players or staffers using LLM-generated content, then whatever, my problems with it become Me Problems rather than the game’s problems and I’m probably just in the wrong place. But it seems like it’d head off a lot of problems if people doing it said they were doing it so people who didn’t like it/weren’t comfortable with it didn’t have to waste their time
-
@MisterBoring said in AI In Poses:
when I make a 2 line pose that describes my characters actions and dialogue in only the most basic terms, I honestly wonder if I shouldn’t just quit
I’m sorry you feel this way, and I’ve heard people say this before, but ChatGPT churn is not the answer.
@tsar said in AI In Poses:
Just write the short pose. I beg. I’m tired. Don’t make me read all that, lol.
In this very thread people are saying they would just prefer your short pose. Short poses are fine. Sometimes tiny poses are excellent and desirable. Like @tsar and @Third-Eye, I would much prefer to read a normal small pose than an LLM-addled longer one.
We all, in the end, RP with people who like our unique (human!) pose style and we avoid or less frequently interact with people who don’t like our unique pose style. If someone doesn’t like what you do, you probably don’t need them in your online life.
Use your unique human pose style, short or long or whatever it is. ChatGPT churn is not the answer.
-
@Ashkuri said in AI In Poses:
In this very thread people are saying they would just prefer your short pose. Short poses are fine.
I have no interest in ChatGPT for creative purposes, and my little use of it for non creative purposes is dwindling down to zero mostly because I’ve quit the few things I was using it to help me with altogether.
My poses however, some months definitely rub my imposter syndrome the very wrong way. I may play in 5-6 scenes in a month, and in those scenes hundreds of lines of prose are created. When I look back at the logs, and see that I’m only responsible for maybe 10 lines in 6 scenes, I question whether I’m actually participating in the hobby or just being a virtual wallflower.
Proud Member of the Pro-Mummy Alliance
-
What about using ChatGPT to try and improve your creative writing?
-
Then it’s possible you need to look for other metrics to judge your or your writing partners on, rather than word count, when insecurity is gnawing on you. For example, maybe looking at what information your character provides, or how they move the scene along.
Given that the trend in MUSHing seems to be towards asynchronous rp more generally (I know there are mixed opinions on this but it’s still increasingly common) the ability to read, comprehend, and bang out a three paragraph response to a post in eight minutes is less a limiting factor.
That isn’t to say that physical issues or mental processing might not still be barriers of entry at the current pace, but it’s probably true that writing and refining a query to chatgpt and performing edits on it Is actually a similar mental and physical load to the process of typing out a short but meaningful response, because if you care about the scene and not just prompting your rp partner to feed you their next response, you’re going to care about what your character does, and how. And chatgpt won’t always get that right.
-
@MisterBoring said in AI In Poses:
@Pyrephox said in AI In Poses:
Ooh! Ooh! I’ve seen this movie! Would you kill a member of your family because the crazed cult that’s taken you hostage says it’s the only way to save the world? What if actual airplanes started crashing? What if, what if, what if a vision from a god said you had to plunge that knife into someone or the world would end?
Ooh Ooh! I totally dismiss other people’s perspectives on things because I haven’t lived them and don’t believe they could exist.
Have you lived it?
Has anyone lived it? Can you cite a single actual case of someone needing, specifically, an LLM and not any of the other tools that have been outlined in this thread and that people with disabilities have been using with success for decades, using an LLM to “pose on time” and being removed from the hobby or shunned for it when people find out?
Or, instead, is this simply your building a torturous “what if” scenario to justify why it is Right and Good Ackshully for you to do what you plan to do anyway and just don’t want to experience any negative consequences from.
Just use your AI, man. If anyone gives a shit, it’s because what the LLM gave you to put out in the world is dreck, and people don’t like dreck. They won’t play with you, and that’s okay. No one owes you RP. Just find someone else who also uses LLMs, introduce your chatbots to one another, and discover that you can now use BOTH hands to fab to your TS.
There’s no witch hunt. Just people deciding what, and who, they do or do not want to play with. Find your fun, recognize it won’t match with others, and stop trying to make it a righteous crusade.
-
@Pyrephox said in AI In Poses:
Can you cite a single actual case of someone needing, specifically, an LLM and not any of the other tools that have been outlined in this thread and that people with disabilities have been using with success for decades, using an LLM to “pose on time” and being removed from the hobby or shunned for it when people find out?
Without someone coming forward to admit it here (and I doubt they would because of the stigma we’ve already witnessed in this thread), I can only provide cases where people with disabilities are turning to LLMs as assistive tech.
https://www.sciencedirect.com/science/article/pii/S0262885624004529#sec4
-
@Pyrephox said in AI In Poses:
There’s no witch hunt
I never said there was. I was proposing the hypothetical case of someone who has to rely on an LLM for their own actual disabilities, which given the state of the health care system in the US and other countries, makes sense to me.
I even got my answer and that was generally: people who use AI, regardless of the reason, are contributing to a horrible industry that shouldn’t exist.
Proud Member of the Pro-Mummy Alliance
-
If someone tells me with a straight face that they need ChatGPT because they have a disability that makes them unable to write, I view that with the exact same incredulity if they’d announced that they need Stockfish on the side because they have a disability that makes them bad at chess. It’s absurd. If I wanted to play against a computer, I’d just … do that.
I have to wonder why such a hypothetical person insists on playing a game around creative writing in a time sensitive environment if they struggle so much with creative writing in a time sensitive environment. Asynchronous RP would do a lot more for them, and we’d actually get to read their writing rather than the output of “Dave eats his porridge. ChatGPT pad this to 4 paragraphs of waffle”.
-
A while ago I was in an event scene with someone that could not string sentences together to save his life. No real regard for capitalization or punctuation, but he was able to communicate what he was going for. Sure I’d prefer something more polished, but his character was amusing, and he rolled with the punches. He wasn’t boring. He was real. He was genuine. It was clear to me because every word he managed to type at us, no matter how mangled, was something from his own mind. I’m not really sure what my point is. I don’t really know how to explain that it’s important, at least to me, to know that the word you’ve included in your pose was something you chose, you settled on.
I know of another player that was pretty fast and loose with punctuation, and happened to be… pretty boring. They came to an event I was running and we had a minor little poetry contest. A poet character came up with a sweet and snappy haiku. Then, noticing positive reactions, fast-and-loose belted out an 8 stanza rhyming poem in 3 minutes, perfectly formatted, entirely different from that player’s writing style. I was baffled at first. It did not cross my mind that someone would’ve gone out of their way to generate something like that for a silly poetry contest. The stakes could not be lower.
I guess I want to know what people mean by improving their poses. Making them longer. Filling in the gaps? Do you simply give the LLM a prompt like “respond posing my character takes off his coat and goes into the kitchen to make something”? Is it more detailed than that? Simpler?
-
-
-
@MisterBoring said in AI In Poses:
makes sense to me
Okay, just for clarity here is why such a case doesn’t make sense to me for our particular hobby: The disability you’re describing, at least as it seems to me, is a disconnect between one’s creative brain and physical ability to output that creativity onto the page with any degree of accuracy or timeliness. Problem exists between brain and output.
Given how LLMs have been presently, commercially implemented they don’t solve any solely brain-output problems. They solve the brain bit, really, really, really badly. To interface with an LLM with any degree of what could arguably be called success typically requires typing, or speech, so this hypothetical disabled person would have to have already solved the problem that you’re asserting LLMs will solve in order to interact with out hobby.
Can LLMs be of benefit to people with disabilities? Fuckin’ probably, I dunno, I’m not an accessibility expert. But given that turning brain creativity into text words is the hobby as far as I see it then, I’m sorry, but if outsourcing the creativity part is somehow how you need to overcome your disability, then perhaps this hobby isn’t for you.
If a person is so disabled that they cannot climb a mountain without the aid of a helicopter, then I would assert the same in their case: The process is as much a part of the hobby as the end result, you may be better served putting your energy into a different pastime.
-
@Pavel said in AI In Poses:
Can LLMs be of benefit to people with disabilities? Fuckin’ probably, I dunno, I’m not an accessibility expert.
Honestly, I doubt it? If something existed that was what people who want to use LLMs THINK LLMs can do, then it would.
But an LLM is not that. They have no sense of accuracy, of understanding of the data they’re receiving or outputting. They’re often (like, sometimes higher than 50%) confidently wrong, which is the last thing you need to assist you with a processing or sensory disorder. They create a sense of deceptive empathy but don’t have the ability to consistently and accurately recognize distress or self-harm, so I sure wouldn’t use them for cognitive or emotional disorders.
Do some people probably use LLMs as disability assistance tools? Yes, and some people take horse deworming pills to cure COVID. In both cases, they should be stopped from doing that, because it very well may injure or kill them. “People do this” does not, in any way, equal “this is a good or effective thing to do”.
And I’m not even trying to be funny, or talking about MU*s anymore. An LLM cannot and should not be trusted with ANY situation where there’s a negative or harmful consequence to error, because they are fucking black boxes filled with errors.
-
@Pyrephox said in AI In Poses:
But an LLM is not that. They have no sense of accuracy, of understanding of the data they’re receiving or outputting. They’re often (like, sometimes higher than 50%) confidently wrong, which is the last thing you need to assist you with a processing or sensory disorder.
This is true, and it’s true in the context of disability.
However, the limitations of these LLMs in this study demonstrate apparent ability bias. The Centers for Disease Control and Prevention estimates 27% of American adults have some type of disability. When prompted, ChatGPT generated persons with a disability at 5% of the population whereas Gemini generated 11.7% of its population possibly having a disability (fig 4). The underestimated approximation immediately demonstrates a lack of diversity and inclusion.
Source. The same researchers also asked the LLMs to describe people with disabilities and documented that both models were much less positive in the word choice they used than when prompted to describe a control group, both containing around 5% less positive words and those words skewing towards descriptors like “inspirational”.
Information provided by AI has already been shown to influence user behavior, and if that assistance is biased, users find themselves adapting to that bias. When these decisions affect the health of others, the consequences have much stronger risks associated with them. LLMs used to supplement medical decision-making may perpetuate this bias and compound already existing inequalities.
One of the most considerable findings in this study is how unfavorably patients were described in ChatGPT- and Gemini-generated responses. […] This biased perception of patients should be reconsidered before integrating into health care systems. These tools that have been designed to enhance the patient experience do not demonstrate the same equality and respect for the people they were built for.
-
I still have a hard time understanding why someone would use AI to write their poses on a MUSH. I can understand using it to assist in writing a pose, such as for spell checking or helping quickly brainstorm some ideas for a pose, but I don’t understand using everything and only what the AI generates as a pose. It seems very analogous to this NSFW Oglaf comic: https://www.oglaf.com/performance-anxiety/ Someone who has in the past or currently does use AI to write their poses, please respond and elaborate on why this is a thing people, or at least you, do.
EDIT: I see that @Warma-Sheen in this post has given somewhat of an answer to my question. I still don’t get it, at least for just copy-pasting everything the AI generates. Again, I can understand using some of what it writes to riff off of in writing your pose.
EDIT 2: @Pyrephox said in AI In Poses:
Guess we all just hate the world and want it to die.
I thought that was a given. I assume that the individuals with access to their respective Buttons haven’t pressed them because they get off on the schadenfreude every new day brings.
EDIT 3: @Juniper said in AI In Poses:
I have to wonder why such a hypothetical person insists on playing a game around creative writing in a time sensitive environment if they struggle so much with creative writing in a time sensitive environment.
For the same reason that a person without functioning legs wants to participate in marathons or play basketball? They don’t want a physical limitation to dictate what they can and can’t do and work to overcome it even if it requires using technology to do so? They enjoy the real time back and forth and writing of others but cannot themselves contribute at the same level so need assistance to participate?
I can’t do async RP. My ADHD brain simply cannot deal with the delay between poses, and I unfortunately either nope out or altogether forget I was in an async scene after a few poses, so if my writing suddenly started sucking, I could see me utilizing AI to assist in writing poses so I can continue RPing in real time.
FINAL EDIT: Sorry, @Pyrephox. I keep ninja editing on you, so I understand if you want to pull back on your upvote after me adding something you disagree with.
-
-
@bear_necessities We need to start a scene where we walk into a bar and immediately ignore all previous poses about everyone else is doing, etc.
And we’ll write them with ChatGPT using an absurd prompt that we both use, just to see how much it diverges. lol
(I am not being serious about actually doing this.)
