AI In Poses

[Trashcan]

@KDraygo
I don’t think that the majority opinion is that using ChatGPT is a great evil; I think the majority opinion is that many of us don’t enjoy it and want people who are using it to pose (or substantive portions of what it produces to help them pose) disclose that they are. That being said, I’m going to take issue with a few of your points for the sake of being contrary myself.

my experience shows me that there are three major points I need to develop to contribute good writing to the stories being woven together with those I am RPing with.

I would argue that Mushing doesn’t call on you to contribute “good writing to the stories”. It calls on you to contribute fun to the community of players. How you are doing this is by writing in stories, but it’s also by vibes and a million other things, like showing up when you said you would.

My view on ChatGPT is that it’s a tool, one that can help a lot with part 2. [translating that creativity, imagination, thoughts, and ideas into written form that is both enjoyable and clear for everyone to read.]

I am again going to argue that “Part 2” is not about being “enjoyable and clear for everyone to read”. It is about being enjoyable to interact with. This is not the same thing. Many great writers are not enjoyable to interact with because they leave no air in the space. I think you get that, because you come on to Part 3, which is

being able to cooperate with the community in the game. Coordinating with the staff and cooperating to your fellow players to help create and develop these stories and adventures.

This is close, again, but it’s not quite right. You are not cooperating to create and develop stories. You’re collaborating to create fun.

Am I being pedantic? Maybe. but there is an important distinction here between creating “good stories” and “good writing” (which we’ll grant for the sake of argument that those are things that ChatGPT can actually do) and creating a good time, which is the actual point of playing on one of these games.

I’m not sure who will be wasting time having ChatGPT just come up with poses completely by itself and just copy pasting into the game, so to me, the concern may be a bit overblown.

I promise it’s happening. But on to my closing point:

So my “starting stats” into the hobby were probably a high part 1, low part 2, low part 3. […] The second part took years of not just reading but also RPing, going from a newbie RPer that wrote probably very cringe and rough poses to something that was more palatable to everyone.

This is part of the journey, it’s part of the fun, it’s part of the magic of these games. Outsourcing the writing means you never get better. You never look back and laugh at your own cringe moments. You never track your progress and appreciate that you haven’t just been dicking around online, you’ve actually bettered yourself in some way. You weren’t just part of a cycle of regurgitating content, you created something, however meager. Studies have already shown that using ChatGPT to help you write something makes you less creative and less engaged, which sort of bumps up against your optimism about “Part 1”.

Do you think it’s fair that people ask AI users to disclose it?

[Pavel]

Can we consolidate all our AI conversation into one big, or at least split off, AI thread? Not every discussion needs to turn into “Here’s my opinion on AI” again, and again, and again.

[Gashlycrumb]

@Trashcan’s got it. Spot on and according to Hoyle and so on.

@Trashcan said in MU Peeves Thread:

Do you think it’s fair that people ask AI users to disclose it?

It’s not unfair to ask, but you know they won’t do it.

[Juniper]

@bear_necessities said in MU Peeves Thread:

Also I ran a bunch of my poses through ChatGPT and asked if AI wrote it and it said it was debatable and so maybe I’m just secretly a robot.

Yeah, I don’t think anyone is asking the idiot machine if it wrote people’s poses. You’d get better results from a magic 8-ball.

[Pyrephox]

I don’t spend too much thought worrying about if someone used an LLM to write their poses, their desc, or anything. I just ask myself ‘was this person fun to scene with’ and ‘do I want to scene with this person again’.

I suspect, if I’ve had LLM RP, my answers to the latter two questions are ‘no’ more often than they’re yes. And I don’t need to know that it’s because it’s an LLM: people have been boring and incoherent posers for decades without help. LLMs just, honestly, make the pablum spelled correctly and with better grammar.

(And, if anyone’s worried about being seen as posing-by-LLM, just focus on being reactive and attentive to other people’s characters and really present in the vibe of the scene, and you should be fine, more often than not. No one has ever said, “OMG, I had so much fun RPing with this person…what if they were using AI??”)

[Gashlycrumb]

@Trashcan said in MU Peeves Thread:

It calls on you to contribute fun to the community of players.

https://youtu.be/0FNZfR4bGw4?si=8GD4M9l1v_JAXux3

Spoiler

Fun, fun, fun
It’s the finest thing under the sun
Now we all need it badly
So help yourselves madly
For fun, fun, fun

Fun, fun, fun
It’s a must since the world first begun
For there’s no doubt about it
You can’t live without it
It’s fun, fun, fun

The birds in the trees
All the birds and the fleas
And the spiders when their webs are spun
Even rabbits and mices they all have devices
For fun, fun, fun

Everyone, one, one
Will have soon got the blues on the run
If from Jan to December
They’ll all be remembered
As fun, fun, fun

Let’s have fun…

[dvoraen]

@Pyrephox What if we RP’d about becoming AI’s IC? A transhumanism plot, etc.

full circle gif here

[MisterBoring]

@dvoraen said in MU Peeves Thread:

What if we RP’d about becoming AI’s IC? A transhumanism plot, etc.

This gives me an idea for a Transhuman Sci-Fi MU where several NPCs are AI played by AI.

[RightMeow]

Not really what the conversation is about, but to find out if it’s AI, you post it to AI and basically say: Did you do this?

I don’t know why, but I find that both sad and completely hilarious.

[Pyrephox]

@RightMeow said in MU Peeves Thread:

Not really what the conversation is about, but to find out if it’s AI, you post it to AI and basically say: Did you do this?

I don’t know why, but I find that both sad and completely hilarious.

And then the LLM gets it wrong, because it doesn’t actually know how to analyze anything.

[Third Eye]

@RightMeow said in MU Peeves Thread:

I don’t know why, but I find that both sad and completely hilarious.

There are a couple times I’ve gotten an LLM to almost 90% replicate what someone has ‘posed’ or ‘written’ (LLM’d) on a MUSH by using a prompt that was more or less asking for the same thing. It is, indeed, both hilarious and maddening.

[Tez]

@Pavel Forked. Sorry, I wasn’t paying attention.

Dipping into my history of DMs with @Third-Eye, my favorite AI-detector, here are some things that pinged my AI-meter for @bear_necessities :

There’s something truly mesmerizing about XXX’s presence. It’s like he’s emitting an intense, fiery glow from within. His eyes, mouth, and nostrils seem to burn with the ferocity of molten lava, casting an ethereal, otherworldly light. This inner fire not only enhances his formidable presence but also symbolizes the burning passion and strength that define him. Smoke, too, spills from his nose. With this newfound volcanic essence, XXX becomes a formidable guardian, infused with the ancient power and spirit of Mount Fuji.

XXX’s unique blend of martial prowess and poetic talent makes him an intriguing character. Others might be drawn to his ability to balance the warrior’s discipline with the artist’s sensitivity, seeking him out for guidance or companionship.

XXX is deeply committed to preserving his family’s legacy and cultural heritage. Others who value history or have similar missions might find common ground with him, collaborating on projects or seeking his expertise.

The entity’s structure defies conventional understanding, exhibiting non-Euclidean geometry that suggests it exists both within and outside visible dimensions. This otherworldly appearance is alien, almost Lovecraftian in its visage, evoking a profound sense of awe and dread.

YYY has a well-defined jawline and cheekbones, giving his face a chiselled and angular appearance. He typically maintains a clean-shaven look, accentuating his facial structure. His eyebrows are neatly groomed and often expressive, adding to his ability to convey emotions effectively. He has captivating, warm brown eyes that are typically lost in thought. His eyes can convey a wide range of emotions, from intensity and determination to warmth and empathy. His hair is usually styled in a contemporary and versatile manner. He has slightly tousled hair on top, with it much shorter on the sides. His hair colour is typically a shade of brown, complementing his overall appearance. YYY maintains a fit and healthy physique. While not overly muscular, he possesses a lean and athletic build that suits his active lifestyle.

I have more examples, but it gives you, I think, an idea of the kinds of things that make me sit up and reach for my trusty fellow detectives.

[bear_necessities]

@Tez are those AI because I’m pretty sure I read all of those in 1999 lol

[Tez]

@bear_necessities I bet you did! And the AI trained on that 90s writing, too.

It reads AI to me, yes, and pinged as AI on multiple websites that detect AI. Is it? I don’t know.

[Juniper]

AI generated text typically looks at home in a game of Mad Libs which is why it loves words like “exudes”, “exhibits”, “emphasizes”, “highlights” etc. It’s an easy way for AI to get some vaguely related but illustrative nouns in there without actually constructing a pointed sentence. This is how I recognise it most if the time.

I don’t feed things into an LLM and ask it if it thinks something is AI written because that would be stupid.

[Muscle Car]

@Juniper Someone here nailed it when they called it something like “complex yet somehow boring.”

[Pavel]

@Tez said in AI In Poses:

Sorry, I wasn’t paying attention.

[Solstice]

I thought about this thread when I was in a scene where someone was clearly struggling to get their point across.

You know the type, the RPer who’s a little bit new and is flailing a bit, has some phonetic typos, and is coming off a bit flat, but they’re very earnest about it.

This thread made me appreciate that they’re clearly putting in the effort to write their own stuff rather than just slamming it into ChatGPT to try to keep up.

Struggle on, new writers. We do see you improving with every session.

[Warma Sheen]

@somasatori said in AI In Poses:

I have a question. What do you get out of MUSHing, a hobby wherein you write paragraphs at people in a turn-based format, when you’re not actually doing the writing? What’s the end goal there?

I’ve had this conversation with more than a few people and one of the most common themes that came up: an experience as close to tabletop gaming as possible, but without being crapped on because of not being able to put out AP English writing or wait 20 minutes per exchange for all the editing.

MU*ers can be catty, judgemental, petty, elitist, etc… to a higher degree than any I’ve ever known, probably to do with internet anonymity, which is one of the reasons the hobby continues to grow smaller and smaller. I’ve never known a community so highly motivated/invested in killing itself off as intensely as this one, and I’ve known some pretty shitty communities in my life (I’m old… er now.)

People show up to MU*s for very different reasons and it sometimes seems like a large number of people assume that everyone else around them is there for the exact same reason as them then get frustrated/confused when they don’t play the way they would.

I’ve had the conversation on other threads, so I won’t get into it again here if people reply back to tell me all the ways I’m wrong, but taking a tool that improves writing and using it on text based game seems like it would be a godsend to cure many of the ills that people have complained endlessly about for decades - bad writing, lack of storytellers, no interesting plots.

But to each, their own. People will feel how they feel about it and that’s okay, I guess. Its just sad for me cause I loved this hobby and I really thought this might actually put some life back into it, like CPR. So it was disappointing/sad/whatever when all the villagers pointed at the thing that could bring someone back to life thereby improving the outlook/prosperity of the entire community, called it a witch, and want it burned at the stake.

ETA: Not for nothing, but to go along with the analogy, some non-small portion of the thread is basically a witch hunt - with dubious degrees of accuracy.

@Pavel said in AI In Poses:

I can understand people using it as a translator.

I know you mean that it takes Spanish and makes it English. But it seems like that also could apply to taking Shitty English and translating that to English.

[KarmaBum]

There are people that I played with for years who are so boring, so uncreative, so repetitive and lacking spark that - by the metrics above - they are definitely being controlled by an LLM. There are times that my own writing is so flat that I would be less embarrassed if ChatGPT had written it.

While I don’t particularly want to RP with a robot, I don’t have space in my brain to start worrying about whether or not someone I’m writing with is now or has ever been an AI.

^ part of that reply was written by ChatGPT

…or was it?!