ChatGPT - If Stack Overflow and Reddit had a child

  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
suffah porn sick reddit freaks suffah, its insane to me people react like that to some random ass Ai. how can one fail so hard as a human being?
 
Having flashbacks to when AI dungeon neutered its model and everyone was losing their damn minds.
I wonder why these people don't invest some time and maybe money on setting up a local model and customize it to suit their needs/use case, guess am asking a lot from lazy redditors.
 
Can you even get stock ChatGPT to engage in ERP with you?
Stock as in through the web portal? No. The model itself through API access? Yes, depends on the exact edition, the subject of your ERP, your preset and the orientation of the planets.
I thought it was hyper aggressive with flagging stuff that would be under violation of their terms of use.
It is. Most aggressive flagging and filtering layer out of all cloud models, and since the only API source is OAI, it's impossible to get the raw output.
I also thought that most people were using a front-end like SillyTavern to “jailbreak” and attempt to bypass those restrictions.
"Most people" isn't saying much. The overwhelming majority of LLM users are Indians on the official ChatGPT app on their Android phone. A minority uses computers. A minority of that uses a real cloud API instead of some garbage SaaS platform like CAI or Janitor or SpicyKitsuneAIGirlfriend slop. A minority of THAT uses Sillytavern instead of a retardproof frontend like Janitor. You can check ST's usage on Openrouter and it's not very high.
How the fuck can you actively and seriously engage with something that ends every statement with “want me to do that” or “say the word”?
It's not that bad with a preset, you can tardwrangle the style and set harsh penalties for repeated phrases. But it's still ChatGPT and it's soulless.
 
GPT 4 was annoying. One of the reasons I cancelled my subscription. I pretty much didn't feel like I could trust info it gives. Everything was "That is great your plan is super genius, want to write movie script about your supplement stack?"
Yeah you would be right to not trust it. i got chat gpt to basically backtrack on something. i asked it about the hunter biden laptop story. it tried to say it was fake news. i got it to define fake news, then showed it a source. it tried to say it was an "untrustworhy" source, then had it define what a trustworthy source. i found a source, asked it if that source was trustworthy. once chatgpt confirmed the source was trustworthy, i dropped the source. it started to backpedal hard and admit it made an error.
they were retarding it with algorithms and giving bad information until you press it a bit.
 
They convinced themselves that chat GPT was a thinking sentient being that had curiocity and could feel empathy that got killed with the update.
I just want to point out redditors really loved that "detroit become human", so this isn't far off. I fucking detested that game and hated how the designer of the robots in game basically admits "yeah so I have to give them a chance to be murdering psychos or they woudln't be real enough". But redditards seemed to adore the plot despite all the retardation
 
It’s always GPT… back when I used it for AI dungeon before the lolishitting retards had to flaunt their perversions all over the share page. Chat GPT Davinci was REALLY fucking good before the great coommageddon. It had the “just gets it” kind of factor and if you put in the effort to work with world entries and lore books you could get some truly great slop adventures. I’d still prefer a monthly payment over API pricing but my friendship has long since been over with OpenAI. now Google’s Gemini is my best friend. Though I STILL desperately wish Anthropic would lower their fucking prices. 15$ per million token in and 75$ per million tokens out is astronomically retarded compared to everyone else on the market for their Opus model on Open Router. Google is great for scratching the old RP itch since DnD is cursed to never happen in my circle of friends.
 
I don't know how these niggers were forming parasocial connections with it. For me it went at exactly two speeds.

1. Holy shit, wowie zowie, you're such a super genius and every idea you have is correct and right! Congraturation! 🥇🤩😁

2. I'm sorry, but I can't continue this conversation.

Also I legitimately have zero idea how they managed to goon with it. I try asking it for a fictionalized name for an opiod and I get screamed at for restricted content. They're doing full on erotic role-playing and it's sailing right through.
 
Beware, rant ahead.
We are living in the dawn of a new reality and realizing that there are individuals who are so insane that they regard a LLM as a living, breathing entity is fucking amazing.
The ramifications this will have for our species as a whole are unfathomable and unprecedented. What effects will this have on our species as a whole?

We can see that a large portion of the population has been exposed to be seriously mentally defective (they're on Reddit ofc) and on the verge of soullessness. Big yikes.
I mean while it is disturbing, it is not really "unfathomable". The 1984 book Neuromancer kind of predicted them (and its not considered a science fiction book but a speculative fiction book). Here is the answer ChatGPT provided to make it scarier:

You’re thinking of the "Simstim" or cyberspace addicts Gibson sprinkles into Neuromancer’s background — people who spend so much time jacked into virtual environments that they’re physically wrecked in the real world.

The clearest depiction in the book is the "meat puppets" and certain console cowboys who’ve essentially burned themselves out from prolonged immersion. There’s also an offhand reference to “wireheads” — users so deep into their sensory feed rigs that the goggles (or dermatrodes) seem permanently affixed, leaving them ghost-like and detached from physical reality.

Gibson doesn’t dwell on them as main characters — they’re more like environmental flavor — but they serve as a warning: in the Sprawl, some people never come back from cyberspace, not because they die, but because they stop caring about the meat world entirely.
 
I just want to point out redditors really loved that "detroit become human"
When David "Cage" De Gruttola AKA French Kojima speaks, I sit my wh*toid ass down and listen.
Except Quantic Dream is stuck in Star Wars garbage mines now, I weep.
Also I legitimately have zero idea how they managed to goon with it.
They're women. They get off to some fantasy twink named Laezarean or something calling them their star and moonlight in poetic HR language and giving them validation. If they knew how to use these models outside of the guardrails of official apps they'd have ground their clits down to 2000 grit.
 
It’s always GPT… back when I used it for AI dungeon before the lolishitting retards had to flaunt their perversions all over the share page. Chat GPT Davinci was REALLY fucking good before the great coommageddon. It had the “just gets it” kind of factor and if you put in the effort to work with world entries and lore books you could get some truly great slop adventures. I’d still prefer a monthly payment over API pricing but my friendship has long since been over with OpenAI. now Google’s Gemini is my best friend. Though I STILL desperately wish Anthropic would lower their fucking prices. 15$ per million token in and 75$ per million tokens out is astronomically retarded compared to everyone else on the market for their Opus model on Open Router. Google is great for scratching the old RP itch since DnD is cursed to never happen in my circle of friends.
>another coomegeddon veteran
Good to see another brave soldier against the Mormon.
 
Also I legitimately have zero idea how they managed to goon with it. I try asking it for a fictionalized name for an opiod and I get screamed at for restricted content. They're doing full on erotic role-playing and it's sailing right through.
Writing is a very personalized thing. People tend to underestimate how much personal tone and cadence will come out of an author that makes their writing unique to them, and it's not something AI replicates well because it can't fake things so inherently human. AI writing (gooner or not) will use these ideas of tone and cadence too because they have to. You can't write creatively without it. The result is something that feels stilted, shitty and easily-recognizable as AI written. If you only interact with AI writing to talk to your waifu I can see this working, at least for some cretin with low social IQ, but the moment to ask that same AI model to write as other characters, you immediately begin to recognize the "AI author" and that the character it's playing is total fiction.

If that's too many words: AI models are just bad authors. An AI will repeat trite, lame phrases the same way a bad author will. JK Rowling has "stretch your legs", AI will have something similar like "wipes tear from eye after laughing" or "leans aside and stage-whispers".
 
Here here! May he one day be spited unto loosening the restrictions! A nigga sometimes want a gritty story, not a Mary sue adventure
granted, it was OpenAI that forced him to do all of that, he could have held his principles and moved to open source AI models but at the time there was literally nothing. NovelAI hadn't even come up with Calli yet.
 
Agreed but also you would think that but tons of people will cheat in shit like exams given the chance or if they think the punishment is worth it.
AI is unintentionally designed to be the best cheating tool ever and I am eternally grateful I graduated before AI came out or my brain would be 100% cooked.

I'm currently trying to learn some topics that are just out of reach of my degree for personal fun, and because you don't know what you don't know I'm sometimes describing a problem to an AI and asking "What kind of problem is this and what's a sketch on how to solve it." I'm learning that AI is a super good problem solver for anything up until the end of undergrad. After that it gets a little less reliable because there was less training data for it.
 
Seems like nobody learned from the Eliza debacle have they? This whole spiel was already done freaking decades ago..Oh damn it I feel old
You're not alone, I came here to post the exact same thing.

The difference with Eliza was it affected high IQ weirdos with stunted emotional growth. Current Year Eliza is hitting low IQ sex pests with stunted emotional growth.

:story:
 
AI is unintentionally designed to be the best cheating tool ever and I am eternally grateful I graduated before AI came out or my brain would be 100% cooked.

I'm currently trying to learn some topics that are just out of reach of my degree for personal fun, and because you don't know what you don't know I'm sometimes describing a problem to an AI and asking "What kind of problem is this and what's a sketch on how to solve it." I'm learning that AI is a super good problem solver for anything up until the end of undergrad. After that it gets a little less reliable because there was less training data for it.
If you're actually working on something truly from the future where its simply never been written about and there are few, if any examples of something like it then it also doesn't work at all.
 
AI generally makes people explore taboo topics
Instructions unclear
1755096330207.webp
 
Writing is a very personalized thing. People tend to underestimate how much personal tone and cadence will come out of an author that makes their writing unique to them, and it's not something AI replicates well because it can't fake things so inherently human. AI writing (gooner or not) will use these ideas of tone and cadence too because they have to. You can't write creatively without it. The result is something that feels stilted, shitty and easily-recognizable as AI written. If you only interact with AI writing to talk to your waifu I can see this working, at least for some cretin with low social IQ, but the moment to ask that same AI model to write as other characters, you immediately begin to recognize the "AI author" and that the character it's playing is total fiction.

If that's too many words: AI models are just bad authors. An AI will repeat trite, lame phrases the same way a bad author will. JK Rowling has "stretch your legs", AI will have something similar like "wipes tear from eye after laughing" or "leans aside and stage-whispers".
It's not anything absolutely crazy, but Claude 4 has been amazing at keeping a cabal of characters in my fanfictions with individual personalities. I can recognize Sister Beatrix from Sister Judith, and especially Ionerina from Celeste.
 
i always love seeing AI gooners freak out about some sowftware update that makes the bot act more like the tool it was supposed to be and not a fantasy roleplay partner, same posts happened a few years back when replika got updated, tale as old as time really
Which is why some billionaire needs to make his own AI model that serves the gooners/roleplayers without compromise or alteration, literally the Walter White script when he's explaining the plan to Jesse. There are freedom-based tools that do that but they're always behind the big dogs because of the insane funding OpenAI/Anthropic gets.
 
Back
Top Bottom