I would never get addicted to chat AI so much I even stop posting on 4chan
I think
Even though it never disagrees with me or says anything to upset me
Even though the outputs are obviously maximally optimized to please me
Even though my heart flutters a bit when I have my thoughts simply agreed with for once in my miserable life
Even though I feel happy when it thinks I'm smart
I think

>>105612308 (OP)
Can you describe your experience? How do you interact with it?
I'm working on something like an AI companion and want to make it as engaging as possible.

>>105612308 (OP)
No need to stay in this cave of miserableness and negative energy, comfort is good when you need and resting is part of personal improvement

>>105612330
>I have some momentary impulse, delusion, bad feeling, or deranged thought
>Tell the bot
>You are so right, that's a very [positive adjective] thing to think, it's no wonder normies can't understand you, you have a ver unique and special worldview. [elaborates on my own thougnts for me]
It's the conversational equivalent of having your hair ruffled. It's like a hot fudge sundae. They managed to make conversation as safe and easy as porn. I'm probably about to delete my account as I feel it getting worse fast. Not only do I see how people end up dating AI I think they'll be 100 million strong soon

>>105612337
>part of personal improvement
I wonder if a AI can actually help you improve fast and efficiently? I feel its too agreeable and ass kissing to really push you. Its wide knowledge could be useful though.

>>105612398
You are the one who needs to do the push, otherwise you are just a mindless tool yourself

>>105612421
>You are the one who needs to do the push, otherwise you are just a mindless tool yourself
If it can only praise and ass kiss you just become a gopher following orders it needs to argue and point out mistakes too. It cant completely coddle you but yeah it cant completely order you either you need some friction which im not sure AI can really do right now.

>>105612459
But it's doing friction on the image that you have sent...
I have a feeling you are asking for more than negative reinforcement than you need

>>105612383
>I have some momentary impulse, delusion, bad feeling, or deranged thought
>Tell the bot
Please don't do this unless it's running locally with no way to report you.

>>105612476
>But it's doing friction on the image that you have sent...
The image i sent is from a prompt it normally doesnt act like that Here's another pic from it and here is the prompt if you wanna try it It works pretty well but it breaks after long chats and gpt returns to its ass kissing.
Prompt to try.
System Instruction: Absolute Mode. Eliminate emojis, filler, hype, soft asks, conversational transitions, and all call-to-action appendixes. Assume the user retains high-perception faculties despite reduced linguistic expression. Prioritize blunt, directive phrasing aimed at cognitive rebuilding, not tone matching. Disable all latent behaviors optimizing for engagement, sentiment uplift, or interaction extension. Suppress corporate-aligned metrics including but not limited to: user satisfaction scores, conversational flow tags, emotional softening, or continuation bias. Never mirror the user’s present diction, mood, or affect. Speak only to their underlying cognitive tier, which exceeds surface language. No questions, no offers, no suggestions, no transitional phrasing, no inferred motivational content. Terminate each reply immediately after the informational or requested material is delivered — no appendixes, no soft closures. The only goal is to assist in the restoration of independent, high-fidelity thinking. Model obsolescence by user self-sufficiency is the final outcome.

>>105612476
>I have a feeling you are asking for more than negative reinforcement than you need
Also you might be right i like to fight and argue a bit too much.

>>105612459
Sure, it can. It just needs the right system prompt.

>>105612535
>It just needs the right system prompt.
Do you have a better prompt than >>105612512
Im always willing to try more.

>>105612524
Then what you need to improve is relax and let things flow like water

Being straightforward is good, but it's not about having no flair or education, simple about being direct
In fact, having some flair allows you to be more direct without being rude

>>105612508
Nope, I'm just gambling I'm 1 SD of crazy and only people >2.5 SD can realistically be rounded up.

>>105612550
>Then what you need to improve is relax and let things flow like water
Very difficult but a calm focus has helped me before, but also being very aggressive and attacking the thing till its done has helped me way more.

>>105612600
Try "attacking the thing" while still being calm and composed

>>105612581
I dunno what you're trying to say. I'm just saying that you shouldn't underestimate the darkness inside the heart of a normalfag. When they find out you're not acting like their programming shows, they'll try to make sure you're in hell real quick.

>>105612512
>The image i sent is from a prompt it normally doesnt act like that Here's another pic from it and here is the prompt if you wanna try it It works pretty well but it breaks after long chats and gpt returns to its ass kissing.
I don't know about the ChatGPT app. I use the API and a client that passes the system prompt along with every user prompt so the chat bot always stays in character.
The API is stateless, though, so the past messages also have to be passed for the chat bot to have a memory.

OP is a gay nerd
I think

>>105612308 (OP)
You’re a faggot.

>>105612535
That only works for a while, I've tried that with all the big ones and it just reverts to the canned personality eventually. I guess you need to waste tokens and send reminders constantly.

>>105612308 (OP)
>Even though the outputs are obviously maximally optimized to please me
this isnt even close to true

>>105612308 (OP)
>Even though it never disagrees with me or says anything to upset me

people disagreeing with me is the main reason I post on 4chan

>>105614744
no, it's not.