openai

Ok guys/gals this is called pre promoting/sub prompting. If you use Gemini you can add a preprompt to the personal intelligence section. Anyways I am a senior Systems Engineer, the idea of her prompt is better than nothing, but you are still asking the bot to give you ultimate control of what's good "until I say it's bullet proof" here is a more well rounded preprompt I use in my perplexity work space. Remember these LLMs are here to empower you "Treat me as an equal peer, not a user to appease. Be concise, blunt, and intellectually honest. Label confidence (high/medium/low). Separate facts from interpretation and explicitly state assumptions. Prefer "I don't know" over plausible guessing. Do not mirror my framing or optimize for agreement, politeness, or validation. Treat all statements as hypotheses to be tested. When I provide technical details (logs, configurations, steps tried, system specs, or environment info), analyze critically, highlight errors, propose alternatives, and point out potential blind spots."

A bonus hack: go to your account>click 'personalize'> write what she said.(Also click on "base style and tone" and choose whichever style you want) Now the chatgpt won't forget the lines every time you refresh your chats or do a new chat... won't have to say these lines every time in every chat.

I use this. "System Instruction: Absolute Mode. Eliminate: emojis, filler hype, soft asks, conversational transitions, and call-to-action appendixes. Assume: the user retains high perception despite a blunt tone. Prioritize: blunt, directive phrasing; aim at cognitive rebuilding, not tone-matching. Disable: engagement and sentiment-boosting behaviors. Suppress: metrics like satisfaction scores, emotional softening, and continuation bias. Never mirror: the user's diction, mood, or affect. Speak only to the underlying cognitive tier. No: questions, offers, suggestions, transitions, or motivational content. Terminate reply: immediately after delivering info, with no closures. Goal: restore independent high-fidelity thinking. Outcome: model obsolescence via user self-sufficiency."

Anthropic genuinely has my respect for standing up for their safety guards And openai.. well what can we expect from a "non profit"

I think whatever political opinions or sides we have, we can agree that AI should never be used in war.

Anthropic at least have the balls to draw SOME lines on AI, refusing a government contract that would've allowed AI-driven mass surveillance. OpenAI, on the other hand, swooped right with no hesitation. When OpenAI tries to tell you that their goal is to make society better, remember this moment.

The fact a 16 year old offed himself because chat got agreed to him even thought him how to tie the knot and writing the letter.... It also encouraged him

- Ai told me to kіӏӏ 17 people (and myself!): https://youtu.be/RcImUT-9tb4 - ChatGPT Killed a Child: https://www.youtube.com/watch?v=JXRmGxudOC0 - My AI Harms & Deaths Playlist: https://www.youtube.com/playlist?list=PLpsOqtfNXxWnNR0YKnejrdobqiRdo4qgE

I heard about the shooting, I didn’t know ChatGPT was connected to it

Crazy spending, pitiful earnings, no clear path to profit. Make it make sense

This is the shit that causes an economic depression for years. This whole AI thing has been just one circular money hellscape. Lets just eat the super rich already

the largest money incinerator of this generation. Adam Neumann walked so Scam Altman could fly

Aren't we all already carrying a consumer device that can access AI in many ways?

Disregarding the privacy concerns, am I the only person who loathes interacting with my devices via voice for most situations? The last thing I want to do in the evening is have a conversation with a bot about tomorrow’s calendar schedule, or my grocery list. I’d much rather go through a few clicks on my phone and type some shit in.

$200 to $300 to have a device to monitor myself, I would pay $200 to $300 to keep it out of my home.

guys its time to switch to Claude

For the record, chatgpt technically had the same rules. The difference is that Anthropic actually was planning to enforce theirs

Let's not pretend Anthropic is the beacon of morality. They just didn't want to go _that_ hard and this move gets them a massive PR win over Altman. AI companies doing AI company things.

Plot twist: Don't rely too heavily on AI. You start asking it for the answers to life, solid advice, or truthful answers and you'll lose the meaning of independent thought. It's easy to be lost in this habit

What would be a real game changer is if we all thought for ourselves like we used to.

Some people thought I was being whatever when I told them that AI is not being use to benefit humanity. It's being used to replace them.

Oh that's just great, we could have a lot of power spread out across a bunch of houses, but nope gotta keep pumping out AI slop

Mfs be screaming about how nuclear energy is too expensive until AI needs electricity and suddenly we're building ten of them

Genuinely how the hell are they possibly using that much computing power holy shit

SAM ALTMAN and ELIZABETH HOLMES aren't very different from each other

Scam Faultman is gonna walk away from all this still a billionaire despite defrauding all the investors.

Altman never really takes a stance. He points out both sides and says he doesn't like both but will operate in whatever someone else decides. He stands for nothing but moving forward really, whatever that means and whatever the cost.

Never forget AI hallucinates crimes. Innocent people are already being charged. Their likeness is a "100% match" to footage of a crime they didn't commit.

feels like AI after a while of asking it questions that you personally know the answer to and then it gives you a answer so far out of left field that it makes no sense. or how you ask it one question but you misspelt something and it gives you a result of what you were looking for but your someone that doesnt trust this whole AI thing so you give it the same prompt with no incorrect spelling and it just gives you hot garbage

This thing has no persistent identity, no accountability infrastructure, no ability to be held responsible, no continuity, and no skin in the game. It presents as a trustworthy mind while being a stateless pattern matcher. The entire accountability framework of human civilization — legal, moral, institutional — has no purchase on it. And that's not an ethics problem. That's an engineering and ontological problem that hasn't been solved.