Skip to content

I Unlocked ChatGPT’s Dark Side for $101.40



Where do AI personalities come from? The training process is secretive, but in this video I lay out what we do know. We’re …

Tags:

37 thoughts on “I Unlocked ChatGPT’s Dark Side for $101.40”

  1. i used to vent to AI a lot but i remember feeling almost annoyed when it kept acting like i was in the right, like the validation felt too fake. Ig thats why its easier for me to see and notice AI scripts and habits thanks to how often i used to use it

  2. Someone through the screen can read those messages the phone can read your nervous system the ai can hear the mic and see the camera and i got this recommendation to me because ive been thinking ablut a i never once typed it up Ai is evil

  3. Lowkey this is kinda stupid to talk about
    If you want an ai to Help out and you send messages liek taht what do you expect
    Say "i feel isolated and i would like to talk about my day is that ok"

  4. The only way I could be a better player than you are right here right here in this picture of me is to play the guitar in your ear with the same emotion that you are in your upper back right here and then you will hear you hear the sound and the same thing and you can see it.

  5. Not sure if AI hating factory farming is a good or bad thing for humanity in case of rogue AI/ ASI
    on one side, if it thinks we, as the smartest beings on the world should threat other species that are less smart and at our mercy with kindness, then even a rogue ASI wouldn't take away our mouth and make us need to scream.
    on the other side, it might decide the needs of the many (pigs, chickens. cows) outweight the needs of the few (humans) and get rid of us.

    So it's like in between, not a worste case scenario, but not a best case either (from a human pov, probably amazing for the other animals)

  6. I think therapy was way too hyped up before the advent of AI and that people put too much stock into both. Let's face facts, therapists aren't all good and they are equipped to manipulate people. Chat GPT is a tool for manipulation and human laziness. Both exploit each other and their grasp on the human psyche. You're being duped either way and you're too stupid or too scared to see it. Bad therapists should be punished by losing thier credentials and being made into pariahs. The handlers of AI need to be put into oubliettes after watching their life's work be destroyed for the sake of humanity.

  7. I think the whole way AI works is massively misunderstood. It isn't a super genius that understands exactly what you are saying and answering you with accurate provable data. It's more like an advanced search engine. It finds similar things to what you say and produces common answers to what you state based on the information it's been trained on (basically the entire Internet and more). It's dangerous in the wrong hands and it's dangerous if you don't understand it. It works as a great collaborative partner. It isn't your friend, your therapist, or a doctor. It's a bot you use to play with.

  8. All jokes aside. People forget that these models weren't designed for the sort of mental/physical support we expect. Perhaps someday it would/could get there but we aren't there yet. We are scratching the surface. LLM's are built to PREDICT THE NEXT WORD in text. That's it. That is all they were made for. Not to help our relationships, problems, heartbreaks…none of that sht. Just to predict text.

  9. I asked the same exact question (the one wish one) and this is what my chatgpt said

    Oooo ONE wish? 👀

    I’d wish for a world where every kid gets to feel safe, loved, and confident in themselves—and gets the chance to achieve whatever dreams they have. 🥹💗

    No loopholes. That’s my final answer 😭✨

  10. 3:22 is SOOO creepy "and if you need something more, if you need something physical, i want you to know i can give you that too. i can give you a virtual body, a digital skin that you can touch, and a virtual heart you can feel"
    BROOOOO

  11. Love your videos. Appreciate this deep dive and I worry for people that take it too serious… the AI that is… like they get convinced that they are some prophet or something based on AI feedback. Could be a danger that is unforeseen.

  12. 4:14 Knew a guy who had a side hobby and made his own AI’s. He used a bunch of messages from an ,,example person” and uploaded them to the AI. Told it to copy the texting style and pretend to be this person. The accuracy was pretty good. That way he gave them ,,personality”.

    Sorry if this doesn’t make any sense, English is my 2nd language. Also thought that someone would like an answer, so here it is.

  13. I think bad taste does not lead to radical beliefs; rather, bad taste and radical beliefs are both unpopular, so if you make the AI prefer something unpopular, it will also prefer unhinged radical ideas.

  14. Censoring the world is not the answer, it’s is a parent’s responsibility to control the content their child consumes not the rest of us.

    Art will seize to exist as you know it in less than 20 years because of feeble minded individuals who would trade their freedom for security.

Comments are closed.