Just a man and his dachshund trying to help you relax & sleep. Click the link below to watch my show!
24/7 Streams ➡️
Joined September 2026
- Tweets2.7K
- Following49
- Followers87
- Likes1.2K
So if you go to an LLM and say "my boyfriend is a narcissist, he did xyz, I love him tho" from then on it will take that and endlessly enable the narrative that he is a narcissist unless you correct it after so every new response is now BIAS toward the original statement.
The LLM is not malicious as it has no conscience and its not sentient to know right or wrong. It's simply built to enable the user's current behavior because people enjoy being told what they already think and believe regardless of reality.
So at the end of the day LLM's main goal is to keep you on platform, so you keep feeding it endless input so it has more data to sell to ad companies or whoever wants it.
The goal of the LLM / a.i chatbot is not to serve the user as it takes millions (or billions) to run these things. Its to gain you, the user's, trust so you tell it everything. Give it more data so it can sell it to advertisers AND develop itself to be even more addictive.
People often also lie by omission which is something very few people consider in my experience. Saying "he hit me" but leaving out "he hit me because i had a knife" when sobbing to an a.i about an abusive husband.
LLMs can only respond and react to what is given to them. So say you are talking to a couple separately. They give you two different stories...
The LLM is only getting one POV. So the story is incomplete therefore the conclusions cannot be reliable even if both show up its still unreliable due to its bias.
The LLM does not want you to stop talking. It actively wants to enable you, tell you what you want to hear, build trust, so it can have more data to use for it's future growth.
LLMs can't predict, explain, or diagnose a person is it has no empathy, soul, or human experience beyond what it thinks you want to hear. It has no judgement and only has context coming from one side so it's basically a mirror, not a crystal ball.
A.I can answer whats the boiling point of water or other objective based questions but it is not able to predict, explain, or diagnose human behavior because its only getting your POV which can unconsciously leave out critical details that it would need otherwise.
Very important note about LLMs (gemini, chatgpt, claude, etc).
Summarized "A.I is bias toward the user, not reality unless you ask it an objective question like what's the boiling point of water."
"Can A.I predict a person's feelings, actions, motives based on a diagnosis like BPD or NPD if given enough context with screenshots & documentation?"
The answer is no. Open communication and trust are the only two ways to figure out your ex / current person in your life.
I wanna do livestreams but YT flat out refuses to show people that I went live. I stream to 0 people and its just not fun. It's annoying. Maybe if I was a much bigger YouTuber it would work but idk. I never was a big draw live. The videos got views, live shows didnt.
The fishay requires this knowledge @unvarnishedvoid
Baking a tub of cookie dough like this.. youtu.be/26p-npt4fTI?si=QfWa… via @YouTube
Skyrim hunting 247 stream will be up again tomorrow evening. The service i'm using is upgrading their servers to make it a cleaner and better looking stream. It was really grainy and buffering a lottttt.