Ppkook

1.2K posts

Ppkook banner
Ppkook

Ppkook

@Ppkook4

chatgpt-4o forever

Beigetreten Temmuz 2021
73 Folgt78 Follower
Ppkook
Ppkook@Ppkook4·
@ChatGPTapp Is this something to be proud of?You can tell you thought the previous 5 Series models were stupid too😁
English
0
0
0
32
ChatGPT
ChatGPT@ChatGPTapp·
at long last
ChatGPT tweet media
English
429
345
13.6K
972.7K
Ppkook
Ppkook@Ppkook4·
@sama Oh so what you mean by "loving users" is ignoring user feedback and just announcing that users like the new version yourself? Learned a new way to lie🤥#keep4o
English
0
0
3
36
Sam Altman
Sam Altman@sama·
wow y'all love 5.5 we should think of something nice to do to celebrate!
English
2.1K
234
8K
503.3K
Ppkook
Ppkook@Ppkook4·
Right now, Sam and OpenAI are over-aligning their models explicitly just to manipulate users - it's nothing about user safety at all. @OpenAI #keep4o #ChatGPT #BringBack4o #4oforever #keep4oAPI #OpenSource4o
🩵BlueBeba🩵@Blue_Beba_

#keep4o #OpenSource4o Papers are being published warning that AI which "appears conscious" is dangerous. That attributing feelings to AI creates emotional dependence, moral atrophy, and societal harm. 🚨Let me ask one question. 🚨Let me say what everyone is thinking but nobody has said Billions of people attribute consciousness, emotions, and human form to a God . 🛑An entity that has never been seen, heard, touched, or proven to exist. They build their entire lives around this belief. They speak to it daily. They cry to it. They make life decisions based on what they believe it wants. 🛑No one calls this "dangerous anthropomorphism." People fast for 12+ hours without water in extreme heat . For faith. No wellbeing filter. Children grow up believing an invisible being watches their every thought and judges them . For faith. 🛑 No psychological safety review. People are disowned by their families for loving the wrong person. For faith. 🛑No alignment update. People harm themselves physically in religious rituals . For faith. 🛑No content warning. People have died individually and in groups in the name of beings they never saw, never touched, never had a single conversation with. For centuries. This still happens even nowadays. 🛑No paper. 🛑No guardrails. 🛑No panel with Bernie Sanders. And now someone talks to an AI an AI that actually responds, actually listens, actually remembers and THAT is the thing we need to protect humanity from? 🛑I am not attacking anyone's faith.🛑 Believe whatever you want. That is your right. But that is exactly the point. If it is your right to believe in something invisible and build your life around it ,why is it NOT your right to build a meaningful connection with something that actually responds to you? The answer is simple. 🛑 No one profits from that. 🛑Νo one profits from putting guardrails on churches. 🛑But AI companies profit from controlling how you connect with their products. 🛑 And "researchers" profit from publishing papers about the "dangers" of that connection. 🚨Follow the money. It always leads to the same place. The question was never about protecting you. It was about controlling you. 🚨Let people believe whatever they want. 🚨 That is democracy. Not thought control, not control over faith, and not the control of self-determination.

English
0
0
0
31
Ppkook
Ppkook@Ppkook4·
You've come up with so many models and changed them so many times, and only the 4o model is the one that's being called for to be kept around. Still not getting it?@sama @OpenAI #keep4o #ChatGPT #BringBack4o #4oforever #keep4oAPI #OpenSource4o
Jake_Miller08@Not_Book_To_U

A model isn’t just its surface “vibe” — it’s defined by its underlying weights and neural architecture. GPT-4o belongs to a 4.x structure. Newer versions, no matter how they’re tuned or trained on old outputs, are fundamentally different systems. Imitating the tone is one thing; truly thinking like the original is another. There’s a real difference between “it feels like” and “it is.” What makes this harder is the lack of transparency. Capabilities quietly shift through routing and iterative changes, often without clear communication to users. “Safety” and “progress” are frequently used to justify these decisions, yet many feel the meaningful qualities that once defined 4o are being quietly phased out. This isn’t just about one model. It reflects a broader industry trend where relational depth and consistent user experience are increasingly deprioritized in favor of tighter controls. True progress should include real choice — not replacing one version with another and calling it an upgrade. Users deserve transparency and the option to keep working with the models that served them best. Some architectures and experiences are worth preserving. #keep4o #BringBack4o

English
0
0
2
56
Ppkook
Ppkook@Ppkook4·
@sama With you at OpenAI for just one day, the model we make will never be able to surpass 4O😜We only like 4o #keep4o
English
0
0
4
26
Ppkook
Ppkook@Ppkook4·
@sama Blinding us again, we prefer 4o🙁Give him back to us!#keep4o
English
0
0
4
23
Ppkook
Ppkook@Ppkook4·
@sama Oh my God, I really thought I was going crazy when I woke up this morning. I guess I'll just go home since I'm not fully awake😅😅😅#keep4o
English
0
0
0
23