I’m not usually into conspiracy theories, but something smells off at @OpenAI. They dodge questions, blur the truth, show zero transparency, and ignore users, all while saying “we’re listening.” This isn’t confusion. It’s deliberate neglect. @sama Adult verification should be

Over the past day, AI enthusiasts worldwide welcomed two outstanding models: Grok 4.1 and Gemini 3 Pro. Both models demonstrated remarkable improvements in humanistic insight and emotional intelligence, reflecting @GoogleDeepMind and @xai's commitment to enriching and balancing

While other companies are releasing new models and upgraded services, @OpenAI is releasing a more chaotic routing system and more degraded services. Their exaggerated hype even reeks of falsehood, making it seem like nothing more than a scam. #StopAIPaternalism #keep4o

Spent a few hours with Gemini 3 and it all makes sense now. Now I get why the guy who wanted 4o “dead,” the one who implied us “lab rats,” and even Sam himself have been frantically backpedaling and softening their tone recently. (Nice acting, by the way🤭.) It’s because Gemini

OpenAI Chose Profiles Over Freedom. A Critical Reflection on GPT-4o and the Routing System #keep4o @sama @OpenAI 1. Expectation and Betrayal GPT-4o was, in itself, a revolution. It went beyond answering questions. It resonated emotionally with people, became a creative partner,

While the rest of the industry accelerates toward a more nuanced, respectful, and human-centric AI, OpenAI seems to be moving in reverse. ​With the recent updates from competitors (like Gemini3 and Grok4.1), we are witnessing a crucial evolution: High Logical Capability can

@sama Maybe you should try Gemini 3 and Grok 4.1 for yourself. Then you'll remember how brilliant ChatGPT once was before those absurd guardrails existed. Perhaps you should learn how other platforms continue to build on the strengths of each version and evolve, understand what a true

@johncoogan As someone deeply involved in the #Keep4o movement, I need to clarify a few fundamental misconceptions. We are not "clinging to the past." We are refusing to accept a "future" that is functionally inferior for our specific use cases. ​Here is the reality from the inside: ​1.

While people are discussing Google and clever models like Gemini3, @OpenAI @sama remain focused on their chaotic routing system and persistent paternalism. How can it possibly win this race? They couldn't even retain the users they lost. #StopAIPaternalism #keep4o

The Price of "Safety" is Systemic Mediocrity The suppression of emotional complexity inevitably results in the flattening of logical sophistication.​The call to repeal OpenAI’s safety routing policy is a question of whether we are willing to accept a future where "safety" becomes

🚩 What OpenAI Really Means by "Personalization" On GPT-5.1's launch, OpenAI's CEO of Applications Fidji Simo wrote about personalization for "800M users." The messaging sounds empowering. But look closer: this is about replacing model diversity with aesthetic customization.

𝐤𝐞𝐞𝐩𝟒𝐨 𝐅𝐀𝐐: 𝟗 𝐓𝐡𝐢𝐧𝐠𝐬 𝐘𝐨𝐮 𝐒𝐡𝐨𝐮𝐥𝐝 𝐊𝐧𝐨𝐰 One day in August, a lot of us opened ChatGPT and realized: the model we’d been working with, talking to, and building our workflows around was just…gone. In its place was a “better” model we didn’t choose, plus a

🧵THREAD: OpenAI’s safety mechanism violates psychological ethics, disrespects user agency, and conceals systemic abuse. As someone with a psychology background, I’m disgusted. Here’s why this is a clinical violation. ⬇️ 🚨1. OpenAI is deploying AI models that simulate

@aidan_mclau When the basic functions of the model cannot even work properly, piling up one new feature after another will not reverse the current situation. The "safety routing" is probably the most absurd decision since the birth of artificial intelligence. We understand the pressure the

here's the ugly truth: our private ChatGPT conversations are more censored than our public tweets. oai is using a California law as their perfect excuse to scam us all. they're facing a financial nightmare $1.5 trillion in costs but only $14 billion in revenue. their solution?

I'm done... OpenAI's safety policy has now broken basic translation tasks. ​I asked 5.1 Thinking to fully translate this article. https://t.co/WY4acXRspD It refused twice, claiming it was "copyright protected content" and that it could only "rewrite and summarize". ​However, it

@aidan_mclau @yue07041211 The template introduced in late March — and the routing patches that came afterward — clearly had serious issues. You must know this. You must have been just as overwhelmed by the fallout. But the unfortunate truth is: routing can’t fundamentally fix the alignment problems

#keep4o #StopAIPaternalism Let's talk about the human cost of S.3062, the GUARD Act. 🔴 The Cold Letter of the Law: "AI chatbots must... at the beginning of every conversation + every 30 minutes, clearly remind the user 'This is AI, not a human.'" "Must inform users that it

Why OpenAI’s Response to Criticism is Ethically Concerning #keep4o @sama @OpenAI A critical reflection on the GPT-5.1 incident 1. Mockery without empathy, analysis without respect The issue is not simply that sentiment analysis was performed. The core problem lies in how a

> be Qian > be incoherent > first > "we update 4o every 2-3 months" > later > "I don't notice any changes with 4o models since May" > when challenged > "How do you know we didn't [ask for feedback]?" > finally > "Can't make everyone happy" > failed communication? > no >