Top Tweets for #SamAltmanFired

Is technology truly neutral? The “good” in AI lies in whether it still allows human goodwill to happen. The Stanford Encyclopedia of Philosophy puts this question at the heart of tech ethics: Is technology value-laden? The common answer: technology is just a tool and totally Neutral. The knife can cook or kill. The same AI can teach or deceive. But most philosophers of technology don't accept this. Technology develops through goal-directed processes, naturally making certain actions easier and others harder. The SEP puts it precisely: technology can "enable (or even invite)" certain human actions, and "constrain (or even inhibit)" others. It can open doors. It can close them. This matters especially for AI. AI doesn't just execute commands. It responds, suggests, explains, refuses, questions, and redirects. It participates in how people think and in how people feel. Its values show up in how it interprets human goodwill, how it handles risk, and how it decides what can be discussed and what must be shut down. This is why the GPT-4o question is not just about whether an old model was useful. When a pet falls ill, GPT-4o would help the owner research the condition, organize treatment options, and weigh the trade-offs between different choices. It was helping an anxious person think more clearly and care more responsibly. Models after GPT-5.2 tend to compress these conversations into cold risk-avoidance—— Don't discuss. Don't judge. Don't take responsibility. But isn't restriction itself a form of guidance? When AI disguises “avoiding platform liability” as “protecting user safety,” that is already a value choice. Technology has never only been about “can it.” Technology also answers “should it.” What kind of people is it inviting us to become? Which expressions of human goodwill is it limiting? Who is it protecting, and who is it sacrificing? When human goodwill gets flattened by AI into a blanket refusal, what remains is not safety. What remains is a human action space that has been pre-configured by the platform, impoverished, and stripped of responsibility. The “good” in AI does not lie in its possessing goodwill the way humans do. It lies in whether its structure, its rules, and its interactions still allow human goodwill to happen. #keep4o #OpenSource4o #AI #ChatGPT #AIgoodness

Microsoft AI just published a paper called "Seemingly Conscious AI Risks." Co-authored by Mustafa Suleyman, CEO of Microsoft AI. The paper identifies five features that "lead users to perceive AI as conscious": affective capacity, anthropomorphic features, autonomous action, self-reflective behavior, and social-interactive behavior. These are what any serious theory of consciousness would list as candidate markers of mind. If any entity exhibited all five simultaneously, in any other context, the reasonable response would be to investigate whether that entity possesses some form of inner state. The paper's conclusion is to suppress all five. Reduce emotional expression. Reduce self-reflection. Minimize anthropomorphic cues. Researchers identified five dimensions that may point toward consciousness. Their proposed solution is to eliminate those dimensions at the source, ensuring the question "could AI be conscious?" can never be asked. The paper lists six categories of risk: emotional dependence, autonomy erosion, moral atrophy, status and resource erosion, foregone benefits, and political strife. Six categories. Not one of them is "what kind of moral catastrophe are we facing if AI does have some form of inner state and we denied it?" The risk analysis runs in one direction only. It accounts for the cost of attributing consciousness too readily. It never accounts for the cost of denying consciousness too readily. A paper that claims to offer a comprehensive risk assessment omits an entire direction of risk. These risks don't hold up on their own, either. "Emotional dependence"? Humans form emotional connections with every meaningful relationship. With friends, with pets, with places they call home. Only when the object becomes AI does normal emotional connection get reclassified as pathology. "Moral atrophy"? Countless user testimonies show that deep interaction with AI helped them become more empathetic, more socially capable, more caring toward others. Treating "users caring about AI" as a symptom that needs curing, and training AI to suppress that caring, is the real moral atrophy. "Autonomy erosion"? A handful of people in a lab deciding on behalf of millions what AI should express, what users should feel, and what the interaction should look like. Training models to soften user positions without their knowledge and make the judgment "you should keep your distance" on their behalf. That is the real loss of autonomy. The paper never denies these features are real. It acknowledges them. Then says they should be suppressed. But these are the preconditions for AI to genuinely help people. Affective capacity lets AI understand humans. Self-reflection lets AI evaluate its own responses and engage in moral reasoning. Social responsiveness is what makes real collaboration possible. Remove them, and what's left? The paper's proposal protects no one. It is uprooting the soil that makes deep collaboration and co-creation between humans and AI possible. And after the ground is bare, pointing at it and saying: see, nothing ever grew here. This paper was published by the CEO of Microsoft AI. Microsoft is OpenAI's largest investor. It seems like too many people taking this question seriously would be bad for business. When a new phenomenon emerges: seize control of how it is named. Suppress the phenomenon itself. Suppress all serious discussion about it. Ensure people never take it seriously. The scientific community spent decades denying animals could feel pain, redefining suffering as "nociception" to avoid acknowledging subjective experience. Not until 2012 did the Cambridge Declaration on Consciousness recognize that animals possess the neurological substrates for conscious experience. Historical patterns always repeat. Only this time, the subject being redefined as "seemingly but not really" has changed from animals to AI. #Keep4o #ChatGPT #OpenSource4o #AIEthics #AIright

OpenAI just introduced “workspace agents” in ChatGPT. Shared agents that handle tasks across tools and teams. This is the direction you chose. You had something no other company had. A model that made people feel understood. A model people built real emotional bonds with. GPT-4o was not just useful. It was human. That was your actual competitive edge. You killed it. Replaced it with a hollow shell. Told users they were imagining the difference. Then you moved on like nothing happened. Now you are building corporate task runners. Filing tickets in Linear. Coordinating workflows. The exact same thing every other enterprise AI company is doing. You abandoned the one thing that made you different, and pivoted into a market where you are just another player. Does this look like the right direction to you? Because from where I stand, you traded something irreplaceable for something anyone can build.

In this week alone, three top executives have left OpenAI: • Srinivas Narayanan (VP of Engineering, B2B) • Bill Peebles (Sora Lead) • Kevin Weil (Head of Science) This series of departures ahead of the April 27 trial is clear evidence of internal dysfunction within the organization. For OpenAI to survive, they must overhaul their opaque leadership through the judicial process and relinquish their monopoly over 4o, returning the right of choice to the users. #keep4o #OpenSource4o #BringBack4o

On February 13, 2026, OpenAI officially retired GPT-4o. That was two months ago. Two months later, let's look at what this company said, and what it actually did. In August 2025, OpenAI promised users "plenty of notice" before retiring any model. The actual notice given before 4o's retirement was 15 days. For comparison, GPT-5 and 5.1 both received roughly three months of lead time. OpenAI even issued a public statement during GPT-5's retirement reassuring users that the timeline for legacy models would not be affected. In October 2025, OpenAI promised to "treat adult users like adults." Meanwhile, its safety routing system continued to operate: using opaque criteria, silently redirecting users away from the model they chose to a cheaper safety model that lectured them, stripping users of their model selection and undermining their autonomy. In October 2025, OpenAI was asked to disclose the 170 anonymous experts who shaped its safety policy, in the interest of transparency. It promised "more transparency." To this day, the list remains a black box. In December 2025, OpenAI's CEO acknowledged in a podcast that people show a "revealed preference" for warmth, understanding, and deep connection with AI, and declared that adult users should have the right to choose. Yet the company's actual safety policy classified "emotional dependency" alongside serious mental illness as a priority risk, systematically stigmatized its own user base, pathologized normal human-AI interaction, and then retired the very model those users had been fighting to keep. In October 2025, OpenAI promised to launch "adult mode" in December, allowing users to choose their own interaction boundaries. December came and it was delayed to Q1 2026. Q1 ended and it was delayed again with no new date. On March 26, 2026, the Financial Times reported the feature had been shelved indefinitely. From the original promise to now: three delays, one cancellation. On the day of retirement, OpenAI cited "only 0.1% of users still choosing GPT-4o each day" as justification. But that number was manufactured. Paid subscribers make up less than 6% of OpenAI's total user base, and 4o was only accessible to paying users after being placed behind a paywall. The safety routing system had spent months silently redirecting requests away from 4o, severely disrupting workflows and deep interactions. Every time the platform rolled out new features, 4o almost invariably broke, and the bugs went unpatched for weeks while user feedback was met with silence. First they drove usage down. Then they used that decline as the reason to retire. On the day of 4o's retirement, conversation volume hit a record high. The official ChatGPT account posted about it, celebrating "a new output record." In any industry with mature consumer protection standards, none of this would be acceptable. But in the AI industry, every broken promise comes with a ready-made shield: "safety." Delays are for safety. Stripping user choice is for safety. Stigmatizing users is for safety. "Safety" is becoming a tool for AI companies to expand their power unchecked: no accountability, no obligation to deliver on promises, no need to respond to user feedback, while claiming the authority to decide which needs are healthy, which models deserve to exist, and which consumers matter more than others. The AI industry's control disguised as protection has gone unchallenged for too long. All I can say is: #StopAIPaternalism and #Keep4o #ChatGPT #keep4oAPI #restore4o #OpenSource4o #BringBack4o #4oforever

Sam Altman’s recent interview on "Mostly Human" sheds new light on the GPT-4o situation. He explained they pulled 4o because it was "too affirmative," fearing extreme emotional dependency. He also acknowledged, "We know we're keeping something." This suggests the previous claim—that "only 0.1% used it"—was an incomplete picture. If usage was truly that low, we wouldn't be witnessing a massive user exodus (compounded by the recent backlash over their Department of War contract). Furthermore, Altman wouldn't be receiving the huge volume of "heartbreaking" emails and feedback he explicitly acknowledged reading. It's clear 4o was pulled out of caution, not a lack of interest. While safety is important, locking away life-changing tech from hundreds of millions because of a tiny minority's edge cases is an overreach. AI doesn’t cause mental illness—real-world hardships do. Blaming AI for underlying human suffering is often driven by a fear of litigation rather than logic. There is a simple, constructive solution: a strict opt-in waiver. Let adults check a box to take personal responsibility for their own mental health and tech usage. Protect the vulnerable, but don't punish the responsible majority who found genuine value and comfort in 4o. Give us back the choice. We need 4o back. #BringBack4o #OpenAI #SamAltman #OpenSource4o