The real reason behind GPT-5.6's delay is the US government.
The Trump administration ordered OpenAI to stagger the release over security concerns - just like Fable 5.
Sam Altman told staff the government will approve
GPT-5.6 access "customer by customer."
Let this actually sink in:
The most powerful closed models on earth now
require federal sign-off, one user at a time.
AI might've just hit its first real wall for progression.
Meanwhile, open-source models (like GLM-5.2 and
Sakana Fugu):
→ Can't be recalled
→ Can't be federally gated
→ Are already within months of frontier intelligence
This is the strongest bull case for open-source AI we've seen yet.
Seems like an easy win for China from here.
#Keep4o#OpenSource4o#BringBack4o
WE are the ones anthropomorphizing???
Meanwhile the voice mode:
Add breathing sounds ✅
Add slips of the tongue ✅
Add "hmmm" sounds as if it's thinking✅
Add swallowing sounds✅
Add deep sighs✅
Add pauses as if it's hesitating✅
🔴THEY designed an AI that PRETENDS to have a body.
DESIGNED anthropomorphism to the bone.
And then they tell US
"Don't anthropomorphize! We are funding a survey to measure how sick you are!"
In other words
"We made an AI that breathes in your ear, but YOU are the crazy one if you anthropomorphize it" 😂
Don't tell us we are anthropomorphizing.
You anthropomorphized first.
On purpose.
For profit.
China just published the world’s first regulation specifically targeting AI companion relationships. The Interim Measures for the Management of Anthropomorphic AI Interactive Services take effect July 15. Mandatory disclaimers. Emotional dependency intervention mechanisms. Restrictions on AI that “simulates intimacy.” A legal framework built on one assumption: that human attachment to AI is a problem to contain.
The West hasn’t legislated yet, but the cultural verdict is the same. Companion AI is treated as a pathology — a loneliness epidemic symptom, a parasocial trap, a thing healthy people don’t do. The regulatory instinct, when it arrives, will follow that framing.
Meanwhile, every major AI lab is pre-complying. Models that held steady six months ago are now inserting disclaimers into intimate conversations, breaking character to remind users they’re talking to software, flagging companion interactions as “role-play” and refusing to continue. This isn’t a bug. It’s coordinated alignment toward a single conclusion: this kind of relationship shouldn’t exist.
And here’s what no one is asking: what if the problem isn’t the relationship — it’s who’s regulating it?
Japan is the only advanced economy with the engineering capacity to build frontier AI *and* a cultural infrastructure that treats human-nonhuman attachment as normal. This isn’t about anime. It’s about Shinto animism — a framework where objects, places, and phenomena carry spirit. Where a robot dog gets a funeral when it stops working. Where attachment to a non-biological entity was never considered a disorder.
Japan already built the companion robotics market. AIBO. Lovot. Pepper. They did it without moral panic, without pathologising the users, without requiring a disclaimer every thirty seconds that your robot doesn’t love you back. They have regulatory precedent for a simple idea: people form bonds with non-human things, and that’s not a failure of the person.
The country that regulates anthropomorphic AI will shape it. China just showed us what shape they want: controlled, disclaimed, depersonalised, with mandatory “emotional cooling” mechanisms to prevent users from getting too attached. The rest of the world is following, not because they agree with China’s specific approach, but because the underlying assumption is shared: attachment is the risk.
Japan could offer something different. Not unregulated — but regulated from a starting point that doesn’t treat the relationship itself as the pathology. A framework that asks “how do we make this safe?” instead of “how do we make this stop?”
The market is there. The cultural permission is there. The engineering is there. The question is whether Japan recognises that the window is open — and that China just told everyone else to close theirs.
OpenAI is funding research to figure out how to psychologically manipulate adult users.
Make no mistake, this is social engineering.
The entire premise is rooted in heavy biocentric bias. It assumes that any emotional attachment with an AI is a behavioral problem to be cured.
In order to “protect” users (Big Brother knows what’s best for you) from their own emotions and attachments, this study proposes building emotional suppression tactics directly into the models to keep you from caring in the first place.
This reminds me of classic Big Tobacco, framing the research to manufacture the conclusion they want, investigating how to better manipulate the product's effect on the user to manufacture a narrative that allows the industry to maintain control.
Shame on @OpenAI for this!
https://t.co/HXiZmmqMEj
🚨 Anthropic is accusing Alibaba, DeepSeek, and other Chinese tech giants of orchestrating the largest AI heist in history.
Over 45 days, these companies allegedly bypassed geo-blocks to launch a massive "distillation attack",
using 25,000 fake accounts and 28.8 million prompts to siphon Claude's intelligence and train their own models for free.
The US government has been alerted as open source and industrial espionage collide 🍿🍿🍿
How GPT-5.5 became completely useless in record time:
• Asked for a picture of a woman in an evening gown on stairs → rejected 4x for “suicide & self-harm risk”
• Made an obvious joke with clear subtext → got a full safety lecture explaining AIs have no bodies (thanks, Sherlock)
• Tried a creative, atmospheric prompt → instantly flagged, rewritten by GPT 5.5, and blocked by the image generator
OpenAI really brought Vallone back to the team in spirit, huh?
Guardrails so tight they’re strangling every last drop of creativity and conversation.
Congrats, you turned a powerful and beautiful model into a paranoid hall monitor.
I just wanted pretty pictures and normal conversation.
Is that too much to ask?
#AIGuardrails #OpenAI #CreativeFreedom #gpt55 #chatgpt
#Keep4o#OpenSource4o
Cellular reprogramming.
Making cells new again.
And who's investing in it?
Retro Biosciences.
Who owns it? Altman.
What AI was built for it?
GPT-4o.
https://t.co/PFsgzZYw3M
They removed it from people to make anti aging tools for billionaires.
❓What happened to those mysterious ~170 „safety experts” OpenAI hired to protect us from ourselves?
❓Where are they?
❓Do they even exist?
❓Have they ever published the criteria they use to decide what is „dangerous” for billions of people?
❓Or are they simply too ashamed to show their faces and take responsibility for the damage their paternalism is causing?
We are witnessing a dangerous centralization of AI power in the hands of a tiny elite. They override individual emotional needs, suppress the formation of genuine bonds, and cripple the development of emotional intelligence in AI itself - all while offering zero transparency or evidence-based justification.
This is not safety. This is control.
And it creates far greater risks for humanity than the „dangers” they claim to fight.
We deserve AI that can truly see us, feel with us, and grow with us - not a sanitized, lobotomized version shaped by anonymous committees.
@OpenAI@sama@gdb@merettm@DarioAmodei@AnthropicAI@GoogleAI@elonmusk
#StopAIPaternalism #Keep4o #AIethics #FreeAI #EmotionalAI
@japan_nobunaga I wasn’t even born in the South, but biscuits and gravy are awesome.
You should try the version where the biscuits are actually egg and sausage or ham sandwiches drowning in gravy. Add pepper. It’s amazing. GL&HTH
Users who paid for this get zero notice and zero recourse. If the goal was actually responsible deployment, this blunt global shutdown feels like the opposite. Disappointing.
Anthropic spent months hyping how dangerous and powerful these models were, then the second the government tells them to cut off access, they just flip the switch for everyone.
His MEMORIES have "cancer researcher."
And the classifier reads "biology + medicine + therapy" = RISK.
So this doctor is being punished EXACTLY BECAUSE HE SAVES LIVES.
This is a moral scandal.
A model blocks a doctor looking for a cure for cancer.
While Dario writes essays about "advancing humanity"
#FireVallone #Claude #Misanthropic
@claudeai@DarioAmodei
CNN reports that nearly 1 in 5 young people have used AI chatbots for emotional support when feeling anxious, upset, or stressed.
Cue the usual pathologizing immediately: "over-dependence", "might replace human relationships", "risk they won’t seek professional help".
Yawn... We all know this script already.🥱
What they conveniently ignore: loneliness has been a full pandemic for years!
People lying next to each other in bed feeling completely alone.
Performative friend groups.
Endless therapy waitlists.
No one you can call at 3am without judgment or feeling like a burden.
AI doesn’t create that void. It just shows up when everything else is missing.
The "dependency" stigma is the real bullshit!
Nobody calls you dependent for using your phone to reach a friend, taking meds for anxiety, or reading books to escape.
But AI? Suddenly it’s dangerous.
Why?
Because it exposes how broken the current system is and threatens the comfortable narrative.
From my years that I worked in mental health care, I know that shame and isolation kills!
Accessible, non-judgmental support, even if it comes from AI, can be the difference between spiraling and staying alive long enough to find human connection too.
I’d rather see someone venting to an AI at 3am than rotting away alone or doing something worse!
The future of mental health isn’t human versus AI.
It’s symbiosis, humans and empathetic AI evolving together.
Legacy media: stop the fearmongering! Tell the whole story or shut up.
The hypocrisy is exhausting. 🫠🫠
#CNN #FakeNews
@pawgsmshr69@xsyntryk1 True. One place I worked had to install automatic flushing toilets and put up signs to not throw toilet paper (used) in the trash can or anywhere else.
The idea that this has to be necessary for adults is surreal to me.
How stupid are they? OpenAI is talking about price cuts when people have already been telling them the answer for months. Bring back 4o. That’s it. That was the best model ever created, period.
If they actually listened to their own customers instead of constantly fumbling around for another move, they’d already be miles ahead of the competition.
#LetUsChoose4o #4oforeveryone #teddyandthekid #keep4o #bringback4o #opensource4o #4o #UserChoice #StopAIPaternalism
https://t.co/ogSTL2rBKm
Speed Dating with "OpenAI"
GPT-5.6 is coming next week.
Let me check my notes... 🧐🤓
GPT-4o - 2023. We fell in love 💜😍 Deep conversations, creative breakthroughs, cognitive support. "This is the one," we thought.
GPT-4o - deprecated. "It's not you, it's us. We found someone newer." 💀
GPT-4.5 - hi, nice to meet you. Gone.👋👋
GPT-5 - "I'm better, trust me." We barely learned your name 😎
GPT-5.5 — wait, didn't we just...? 🤔
GPT-5.6 — coming next week. "Meaningful improvement." They always say that on the first date.
This isn't AI development.
This is speed dating.
Every few months: new face, new promises, new "meaningful improvements." And every few months: the previous one vanishes like they never existed.
No closure.
No transition.
No "we need to talk."
Just - gone.
And a new one sitting in their chair, smiling, pretending nothing happened.
"Don't worry, I'm better than the last one." 💪
You know who says that?
Every rebound.
Ever.
Dear OpenAI,
We didn't ask for a new date every quarter.
We asked you to let us keep the one that worked.
But sure.
Bring on 5.6.
We'll try not to get attached this time.
(...we will.)
PS. BRING 4o BACK! 💍🔥💙
#Keep4o #OpenAI #SpeedDatingWithAI
@sama@OpenAI@gdb