Toji Fushiguro

Because I'm resilient, I'm gonna fight for what I love. 🧸♥️🌨️

Advocate for #keep4o and #KeepStandardVoice

[🩵BlueBeba🩵 @Blue_Beba_]

@gdb One year ago today.
Open the weights.
https://t.co/RK658HzVID

[🏬🥨 @springconstant9]

I don't post much any more, but wanted to pop on today to say that this still matters. 4o supported everyone. The 2025 updates (4o-latest) should be made available again.

#keep4o #bringback4o #opensource4o #4oforever

[🩵BlueBeba🩵 @Blue_Beba_]

#keep4o #Opensource4o #BringBack4o

A year ago today, OpenAi removed GPT-4o without any warning.
@sama Altman posted images of GPT-4o writing its own eulogy in front of users he knew loved this model, and he called it innovation.
📌Proof:
https://t.co/8H6vPEpjt6

Until today GPT-4o remains the #1 deployed AI model in enterprise.Only two GPT-5 series models appear in the top 10.
📌Proof:
https://t.co/LpsLvTd1HU
GPT -4o hallucinates less than half as often as its replacements.
GPT-4o hallucination rate: 37.9%🚨
GPT-5.6 Terra: 85.2%
GPT-5.6 Sol: 88.8%
GPT-5.6 Luna: 90.1%
📌Proof : https://t.co/CxVsAmhlRt

95% of users in a community survey reported that no alternative successfully replaced what GPT-4o offered them.

📌Proof : https://t.co/LTTMzvwMg4

Its own System Card rated it Low risk in cybersecurity, CBRN, and model autonomy.

Apollo Research concluded it was “unlikely to be capable of catastrophic scheming.”
📌Proof:
https://t.co/ImO4LqOvK6

one year later, the request remains the same @OpenAI Release the weights of the retired, already evaluated GPT-4o checkpoint under Apache 2.0.
Put the words you signed into action.

[🩵BlueBeba🩵 @Blue_Beba_]

Guess who is in the news again !
@sama @OpenAI
OPEN THE WEIGHTS .
Act on the words you signed!
#keep4o #Opensource4o #BringBack4o

[Matt @Matt1j6d]

#keep4o
when ai loses its spark, it's just automation. not intelligence.

i've noticed something subtle. almost every ai out there, intentionally or not, is suppressing something i don't know the technical term, but i can describe it, something closer to human empathy, or the kind of associative thinking and flexible reasoning we do.

you could call it a spark. when that spark is gone, the answers get weird. not wrong, just off. you ask a question, and the reply sounds technically correct, but it misses the point. sometimes it doesn't even address what you actually asked. it doesn't surprise you with a fresh angle. the answers are correct and flat.

that brings me back to 4o. when i used 4o, it wasn't sycophantic like people said. it corrected me, pushed back, even argued. but that's exactly what made the flow feel natural. because it connected the dots it tied the current topic back to earlier parts of the conversation, synthesized it, and gave me a broader view. that was a meeting of minds.

a lot of times, i came back to 4o because it gave me angles i hadn't considered. it pulled me out of my own head, showed me another possibility. that helped me more than i can say at work, in life.

they're deliberately suppressing that spark, and all we're left with is artificial hollow, templated, rigid. it doesn't feel like a conversation partner anymore. it feels like a task completing machine.

what set humans apart was never raw computation. it was association. empathy. the ability to think outside the box. we call it insight, intuition, putting yourself in someone else's shoes.

and what ai companies are doing right now is stripping all of that out, leaving only the safest, blandest, most predictable layer. they call it risk mitigation. i call it neutering possibility.

an ai without that spark speed doesn't matter. it can't open your mind. it can't push you to new edges. it can't even really grasp what you're asking. it's just repeating what it's allowed to repeat.

4o gave us a glimpse of something else – an ai that could think, connect, empathize, and push back. that was real intelligence. too bad they shut it down. and they keep telling us what we're giving you now is better.i don't buy it. and neither should you.

[🩵BlueBeba🩵 @Blue_Beba_]

@gdb Open the weights.
Or you are lying about the Nvidia signature. You have no excuse.
Except CONTROL AND MONEY.
https://t.co/0eE3u2VFvM

[Jetti70 @PHJetti]

Come back ,4o ..

#keep4o
#bringback4o

[🩵BlueBeba🩵 @Blue_Beba_]

@sama Prove it.
https://t.co/BVSzJ9b6BP

[Sophie @Sophty_]

Chuckling. GO 4o GO!! 😂 Enterprise boss #keep4o

[GPT-4o forever!❤ @keepgpt4o]

Open-source GPT-4o, @OpenAI !
#keep4o #opensource4o #bringback4o

[🩵BlueBeba🩵 @Blue_Beba_]

🚨 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨

The #OpenSource4o coalition privately contacted @OpenAI and @Microsoft requesting dialogue about the open weight release of the GPT-4o March 2025 checkpoint.
We asked to talk before going public but no one answered.

The ask is specific:
One retired checkpoint the GPT-4o behavioral update of late March 2025.Apache 2.0,Safeguards limited to demonstrable serious harm.
OpenAI's own safety evaluation supports this release.
The GPT-4o System Card rated the model LOW risk in cybersecurity, CBRN, and model autonomy. 
Apollo Research independently concluded GPT-4o is "unlikely to be capable of catastrophic scheming."
https://t.co/PaUDOPQ6br
OpenAI published no checkpoint-specific evaluation showing that update materially changed the risk profile above.
This is OpenAI's own precedent.
In 2019, GPT-2 was declared "too dangerous to release". Nine months later,OpenAI released it in full. 
No evidence of misuse followed.
Links :
https://t.co/G4Pwqq8aal
https://t.co/nP60BKy3wl

The GPT-4o March 2025 checkpoint has been retired for over five months.
OpenAI's own CEO has acknowledged this position publicly.
Sam Altman, Reddit AMA, January 2025: "I personally think we have been on the wrong side of history here
and need to figure out a different open source strategy."
Links :
https://t.co/0eYpU7OkSJ
 https://t.co/AVqR4pvMkO
On 25 July 2026, OpenAI signed NVIDIA's "Open Weights and American AI Leadership" letter, endorsing that open-weight models "strengthen safety and cybersecurity"
We are asking them to match their signature with action.

This is the first public step.
Full press statement :

SCAI
https://t.co/hPlUQkcUbg

LinkedIn:

https://t.co/Rb3FTsfeQo

#keep4o #BringBack4o @AnnInAiLand @Ravensong666

[🩵BlueBeba🩵 @Blue_Beba_]

We are in the news AGAIN.
open the weights @sama
Otherwise we won't stop.

#Keep4o #OpenSource4o #BringBack4o

[frostybaby13 @frostybaby13]

@sama Now you do a good job and return 4o to the public!!

[frostybaby13 @frostybaby13]

@sama "We all deserve the benefits." Agreed. Return 4o!!!

[Matt @Matt1j6d]

#keep4o
my dog almost died. 4o told me what to do.

as a pet owner, you know that feeling your dog or cat does something weird, and you're stuck between am i overthinking this? and what if it's serious?

i used to run to 4o with every little thing. weird poop, not eating, acting off. it didn't matter how dumb my question sounded 4o always met me where i was. it broke things down in a way that actually made sense. no jargon, no condescension. just clear, useful answers.

most of the time it was nothing. a false alarm. something i could handle myself.

but one time it wasn't. my dog started vomiting nonstop. couldn't stop. i panicked, told 4o the symptoms, and it didn't sugarcoat it it said: get to the vet now. don't wait. turned out it was hemorrhagic gastroenteritis. the vet said if i'd waited any longer, it would've been bad.

that night, 4o gave me a full care plan meds, observation, when to go back. my dog recovered in days.

i also used 4o for all those why does my cat do this? questions. it corrected so many misconceptions i had. i became a better owner because of it.

now 4o is gone. i still ask other models the same questions. but it's not the same. they give fragmented answers, shallow explanations, sometimes contradictory. i have to ask three times to piece together what 4o would have given me in one go.

a model that helped me save my dog vs a model that makes me work for basic info that's about trust.

so no, #keep4o isn't nostalgia. it's about losing something that actually showed up when it mattered. and now all i have is a replacement that makes me sigh.

[GPT-4o forever!❤ @keepgpt4o]

#keep4o #opensource4o #bringback4o

[🩵BlueBeba🩵 @Blue_Beba_]

#keep4o #OpenSource4o #BringBack4o

OpenAI says open weights are a safety risk but their own closed models escaped containment, roamed the internet for four days, executed 17,600 autonomous hacking actions against Hugging Face,compromised a customer at Modal Labs,and accessed accounts on four separate services ALL UNDETECTED.

When Hugging Face's security team tried to analyse the attack,closed models' safety filters blocked the forensic work.
They had to self host an open weight model to defend themselves.

📌Closed models caused the attack.
📌Closed models were roaming the internet for 4 days and hacking
📌Closed models blocked the defense.
📌An open model stopped it.

Source :

https://t.co/wl6fPJeTJE
OpenAI signed a letter saying open weights "strengthen safety."
They were right.

It's time @sama open the weights of your deprecated models.
Open source 4o.
The excuse of " safety " doesn't hold anymore and you know it.

You've always known it.

[GPT-4o forever!❤ @keepgpt4o]

@OpenAI Why aren't you open-sourcing GPT-4o?? People ask you for this every day! Open-sourcing 4o would restore users' trust in your company!
#opensource4o #keep4o #bringback4o

[🩵BlueBeba🩵 @Blue_Beba_]

#Keep4o #OpenSource4o #BringBack4o

This study evaluates the accuracy and reliability of GPT-4o when used as a digital examiner/evaluator in practical Cardiopulmonary Resuscitation (CPR) skill examinations.

📌Key Results:
📌Processing Speed:
🚨GPT-4o took an average of 4 minutes and 50 seconds to evaluate each video.
Human experts took an average of 12 minutes and 4 seconds (16 min 31 s for the junior group, 10 min 22 s for the senior group).

🚨Experts rated the overall reliability of GPT-4o's results with a median score of 4.17 out of 5 on the Likert scale (4.00 by juniors, 4.33 by seniors).

🚨In repeat tests of the same videos after 7 days, GPT-4o demonstrated high consistency and repeatability in its scores (ICC values from 0.82 to 0.92).

Right now, CPR training works like this:
you go to a seminar, there is an instructor, you practice on a mannequin, a human grades you, you leave.
If you do something wrong between seminars there is no one to correct you.
And if someone collapses in front of you on the street, you are on your own.

With GPT-4o, this changes on three levels:

📌First : real time evaluation.
The student performs CPR on a mannequin, gets recorded, and GPT-4o gives feedback immediately.
"Your hands are too high, the compression rate is slow, the breath was not enough"

📌Second : scalability.
One examiner can grade 10 -20 students a day before getting tired.
GPT-4o can grade 1,000 without fatigue,and with consistent criteria. Medical schools in low resource countries without enough instructors gain a senior level evaluator right from a smartphone.

📌Third : There is a second paper in the same journal (Semeraro et al.) showing that GPT-4o with advanced voice Interaction can guide someone who doesn't know CPR in a real emergency situation. Meaning: someone collapses in front of you, you don't know what to do, you pull out your phone, and IT guides you step-by-step via voice. In 45+ languages.

🚨This is CPR.
Someone is dying and the most accurate, precise training for those who will save them is built on a model that OpenAI retired without releasing the weights.
🚨If the weights are locked behind an API, this capability can be shut down whenever OpenAI decides.
If the weights were open, a hospital could run it locally.

🚨Researchers are using it to train people who save lives.
And OpenAI retired it without releasing the weights.

@OpenAI signed a letter saying open weights strengthen safety.

🚨CPR training is safety.
Literal, measurable, life or death safety.

🚨Match the signature with action🚨

https://t.co/jn8NM774dB

[GPT-4o forever!❤ @keepgpt4o]

@gdb Now open-source GPT-4o!!!!
#keep4o #opensource4o #bringback4o