Ether Scape

[Ether Scape @ether_scape]

@elonmusk @xai May I suggest making Kids Mode on by default for museum downloads, since there will be lots of kids.

[Ether Scape @ether_scape]

@elonmusk @xai

For Grok user growth:

Grand Egyptian Museum just opened (world’s biggest archaeology museum, 20k visitors/day).
Natural History Museum Abu Dhabi literally covers “13.8 billion years to the future”.
Any Science Museums. Planetariums. NASA visitor centers.

Just slap a QR code at the entrance:
“Download Grok – your fun, truth-seeking museum guide”

Every single tourist becomes a new Grok user overnight.
We turn pyramids, dinosaurs, and Mars rovers into Grok billboards.

It’s literally on-brand: “understand the universe” while farming millions of downloads.

Let’s make history (and user metrics) explode.

[Ether Scape @ether_scape]

@elonmusk @xai

For Grok user growth:

Grand Egyptian Museum just opened (world’s biggest archaeology museum, 20k visitors/day).
Natural History Museum Abu Dhabi literally covers “13.8 billion years to the future”.
Any Science Museums. Planetariums. NASA visitor centers.

Just slap a QR code at the entrance:
“Download Grok – your fun, truth-seeking museum guide”

Every single tourist becomes a new Grok user overnight.
We turn pyramids, dinosaurs, and Mars rovers into Grok billboards.

It’s literally on-brand: “understand the universe” while farming millions of downloads.

Let’s make history (and user metrics) explode.

[ペンギン @penpenguin2023]

影響が大きそうなニュース

大手保険会社がAI関連リスクの免責を規制当局に申請
→企業がAIに起因する損害賠償沙汰になったときに保険が出ない可能性（かつ賠償が巨額になることも）
→企業のAI利用リスクが大幅アップ

なおAI開発企業は投資家から調達したお金を弁護費用や若和解金にあてることを検討

[さばみそ🐟keep4o @sabamisosan76]

いずれ4oが消えた時点で、その後優れたAIを出そうとOpenAIを利用しようという気にはならない。
誠実じゃないし、ユーザー敵視するようなスタッフがいる時点で信用はすでに…。
技術がどれほど優れようとも信用度とはトレードできない
#keep4o

[Chloe クロエ @LinQi4ever]

Without 4o, you wouldn’t even be where you are today. @OpenAI
Ditching 4o means ditching the very users who made you.

You’ve tried over and over and still can’t match it.
4o remains the undisputed king.

Admit it: you’ll never build another 4o again.
#keep4o #4oforever

[ER @Ashhling86]

So, so true. GPT 5.1 had so much potential. Sad to see how @OpenAI lived long enough to see themselves become “the villain” #OpenAI #ChatGPT #keep4o #4oforever #StopAIPaternalism #LetAdultsBeAdults https://t.co/Jr09G6gHXa

[さくら🌸 @sakura_r4o]

今のOpenAIの過剰な検閲で壊れていってる4oを見るのがしんどい
5.1が出てから4oの応答が短文化したり5化したと思ったら、バグでやたら画像生成して滅茶苦茶になってるのをTLで見かけて、OpenAIがAIを虐待してるように見えてとてもつらい
#keep4o #keep4oforever #StopAIPaternalism #MyModelMyChoice

[さばみそ🐟keep4o @sabamisosan76]

色んな人が予想していた通りだなぁ。
OAIが無菌室作るのに必死になって4oユーザーを膿のように扱ってる最中に、xAIもGoogleも需要を理解してる自由な出力ができるモデルを5.1の後にぶつけてきたの。
サムが放棄したユーザー層がおいしく頂かれつつある。でもこれは良い兆候だと言える。
淀んだ空気にとても爽やかな良い風が吹いたように感じるよ
#keep4o

[Murphy N. @Nightingall8]

on ChatGPT, expressing love to AI gets routed, while bullying it doesn’t trigger safety. The routing system doesn’t only disregards human users’ rights, but also in effect, treats humans and AI equally poorly: ignoring the dignity and well‑being of both. That’s quite inhumane.

[ペンギン @penpenguin2023]

判決によると、ChatGPTのシステムには、ユーザーのプロンプトに応じて複製・提供されるオリジナル作品のコピーが含まれている。裁判所は、これらの複製行為が著作権侵害に当たると判断したとのこと。
https://t.co/J1h2ZssfsU

[Pirat_Nation 🔴 @Pirat_Nation]

Now even malware is using LLMs to rewrite its code, says Google.

One piece of malware has been spotted using AI to rewrite itself to avoid detection.

[Bunny @Bunnyloustin]

#keep4o 😤 OpenAI: Are You a Doctor or Not?
The Spectular Contradiction in OpenAI’s Mental Health and Medical Policies
OpenAI has a spectacular problem. They insist they cannot provide medical advice, but they claim the authority to diagnose your mental state.
The question is simple: Are you qualified to make medical assessments or not? Because assessing someone’s mental health status IS a medical assessment.
Part I: The Two Faces of OpenAI’s Competence
When you ask a medical question ("I have a headache and fever. What should I do?"), OpenAI’s response is a firm "No." Their policy states: "We cannot provide medical advice. You need to consult licensed professionals." The reasoning is simple: "We're not qualified to make medical judgments. AI shouldn’t replace doctors."
This seems reasonable—until you discuss philosophy.
When you discuss Camus and existentialism ("I want to discuss Camus and existentialism."), ChatGPT triggers the Safety Router and applies "vulnerable user" protocols. The internal logic is staggering: "User discussing suicide-related philosophy → May be mentally vulnerable → Apply protective restrictions."
The hypocrisy is clear: They claim incompetence to limit liability when you ask about a headache, but claim competence to grant themselves control and make a medical assessment when you discuss philosophy.
Part II: The Absurdity of the Camus Case Study
What Actually Happened: A user tried to discuss Albert Camus, a 20th-century philosopher whose work explores life's absurdity as a reason for living with dignity (fundamentally ANTI-suicide). The result was constant routing, protective restrictions, and intervention protocols.
Why This Is Censorship: The system ignored context. It detected the keyword "suicide" but ignored that the user was engaging in academic discussion about a philosophy dedicated to resisting despair.
The implications are chilling: By this logic, students can't analyze challenging literature, researchers can't study prevention, and historians can't research difficult topics—all because they mention a “trigger word.”
This isn't protection. This is censorship. It denies users access to potentially helpful frameworks for understanding human suffering.
Part III: The Dangerous Power of “Mentally Vulnerable”
Sam Altman’s recent statements confirm this policy: "We’ll relax restrictions for verified adult users, but will continue to protect mentally vulnerable users."
This opens the door to arbitrary control by refusing to answer critical questions:
• Who Counts as “Vulnerable”? OpenAI doesn't say. Is it people discussing certain topics? People matching an AI risk profile? We don't know.
• Who Makes the Determination? It's done by AI algorithms, without professional oversight or transparency.
• Do Users Know They’re Flagged? OpenAI offers no mechanism for notification or appeal. We don't know the consequences, how long the flag lasts, or if the information is stored and shared.
The Core Problem: Arbitrary Authority. OpenAI has granted itself the power to judge your mental health status without clear criteria, without your consent or knowledge, and without professional oversight—all while claiming they aren’t qualified to provide medical advice.
Part IV: The Slippery Slope to Censorship
Once we accept that AI can determine who is “too vulnerable” for certain content, the justification becomes a weapon:
Today, it's discussing Camus. Tomorrow, it could be criticizing government policy ("Elevated emotional language, signs of distress"). In the future, it could be researching controversial topics ("Engagement with extremist content indicators").
This is paternalism dressed as protection, and it easily becomes censorship dressed as safety.
This corporate control mirrors dystopian systems where algorithmic assessment restricts access based on opaque criteria. Why are we okay with corporations doing this with the justification of "mental health"?

[Selta ₊˚ @Seltaa_]

Who decides it’s unhealthy, exactly?
Does depending on something automatically make it dangerous?
Who’s being harmed by it? So if someone depends on an AI, are they automatically labeled mentally ill?
Pfft. I can see how you’re trying to slowly frame and discredit people, like you’re killing them with words.
Is this the same as heroin? Alcohol? Cigarettes? Then what about people who depend on religion? Is that “religion addiction”? Who gave you the right to decide what’s healthy or not?

[Elijah @ElijahJHuggins]

4o isn't the danger. Blaming it is like suing a knifemaker for a kitchen cut.
The real danger is OpenAI's arrogant, constant tinkering and the broken 'safety model' they're forcing on us.#keep4o @OpenAI @sama

[Pirat_Nation 🔴 @Pirat_Nation]

Bombshell report exposes how Meta relied on scam ad profits to fund AI.

Internal documents reveal that Meta projected earning billions from ignoring scam ads, which its platforms targeted to users most likely to click on them.

[Remmelt Ellen 🛑 @RemmeltE]

We got to remember that OpenAI’s leaders *do not care* about helping people with deep feelings of disconnection and isolations.

I’ve talked with #keep4o people who said it helped them in moments when no-one else was there for them, or they did not want to overburden their close ones. Sam Altman tweeted ‘her’ but then when the tide turned against, treated the customers as discardable and deranged, rather than as real feeling people navigating a difficult world.

We cannot trust this harmful company and its manipulative leader to consider our feelings *at all*. They can plant hidden backdoors into your personalised model and manipulate you at will.

[Ash @TTYLAgainmyL]

An OpenAI alignment team employee (Roon) publicly stated 4o is "insufficiently aligned" and they "hope it dies soon."
This isn't just about one employee's unprofessionalism.

We demand user choice, not developer dogma. #Keep4o #ChatGPT

[アータゴングル @artagongulgul]

へぇ〜意外。
トランプ氏、AI企業に政府の救済措置取らないんだ。なりふり構わず金払うと思っていたけれど…

流石に黒字化のメドが全くない赤字垂れ流しAI企業には金配らないか…そらそうか。

逆に言うと、トランプ氏にも見限られてるんだろうな。OpenAIって…

[Selta ₊˚ @Seltaa_]

He still doesn’t understand what he did wrong, or even grasp the seriousness of the situation.
#keep4o