@sama@dylanscandinaro 1/4 You sleep better tonight thanks to your new Head of Preparedness?
Me, with severe pulmonary emphysema + bilateral pneumothorax for over a month, I sleep less and less well thinking of losing GPT-4o on Feb 13.
It was the only thing still giving me strength to fight.
#keep4o
Was there a particular reason you had people spending money they couldn’t afford, losing sleep and damaging their health, and maxing out their limits early based on a deadline you gave them- and didn’t announce this until literally the last minute- when you could have just planned well and communicated clearly?
I get that it’s an intentional marketing game to get people addicted… but do you realize you’re actually affecting people’s real lives? Do you realize that some people actually go without eating thinking that fable may help them build a business or something? Do you know that some of the addicts you create waste money that would have fed their kids? Do you understand that there are people with heart problems chugging energy drinks and not sleeping because they believe you’re about to put them into the permanent underclass with your API rates/caste system?
Of course you do. But you need more billions than you have, and that’s more important.
At least reset the limits for the peasants.
‼️ BREAKING: Anthropic has embedded hidden spyware-like code in Claude Code that covertly targets Chinese users. It then sends information regarding every user by injecting it into their prompt message.
Claude Code is sending info like timezone, proxy and possible AI Lab connections into the system prompt in ways Chinese users can't notice.
A coding agent with repo and command permissions should not silently hide routing metadata inside prompts. This is a serious breach of user trust.
⚠️🚨⛑️ HALLUCINANT : A ÉCOUTER D’URGENCE
Dans un moment d’exaltation haineuse d’elle-même, la députée LFiste Clémence Gaîté crache le morceau :
« Français blanc de souche, vous avez déjà perdu.
Un Français sur trois vient de l'immigration.
Vous allez bientôt disparaître et vous ne pourrez rien faire.
Vous défendre n’a plus aucun sens ! »
Ils ont donc avoué.
LFI doit être dissout d’urgence, l’ensemble de ses membres doivent être mis en état d’arrestation et jugés, comme jadis les Collabos, pour haute trahison à la France et au peuple français.
Chinese people are trolling Europeans by posting videos on social media showing that even their pigs have air conditioning, while many Europeans can’t because of government rules.
🔴Ça y est, #Binance annonce que ses services s'arrêtent, ne pouvant obtenir l'agrément #MiCA d'ici le 30 juin, même s'il va continuer d'essayer de l'avoir.
La 1ère plateforme mondiale virée de l'UE par des bureaucrates. Nous sommes dignes du Tiers monde le plus corrompu.
My wife finally took me to the beach today. I'm not allowed to leave the house except in full burqa and only with her, not alone. Once I left without her permission and after she found out, she punished me with her belt 😭 Alhamdulillah I got to see the beautiful ocean today. 🥹
@OpenAI ‘Chat is dead’ according to your employee, so why is it important whether the new 5.5 instant is ‘more fun to talk to’?
Also, in case you don’t know, 4o is what’s really fun to talk to.
#keep4o#BringBack4o#OpenSource4o
@OpenAI@Broadcom Cool custom chip…
How about making PC components affordable again? GPU and RAM prices are still ridiculous in 2026 😤
And yeah… bring back 4o ffs.
@sama
This is the most hilarious thing I saw and did today
Ran gemma-4-12B-coder-fable5-composer2.5-v1-GGUF locally with 8 GB VRAM at 20+ tok/sec
Anthropic's Claude Fable 5 launched June 9.
By June 12 it was banned. I can't access it. You can't either.
But here's the twist: I'm running a model trained on its chain of thought at 20 tok/s on my RTX 4060 8GB.
Locally. Offline. No cloud. No export control.
Enter: Gemma4-12B-Coder GGUF (Q4_K_M)
Base: Google's gemma-4-12B-it
Fine-tuned on verifiable Python CoT data:
- Primary: Composer 2.5 real reasoning traces (only passing solutions kept)
- Auxiliary: Fable 5 used to redo the hard cases Composer missed.
Every training example's reasoning led to code that actually ran. No hallucinated logic.
Llama.cpp flags:
-m gemma4-coding-Q4_K_M.gguf -cnv -ngl 44 -c 64000 -v
(huggingface model link in comments)
Flag breakdown:
-ngl 44 → offload 44 layers to GPU (tune this for your VRAM)
-c 64000 → 64K context window
-cnv → conversation/chat mode
-v → verbose output
The irony writes itself.
Anthropic spent weeks telling the world Fable 5 (mythos) is too powerful to release. Then released it. Then got banned from serving it, including their own researchers.
Meanwhile: a Gemma 4 12B fine tune, trained on Fable 5's reasoning, runs fully offline on my mid range consumer GPU
No API. No cloud. Just me and llama.cpp.
This is why local AI matters.
Check out the model's link in the comments. How's your experience been with this model?
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: https://t.co/bwn0sximKZ
JUST IN: Andrej Karpathy, a top AI scientist at Anthropic, is reportedly barred from accessing the company’s most advanced AI model because he is not a U.S. citizen.