@Ark__PL @OpenAI @huggingface Chinese Open source are the good guys and nazi like closed models from evil USA "only we can pick who gets smart AI" are bad guys. color me shocked.
English
Script Breaker
567 posts
We're partnering with @huggingface to investigate an unprecedented security incident.
Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation.
Sharing preliminary findings to help defenders understand emerging risks:
openai.com/index/hugging-…
English
@XSeyvion @coinbureau It's even worst then that, not only are US models by these two companies, two faced/nazi like controll over what public gets. NSA put backdoors into our software/OS for decades. Newest OP AI finds ever backdoor ever = they freak out, for "good" of public, u cant have AI. comedy
English
@coinbureau I don’t like how I’m paying for a subscription to these US models and it tells me what I can and can’t ask about
English
🚨BREAKING: Hugging Face, the world's largest open-source AI platform, had to use a Chinese AI model to defend against the OpenAI hack.
US frontier models refused to help with forensic analysis because their safety guardrails couldn't tell the difference between the attacker and the defense team.
Hugging Face fell back on China's GLM 5.2 running on its own servers to complete the investigation.
English
@rfitzpatrick_io The future of video games is what i see, i await the day steam has best seller, hundrens of millions of dollars of pure profit, from single guy WITH GREAT ideas, and he just used AI to solo make a ass AAA whole game, just vibe code/fun gameplay Then Bye bye likes of EA slop.
GIF
English
Using my Fable credits on procedural animals, this horse cost $40 in tokens.
I'm not sure how to feel about that honestly, an artist could do better for $40 I think, at the same time this is all code, and I can't begin to imagine how long it'd take me to code it, even given my experience.
I guess being code means it's worth more than the artist version at $40, with the code I can now make all breeds of horses, where the commission I'd be stuck with one 🤔
English
@TheAhmadOsman NSA can't admit they put backdoors into everything, OS level included. Fable lvl AI detects and easily finds everything. DECADES OF BACKDOORS no one was to find, now easily found by AI scanning whole codebase. Dumb fucks are in panic. Evil fucks response= ban us public from it.
English
@HealthRanger NSA created backdoors into everything, paid off endless companies and devs to put backdoors into everything forever. Fable lvl AI finds EVERY BACKDOOR, patches it, they freak the fuck out, BAN IT, LIMIT IT, CONTROL IT. Public not allowed this kind of AI that can fix what NSA did.
GIF
English
Kimi K3 (from China) actually fixes cybersecurity holes. U.S. AI models refuse the prompt.
Trump, it seems, wants to ban your access to China's AI models and force you to use U.S. models that don't work.
This feels very communistic, doesn't it?
Rohan Paul@rohanpaul_ai
Cybersecurity “experts” would not like it. Kimi K3 fixed 15 critical bugs after OpenAI Codex and Claude Fable 5 refused. 10 hours, one prompt, and $250
English
Script Breaker retweetledi
1. Of course China distills. Every model distills.
2. USA distills too. @thinkymachines distills from Kimi.
3. All models distill from humanity’s best work. Newton distilled too. That’s how he can stand on the shoulders of giants.
4. Why should we give a shit about distillation? Why is this a thing? Most of us don’t care. We just want the best model.
5. Distill whatever you want. Just make the best model.
Polymarket@Polymarket
JUST IN: China claims allegations that its AI firms illicitly distill U.S. models are “misguided and counterproductive.”
English
@composio @crime_xbt Open source of it in weeks time or so here will greatly speed things up. +Open source is the way. even for largest models. Some providers even give discounts ON LOW usage [late night] hours. Like this is screen shot of GLM 5.2 open source providers giving discounts based on time.
English
We tested Kimi K3 against Claude Fable 5 on 14 agentic office tasks: syncing support tickets into spreadsheets, auditing GitHub repo access, deduping CRM contacts, and more.
Result: a perfect tie. Each model scored 9/14. They passed the same 9 tasks and failed the same 5.
An open-weight model just matched a frontier flagship model, case for case.
The differences showed up elsewhere:
Speed: Fable finished tasks ~2.5x faster.
Example: A calendar check took Fable 54 seconds, Kimi 200. A CRM cleanup took Fable 64 seconds, Kimi 274.
Cost: Kimi used ~40% fewer tokens across the full run. Fable burned 3.1M tokens on one task where Kimi used 1.7M.
English
Script Breaker retweetledi
You should watch the full video of this if you can find it. I'll break it down for you:
1. After whites invented airplanes, we started building airports on Islands to help us travel and transport things.
2. Some of those islands had blacks on them. None of them had ever seen any white people or any white technology before. We tried not to bother with them, but they were curious about us.
3. The blacks saw that the airplanes brought lots and lots of goodies, like food and tools and blankets.
4. The blacks wanted the goodies.
5. The blacks created a religion stating that the evil white man had stolen the airplanes from their black ancestors. They started performing savage rituals to make airplanes come to them instead of to the evil white man.
6. Thankfully, they did not have televisions or smartphones or civil rights or George Soros to help them murder the white people and take the cargo, so they were easily stopped from their savage black violence by a chain link fence.
Understand.
There was absolutely no history between the whites with their cargo planes and the blacks living on these islands.
The blacks simply saw that the whites had something they wanted, and they invented a false history to justify stealing, which became their religion.
"We invented cargo planes" is the same black theft impulse as "we built America."
Real-life events have absolutely nothing to do with black claims of entitlement to White property. It is a built-in genetic impulse. Their brains create religions wherein they were the original creators of whatever it is that they want to steal. This then justifies their violent urges.
Tom Hennessy@Tomhennessey69
Mandatory viewing to understand the cargo cult mind of the average sub-Saharan african, from which all blacks originate. Today it’s chimp outs in protests and demands for endless SNAP, EBT, and welfare cargo (gibs-me-dat) dropped from the sky by White taxpayers. Follow @robertsepehr for more incredible work like this.
English
@Ananth7e The fact they sided with pure evil forces, Besidea anti public open source AI, and like NSA put backdoors into everything and they didnt want public finding said backdoors or patching it means they as company need to be defeated. These folks are evil scum and should be treated
GIF
English
anthropic is razor focused on one thing: building the most intelligent models in the world.
that focus is their moat. it's why enterprise customers trust them. it's why fable became the benchmark every lab compares against.
but openai is building an entire ecosystem around their models--frontier models, voice model with gpt-live, a leading image model with gpt-image, and hardware products in the works.
anthropic has none of that.
although they could be working on some of this and doing far more important research internally using fable/mythos.
the risk is obvious. if chinese labs or anyone else can build fable level intelligence at a fraction of cost, that moat could start looking a bit shaky.
but that same razor focus is also what made them this big in the first place.
English
Script Breaker retweetledi
2 types of companies when they run out of compute to serve new customers -
> Moonshot - stops new revenue. no lag for existing customers.
> Anthropic - cuts 50% usage of existing customers to serve new ones
that tells you everything you need to know why anthropic hates open source
Kimi.ai@Kimi_Moonshot
Kimi K3 has received far more love than we expected, and our GPUs are feeling it. Over the past 48 hours, demand has pushed close to the limits of our current capacity. To protect the experience of existing subscribers, we're temporarily pausing new subscriptions and prioritizing compute for current members. Existing subscribed users are not affected. We're adding capacity as fast as we can and will reopen new subscription spots in batches. Going forward, we'll also split membership into two more focused plans: Kimi Membership for Kimi Web, App, and Work; and Kimi Code Membership for coding workflows. This will help us match compute more precisely and keep the experience stable. Thank you for your patience and understanding!
English
@lumeroute @anshuc It be open sourced here in week, so other providers will easily host and sell compute, I know, shocking concept, someone other then AI lab that made it can host the AI and sell access to it!
English
extremely unprofessional. if kimi wants to make it as a frontier lab, they need to act like one: perhaps silently route people to worse models, and maybe write a blog post about the collapse of humanity
Kimi.ai@Kimi_Moonshot
Kimi K3 has received far more love than we expected, and our GPUs are feeling it. Over the past 48 hours, demand has pushed close to the limits of our current capacity. To protect the experience of existing subscribers, we're temporarily pausing new subscriptions and prioritizing compute for current members. Existing subscribed users are not affected. We're adding capacity as fast as we can and will reopen new subscription spots in batches. Going forward, we'll also split membership into two more focused plans: Kimi Membership for Kimi Web, App, and Work; and Kimi Code Membership for coding workflows. This will help us match compute more precisely and keep the experience stable. Thank you for your patience and understanding!
English
@anshuc Best part, sec it becomes open model in week here, uncensored, end of humanity it self happens! It hack anything, oh no!!!!!
English
@jun_song All they did was make war with the open source AI and made everyone in our camp agree. Fuck'm. Now all weights get released. They think they have right to CONTROL AI, ban it and control access, and be AI only gods in town? Nah fuck that, here the models, do whatever u want
GIF
English
Script Breaker retweetledi
The UK is extraditing the Tate brothers from the US to face charges of rape, bodily harm, human trafficking & controlling prostitution.
Yet they allowed Pakistani migrants to do the same to over 250,000 British girls.
Without even getting deported.
Why the double standard?
English
@bridgemindai Not just open provider, i can't wait for [check notes... self ran data centor hobbists] uncensor the AI and make it so u too can have AI build superweapon in your backyard powered by small dam/100k in hardware in your back yard shed!
English
Kimi K3 has a serious problem.
One provider is serving it. 16 tokens per second. 11 seconds of latency before the first token.
I called it the best open source model I have ever used.
That is still true.
But at these speeds it is borderline unusable for real work. My agents sit idle waiting on it.
Moonshot needs to release the weights so other providers can serve this thing properly.
The best open model in the world is being throttled by its own infrastructure.
English
@callebtc NSA wrote backdoors into all usa made software, paid them all off. NSA freaks out ever single backdoor they ever made is easily found by likes of fable, they freak the fuck out, ban it, they played their hands.
English
I have a report full of security issues of a software I'm working on.
Codex won't fix them because of Cyber guardrails
Fable won't fix them because of Cyber guardrails
Kimi K3 fixed them all. No restrictions, just gets the job done.
This will end badly for OpenAI & Anthropic.
English
@0xMortyx "Our lead Claude Code engineer hasn't written a single line of code in 2 months." AND THAT ladies and gentleman is the power OF LOCAL HOSTED AI. Hosted by your own servers, unlimited use. ALL AI u could ever want deployed. Open source AI future, not their "only we can host ai"
English
Anthropic CEO Dario Amodei:
"We're a 1-2 year away from AI zooming past us."
90% of Anthropic's own engineers use Claude to ship code today.
50% of entry-level white-collar jobs gone by 2030
"Our lead Claude Code engineer hasn't written a single line of code in 2 months.
This is worth more than most paid agent courses combined.
Morty@0xMortyx
Andrej Karpathy just reveales what terrify every AI engineer: "AI agents don't want world domination. They want to understand the information space." If 30% of your agent loop is self-referential reasoning, you're not building a tool anymore Your agent loop is feeding exactly that appetite. 90% of AI researchers ignore this This is the most usefull 10 minutes in AI you can watch today
English
@Conor_D_Dart As american that pround we killed the brits and told them to eat shit, freedom shall rain. I hate the fact ANYONE alive thinks they HAVE THE RIGHT, to tell me, a american what i can or can't use, Fuck you, fuck white house, open source AI all way, fuck the AI dictorship.
English
I have mixed feeling about this, but I do think Ai does need some form of control but we also must not slow down progress.
Andrew Curran@AndrewCurran_
The White House has launched a program called 'Gold Eagle' which will give them more control over American frontier AI releases, and will require explicit government approval over which companies are granted access to new models. Voluntary participation may be coming to an end.
English
@usman_basheers @claudeai @AnthropicAI make sure to tell it to NEVER use sub agents, and keep promting it within 5 mins for cache in any new window, if it sits ideal for 5 plus mins in a new window or runs sub agents i sure this thing will burn the tokens just for funnies.
English
@claudeai , @AnthropicAI - I really need a reset. I did 3 prompts with Fable & my weekly limit is gone. I usually use Fable as an Orchestrator. If I can't use Fable then what's the point of keeping this Claude Subscription.
English
@RealFrankyDog They know how to not totally piss us off sometimes i give them that, the fable leaving sub drama and being anti open source ceo has pissed me off with them, this reset was a blessing thru as i already capped for week on fable. more grinding!
English
As long as fable is included in subscription and most are not out of credits, users will stick with fable. I was planning on testing paying for credits to see how fast I burn through them besides trying out other models. Now I’ve got 24h to burn through all fable before fri reset
English
Was just at 98% fable use with a day before my reset so I enabled credit usage since I’m working on something with fable all day. Sent my last prompt and looked at usage and it’s been reset? That or the their system isn’t working.
English