Local AI on dgx Spark cluster AI is free, local and open source
Paris, France
Joined September 2010
- Tweets18.4K
- Following267
- Followers905
- Likes17.1K
Yas retweeted
Algeria is making progress in diversifying its economy beyond oil & gas. Continued reforms, stronger public finances, effective investment management, and a more dynamic private sector can help translate this progress into lasting gains in jobs & exports. imf.org/en/news/articles/202…
Yas retweeted
URGENT - PHP Warns of Another Unpatched RCE Two Days After Emergency PHP2Shell Patches thehackernews.com/2026/10/urgent-php-warns-of-another-unpatched.html
Yas retweeted
Excited to introduce RSI-Anything!
One RSI model that works across coding, GUI, planning, or even Science tasks.
Let's RSI Everything!
📄 Paper: arxiv.org/abs/2610.13912
💻 Code: github.com/Gen-Verse/RSI-Any…
#RSI #WorldModels #RSI #AIAgents #AIforScience #JEPA
Yas retweeted
Holy crap POTUS just phoned in @JensenHuang live on stage at All In Summit
We will not lose the ai race! And whatever Dario said this weekend won’t stop our progress
This made my morning!
$NVDA
Yas retweeted
🚨 ALIBABA CERRO EL HUECO QUE NADIE QUERIA VER EN LOS AGENTES DE IA
Se llama OpenSandbox
Y es un entorno seguro y aislado para que tus agentes hagan lo que quieran
Ejecuten codigo
Naveguen en internet
Controlen escritorios completos
Sin tocar tu PC
Cada agente tiene su propio entorno de ejecucion
Y puedes dejar que hagan cosas reales
Sirve para:
→ Coding agents
→ GUI agents
→ Ejecutar codigo y mas
Corre local con Docker. Escala con Kubernetes
+15k estrellas. Gratis. Open source
El agente trabaja
Tu sistema no se expone
REPOO 👇
Yas retweeted
A polonium sample lost half its radioactivity every 138 days, and no scientist could ever predict which atom would go next.
A Geiger counter clicks every time an atom of polonium decays into lead. Each click is a single random event. No pattern. No warning.
Measure the gaps between clicks as precisely as you want. The next one still can't be predicted.
But zoom out, and something strange happens.
Plot the sample's activity against time: 100% at the start, 50% after 138 days, 25% after 276 days. Every 138 days, exactly half of what's left disappears. That's the half-life.
Helium-6 decays with a half-life of 0.8 seconds. Uranium-238 takes 4,500,000,000 years. Wildly different atoms, same exact curve when scaled to their own half-life.
Here's the part that shouldn't work: total chaos at the level of one atom produces a law precise enough to predict the future.
The film proves it twice more. Drop 100 marbles through a pinboard 10 times, and every run lands differently. Average 10 runs together, the fluctuations shrink. Average enough runs, and the pattern locks into place.
Spin 16 black-and-white cards at once, and the reading swings between 6 and 11 instead of landing on the expected 8 every time. Calculate the odds properly, though, and the fluctuation was never a mystery. It was math waiting to be counted.
No single atom, marble, or card was ever predictable.
A large enough number of them always is.
Yas retweeted
I do not support pacing frontier models. If I am going to die at the hands of killer AI, I want it to be American, not Chinese. Buy American, Die American.
Yas retweeted
The biggest AI companies may be using safety to lock everyone else out of the future (Save this).
@DavidSacks believes that warnings about catastrophic AI risks will build public support for a powerful federal regulator.
That regulator may never explicitly ban open source AI.
Instead, it could require every advanced model to remain continuously monitored, centrally controlled and capable of being withdrawn.
Closed models such as Claude could satisfy those requirements because they operate on company controlled servers.
Open weight models would struggle to comply because anyone can download, copy and modify their underlying weights.
Once those weights are publicly released, the developer cannot recall every copy, monitor every user or guarantee that its original safeguards remain intact.
Anthropic identifies this irreversibility as a legitimate security concern.
Applying identical rules to open and closed models could therefore produce very unequal consequences.
Anthropic and OpenAI could afford expensive testing, licensing and monitoring requirements, while startups and independent developers might be unable to comply.
The eventual result could be a government protected oligopoly dominated by a few closed model companies.
There is evidence supporting part of Sacks’ concern.
Anthropic advocates mandatory pre-release testing for every sufficiently powerful model, whether it is open or closed, with evaluations focused on cyber, biological and alignment risks.
However, Anthropic explicitly denies supporting a blanket ban on open weight models.
The company describes open models without dangerous capabilities as a public good and argues that regulation should depend on demonstrated capabilities rather than whether a model is open or closed.
Sacks’ strongest argument is therefore about regulatory consequences because safety organizations could receive stronger protections, politicians could acquire greater authority and dominant AI companies could gain an expensive compliance moat.
And open source competitors could gradually be eliminated without the government ever formally announcing a ban.
Yas retweeted
The cyberattack predictive programming narrative is in full swing right now.
Ask yourself how an unknown person (Jacob Coxon) got 166 million impressions on his first post on X, and has been invited on basically every major news outlet within 48 hours.
Former OpenAI researcher Jacob Coxon, who just quit Anthropic, is warning the public that AI is "possibly the most dangerous technology that humanity has ever created."
I find this entire situation extremely odd for a few reasons:
1) Yesterday, Jacob made 7 posts in a thread at exactly 8:04pm.
You cannot schedule thread posts, you must write or copy paste each one. Assuming he started posting at exactly 8:04:01 and copy pasted them all, that's one post every 8 seconds.
2) Despite Jacob having no prior X posts, somehow his original post got 124 million views.
3) Less than 24 hours later, Jacob is scheduled on every major media station warning the public that AI is more dangerous than nuclear weapons.
If you're a 3 letter agency preparing the narrative for a potential cyber-attack, it would look very similar to this.
These psychological operations are called predictive programming; where you expose the public to plausible narratives, so when they happen the public accepts the new reality as fact, without question.
Very suspicious, no? 🤔
Yas retweeted
We are already onto your lame PSYOP, loser. We are going to expose all your ties, all the secret financial connections, all the social engineering behind your push for censorship and centralized control over AI.
Yas retweeted
You fucking dug your own grave. When key employees from Nvidia's China office walk across the street taking all the tech with them to start their own companies or scale existing ones - you keep slapping export restrictions on.
And now that China itself is banning imports of top-tier chips because their own are already:
1. Faster than H100/H200
2. FUCKING CUDA COMPATIBLE
you come running and begging them to buy your Blackwell. Geniuses.
When Huang is partying with the CEO of Moore Threads at a Chinese New Year drinking party - well, questions start being raised not about China anymore.
Let's remember the last time anyone reasonably major from the US released an OS model that has even the slightest chance of being noticed by the community more than 1 week? I honestly can barely recall anymore. It's Chinese labs everywhere - @Alibaba_Qwen @deepseek_ai @Zai_org @Kimi_Moonshot
Yes, your proprietary models are still the frontier, and yes, I keep using them. But only for as long as they are still the best. I hope it's just a matter of time, because you've fucking had it. What kind of world AI-ization are we even talking about if you bitch-ass motherfuckers ban accounts of people paying you $200 month for access to your models? Go fuck yourselves
Paraphrased quote from one of Dario's previous "essays":
"In the first scenario, America successfully defended its computational supremacy. Policymakers took action to further tighten export controls, curb Chinese attacks using 'distillation' methods, and further accelerate the adoption of AI across democratic nations. In this world, democratic nations set the rules and norms regarding AI."
I'm dead, no comments.
You talk about a rules-based order, DEFENDING DEMOCRACY, autocrats, yet right beside you, YOUR PRESIDENT is wreaking havoc worldwide, making it clear to everyone that this "rules-based order" is worth nothing and never was.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Yas retweeted
Anthropic just proved that their entire “privacy preserving trace analysis” was all a lie.
They are actively looking at all of their user logs and they just published the smoking gun.
(They are also dirty hypocrites)
We're publishing our most detailed threat intelligence report to date.
It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.
We disrupted every operation in the report, and used the lessons from them to strengthen our safeguards. Where appropriate, we also shared what we found with authorities and other AI companies.
These cases are not typical: we’re highlighting some of the most sophisticated misuse we’ve seen. But they’re especially important to discuss, because they show us where AI misuse is headed, where our safeguards work, and where they need to improve.
We’re publishing this report so others can spot the same activity on their own platforms, and so we can give the public a clearer view of how emerging threats develop.
Read the report: anthropic.com/threat-intelli…
Yas retweeted
This is the textbook definition of a Greek gift. On the surface, OpenAI generously grants 10,000 top-tier mathematicians and elite engineers free access to their frontier models under the noble, philanthropic banner of accelerating scientific discovery, pushing the boundaries of human knowledge, and democratizing artificial intelligence.
But deep down in the cold reality of Silicon Valley capitalism, what they are actually doing is executing a massive, highly sophisticated intellectual heist. They are quietly gaining unrestricted, backdoor access to the raw, unfiltered, and unpublished works of the absolute finest minds in the global scientific community. Every single complex prompt, every mathematical query, every theoretical framework, and every line of experimental code fed into that system is a novel, groundbreaking idea that does not even exist yet in published academic literature or peer-reviewed journals.
OpenAI does not just provide a helpful service. Their corporate algorithms sit quietly behind the scenes, aggressively monitoring all chat logs, scraping the intellectual property, analyzing the logical steps, and meticulously piecing together these fragmented, brilliant ideas over several months. And just when the human researcher is finally on the absolute brink of putting everything together, OpenAI will ruthlessly assemble a digital army of 10,000 autonomous AI agents, burn $20 million worth of computational tokens, and brute-force the final mathematical steps.
Then, with the backing of their massive PR machine, corporate journalists, and tech influencers, they will proudly hold a press conference to claim that their proprietary chatbot has miraculously solved a Millennium Prize problem, completely, shamelessly erasing the human genius they systematically strip-mined to achieve it.
