
Building @zaq__ai: Open source sovereign AI operating system Mentor @IncubTelecom
Paris, France
Joined May 2008
- Tweets2.4K
- Following535
- Followers544
- Likes1.4K
JAi-cool 🇫🇷🇱🇧 retweeted
Just so we are all on the same page:
* The Rust version has hundreds of commits, with multiple optimizations rounds, over several days. The others have 1 or 2 commits from an initial agentic rewrite
* Based on the available documentation, different things were asked from each attempt. The agents.md from Go mentions the Rust version was the canonical example. The Elixir one mentions it was the Rails one
* While we don’t have the full prompts, it seems they were vague, allowing the agents to make non-language architectural decisions that impact results (such as how closely it should replicate Rails’ drawbacks)
It is actually amazing how quickly we can have full-blown rewrites up and running. But if we want to take conclusions from the data presented, it is important to understand the process behind it.
The agents finished porting Campfire to Laravel and Django too. I'm sure there are optimizations that could be found there too, but the whole point is to get a sense of how frontier agents do with different environments out the box. github.com/basecamp/once-cam…
JAi-cool 🇫🇷🇱🇧 retweeted
I am old enough to remember when we were told that Sharon’s extremism was the biggest threat, and that replacing him with a less “reckless” leader would change everything.
Whenever there is a need to rebrand Israel, the same script is used.
PMs change, underlying policies remain.
Replying to @HillaryClinton
- Netanyahu's extremism is the biggest threat to U.S. support for Israel. Even those who support Israel can't defend his government’s recklessness.
It's not too late to change course, but time is running out.
You can read the full speech here: rahmemanuel.substack.com/p/t…
JAi-cool 🇫🇷🇱🇧 retweeted
« Ça m’écœure,.. les Israéliens ont inventé un concept assez original: le cessez-le-feu unilatéral. C’est-à-dire le cessez-le-feu, ça marche pour les mecs d’en face, pas pour eux(...) Qui agresse qui ? »
Le général Michel Yakovleff lâche une intervention sans langue de bois, LCI
JAi-cool 🇫🇷🇱🇧 retweeted
"Mon agent de ciné m'a lâché. Des émissions de télé m'ont m'annulé, et des rédactions ont demandé à leurs journalistes de ne pas écrire à mon sujet".
@AkimOmiri raconte comment ses prises de position en faveur de la Palestine l’ont mis à l’écart du milieu artistique mainstream.
Sans aucun regret puisqu'il a réussi à percer avec son spectacle et son émission sur Radio Nova sans avoir à sacrifier ses convictions 🔥
(@ZawaProd)
JAi-cool 🇫🇷🇱🇧 retweeted
Carbyne est une boîte israélienne qui a pris la main sur les systèmes 911 à New York, en France, au Mexique, en Israël, avec 30 pays visés à terme. Son produit : quand tu appelles les secours, Carbyne peut allumer ta caméra à distance, tracker ton GPS et faire passer ton flux vidéo en direct à de l'analyse IA prédictive. Le tout bâti par d'anciens de l'unité 8200 du renseignement militaire israélien.
Les emails Epstein a montré qu'il avait injecté 1,5 million via sa société écran Southern Trust, il a bankrollé Ehud Barak pour y placer son nom, et il a personnellement consulté Peter Thiel avant de lui arranger un ticket. Reid Hoffman, lui, avait introduit Thiel à Epstein dès 2014. Leon Black apparaît aussi dans la cap table.
Bref, toujours les mêmes investisseurs, ingénieurs militaires, cabinets juridiques...
Le 911 n'est pas un sujet parmi d'autres. C'est l'excuse parfaite pour l'État pour entrer dans ton téléphone, et Carbyne a monté le tuyau.
Ce qui devrait nous faire grincer des dents, ce n'est pas l'IA générative. C'est le mélange entre capital privé opaque, ingénierie militaire étrangère et infrastructure publique sensible, le tout sans qu'un seul élu n'ait jamais rien voté. Les "risques existentiels" de l'AI safety occupent les dîners à San Francisco pendant que la vraie concentration de pouvoir se déploie dans les centres d'appel d'urgence.
Je trouve toujours les régulateurs européens très énergiques quand il s'agit de surveiller un chatbot. Beaucoup moins quand il s'agit de regarder qui contrôle nos appels d'urgence.
JAi-cool 🇫🇷🇱🇧 retweeted
DIRECT : LE SITE DE THALES A GENNEVILLIERS BLOQUÉ
La coalition « guerre à la guerre » bloque le site de l’entreprise et dénonce la complicité de Thales dans le génocide en Palestine : « Israël assassin, Thales complice » scande les militants depuis 8h du matin
📹 @LisaLap_
JAi-cool 🇫🇷🇱🇧 retweeted
Time to build in public
My first commit of Jido Managed Agents
Normally wouldn't make software at this stage public, but trying it differently this time
This is the ROUGH cut - the output of the Ralph Loop and I will refine it over the weekend
github.com/agentjido/jido_ma…
JAi-cool 🇫🇷🇱🇧 retweeted
Terrifying reality check. An Israeli soldier openly admits on camera that they are helping illegal settlers steal Palestinian land and justifies settler violence as "revenge." CNN exposes how the IDF acts as a private militia for extremist settlers in the West Bank.
JAi-cool 🇫🇷🇱🇧 retweeted
Replying to @ClementDelangue @atreides_sf
Let's be real, all closed models profit from open models WITHOUT GIVING BACK.
JAi-cool 🇫🇷🇱🇧 retweeted
Cursor is raising at a $50 billion valuation on the claim that its “in-house models generate more code than almost any other LLMs in the world.” Less than 24 hours after launching Composer 2, a developer found the model ID in the API response: kimi-k2p5-rl-0317-s515-fast.
That’s Moonshot AI’s Kimi K2.5 with reinforcement learning appended. A developer named Fynn was testing Cursor’s OpenAI-compatible base URL when the identifier leaked through the response headers. Moonshot’s head of pretraining, Yulun Du, confirmed on X that the tokenizer is identical to Kimi’s and questioned Cursor’s license compliance. Two other Moonshot employees posted confirmations. All three posts have since been deleted.
This is the second time. When Cursor launched Composer 1 in October 2025, users across multiple countries reported the model spontaneously switching its inner monologue to Chinese mid-session. Kenneth Auchenberg, a partner at Alley Corp, posted a screenshot calling it a smoking gun. KR-Asia and 36Kr confirmed both Cursor and Windsurf were running fine-tuned Chinese open-weight models underneath. Cursor never disclosed what Composer 1 was built on. They shipped Composer 1.5 in February and moved on.
The pattern: take a Chinese open-weight model, run RL on coding tasks, ship it as a proprietary breakthrough, publish a cost-performance chart comparing yourself against Opus 4.6 and GPT-5.4 without disclosing that your base model was free, then raise another round.
That chart from the Composer 2 announcement deserves its own paragraph. Cursor plotted Composer 2 against frontier models on a price-vs-quality axis to argue they’d hit a superior tradeoff. What the chart doesn’t show is that Anthropic and OpenAI trained their models from scratch. Cursor took an open-weight model that Moonshot spent hundreds of millions developing, ran RL on top, and presented the output as evidence of in-house research. That’s margin arbitrage on someone else’s R&D dressed up as a benchmark slide.
The license makes this more than an attribution oversight. Kimi K2.5 ships under a Modified MIT License with one clause designed for exactly this scenario: if your product exceeds $20 million in monthly revenue, you must prominently display “Kimi K2.5” on the user interface. Cursor’s ARR crossed $2 billion in February. That’s roughly $167 million per month, 8x the threshold. The clause covers derivative works explicitly.
Cursor is valued at $29.3 billion and raising at $50 billion. Moonshot’s last reported valuation was $4.3 billion. The company worth 12x more took the smaller company’s model and shipped it as proprietary technology to justify a valuation built on the frontier lab narrative.
Three Composer releases in five months. Composer 1 caught speaking Chinese. Composer 2 caught with a Kimi model ID in the API. A P0 incident this year. And a benchmark chart that compares an RL fine-tune against models requiring billions in training compute without disclosing the base was free.
The question for investors in the $50 billion round: what exactly are you buying? A VS Code fork with strong distribution, or a frontier research lab? The model ID in the API answers that.
If Moonshot doesn’t enforce this license against a company generating $2 billion annually from a derivative of their model, the attribution clause becomes decoration for every future open-weight release. Every AI lab watching this is running the same math: why open-source your model if companies with better distribution can strip attribution, call it proprietary, and raise at 12x your valuation?
kimi-k2p5-rl-0317-s515-fast is the most expensive model ID leak in the history of AI licensing.
JAi-cool 🇫🇷🇱🇧 retweeted
🚨BREAKING: UAE’s🇦🇪 Al Habtoor Group Founding Chairman Khalifa Ahmad Al Habtoor’s open letter addressing Lindsey Graham’s call for Gulf States to join the war on Iran:
I heard the statements of U.S. Senator Lindsey Graham, in which he calls on the Gulf Cooperation Council countries to enter this war, saying that we are also under attack and that we must join the fight. And I say to him clearly: We know full well why we are under attack, and we also know who dragged the entire region into this dangerous escalation without consulting those he calls his "allies" in the region.
We thank God that the United Arab Emirates and the Gulf Cooperation Council countries are doing well, and we have no need for someone who claims he came to the Middle East to save us. The truth is that hasty American decisions are what embroiled the region in a war whose peoples were not party to its decision-making, and its local allies were not consulted before it was launched.
We do not deny the Iranian threat to the region, which has become clear in recent days. We do not trust Iran. But this is a dirty game in which several powers clash at the expense of our region, with no honor in it and no transparency. In this scene, Iran, Israel, and the United States all move according to their own interests, not according to the interests of the peoples of the Middle Eastern Arab countries.
And for this reason, we say it clearly: We will not enter this war to serve the interests of others, nor will we sacrifice our sons in a conflict that could have been avoided through diplomacy and political solutions. We value the lives of our sons, and we do not treat their souls as "collateral damage" as some do. Nothing in this world is more precious than the lives of our sons, and no alliance with any country in the world is worth risking these lives. If President Donald Trump and Senator Graham are prepared to risk their country and the lives of Americans for Israel's interests, that is their choice. As for us, we will not do the same.
Senator Graham says they are "the Arabs' allies" and that we need and benefit from U.S. protection. And I say to him: We do not need your protection. All we want from you is to keep your hands off us.
He also says: "We sell you weapons." As if that were a service the United States provides us. The truth is that this is no favor from anyone. We invest in our security and pay billions of dollars for these weapons, and it is a massive trade and industry built on these deals. In fact, the United States itself is now turning to buy weapons from Ukraine to supply its allies in other wars. This is an industry that thrives on wars and arms sales, not a charitable endeavor.
And the clearest thing of all is what Senator Graham himself said when he spoke about oil. He stated that Iran and Venezuela together hold 31% of global oil reserves, and that the United States could establish a partnership with this share of the world's oil, and that it would be a "nightmare for China"—he even added that if the regime in Iran changes, there would be a "new Middle East," and the United States would "make a lot of money."
Only then does the picture become clear. And only then do we understand why they want this war.
Finally, I say to Senator Graham: You may be a senator in the U.S. Senate, but anyone who hears your statements might think you're a member of the Israeli Knesset, because you defend Israel's interests more than you defend the interests of the American people themselves.
The peoples of our region have the right to live in security and stability, far from others' wars. We also have the right to decide how to protect our interests and security, and to emerge from crises we did not create but which were imposed on us as a result of conflicts between powers vying for influence in this region.
We want peace and stability. We will not accept being forced down the path of war, nor will we accept being fuel for others' battles.
🚨WAR ON IRAN: US🇺🇸 and Israel CANNOT impose surrender, defeat on Iran as a real solution:
‘Whether it’s the war in Gaza that we did not support, but we got burned by, or the twelve day war on Iran, and this current war that we did not opt for, but we will burn by as well. These are decisions by the US, by Israel, not our decisions.
However, I think our role right now is not to say that look, I am going to be part of this war or not going to be part of this war. I think our role right now is to look at the future and see how can we come out of this, and in the UAE🇦🇪 fashion, I think we also, in the midst of big calamities, we look for opportunities.
Any solution that could come out of this, if it doesn’t involve real benefit to Iran and Iranian people, it will not succeed.
This cannot be a defeat, a surrender, a humiliation of Iranians, because there are 100 million people in Iran, those cannot be ignored. They have their own agency, they have their own aspirations. What we need is to have our aspirations and their aspirations aligned.’
-@MABaharoon, Director General of the Dubai Public Policy Research Center, on the latest episode of Going Underground, as the UAE🇦🇪 deals with the regional crisis unleashed by the war on Iran.
Watch the full interview in the quoted post, or watch it on Rumble, link below in the replies👇
JAi-cool 🇫🇷🇱🇧 retweeted
Here's a megathread on the war on Iran that focuses on the deeper dynamics, potential trajectories, and likely outcomes.
A lot will change over the next few weeks - this thread is about what won't.
Stay until the end for an announcement & an invitation.
🤖 Made with AI
JAi-cool 🇫🇷🇱🇧 retweeted
Replying to @theo
The only logical conclusion is that AGI will happen any time now.
JAi-cool 🇫🇷🇱🇧 retweeted
Elixir v1.20.0-rc.2 series is out! Overall, the compiler finds more bugs, for free, and it has never been faster:
* Infers types across clauses, finding more bugs and dead code
* Compiles ~10% faster and has a new interpreted mode (up to 5x faster, scales to the number of cores)
* Modifying a struct definition recompiles fewer files (better incremental compilation)
Please give it a try and share your feedback! Expect at least two additional RCs before v1.20.0!
JAi-cool 🇫🇷🇱🇧 retweeted
Jews don’t need a separate country.
American Jews belong in America.
British Jews belong in the UK.
Canadian Jews belong in Canada.
Turning Israel into “the country of the Jews” pushes the idea that Jews are a separate nationality.
It’s wrong, reckless, and dangerous but nationalist agendas have always been placed above Jewish safety.
- Our Community Memeber
JAi-cool 🇫🇷🇱🇧 retweeted
It's official. It's genocide.
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more


























