0xAzah 🇺🇦 (@azahcrypto) 's Twitter Profile
0xAzah 🇺🇦

@azahcrypto

I love charts. It's green and red 📈. Very beautiful.
My pronoun is : Airbus A400M Atlas.

ID: 949349630231367680

calendar_today05-01-2018 18:39:24

5,5K Tweet

234 Followers

571 Following

Emmanuel Macron (@emmanuelmacron) 's Twitter Profile Photo

L’adjudant-chef Arnaud Frion du 7ème bataillon de chasseurs alpins de Varces est mort pour la France lors d’une attaque dans la région d’Erbil en Irak. À sa famille, à ses frères d’armes, je veux dire toute l’affection et la solidarité de la Nation. Plusieurs de nos militaires

AiBattle (@aibattle_) 's Twitter Profile Photo

Mistral 4 is coming - "Mistral 4 is a powerful hybrid model with the capability of acting as both a general instruction model and a reasoning model. It unifies the capabilities of three different model families - Instruct, Reasoning ( previous called Magistral ), and Devstral -

Mistral 4 is coming

- "Mistral 4 is a powerful hybrid model with the capability of acting as both a general instruction model and a reasoning model. It unifies the capabilities of three different model families - Instruct, Reasoning ( previous called Magistral ), and Devstral -
Mistral AI (@mistralai) 's Twitter Profile Photo

🚀Announcing a strategic partnership with NVIDIA to co-develop frontier open-source AI models, combining Mistral AI’s frontier model architecture and full-stack AI offering with NVIDIA’s leading compute infrastructure and development tools.

🚀Announcing a strategic partnership with NVIDIA to co-develop frontier open-source AI models, combining Mistral AI’s frontier model architecture and full-stack AI offering with NVIDIA’s leading compute infrastructure and development tools.
Mistral AI Labs (@mistralailabs) 's Twitter Profile Photo

🔥 Meet Mistral Small 4: One model to do it all. ⚡ 128 experts, 119B total parameters, 256k context window ⚡ Configurable Reasoning ⚡ Apache 2.0 ⚡ 40% faster, 3x more throughput Our first model to unify the capabilities of our flagship models into a single, versatile model.

🔥 Meet Mistral Small 4: One model to do it all.
⚡ 128 experts, 119B total parameters, 256k context window
⚡ Configurable Reasoning
⚡ Apache 2.0
⚡ 40% faster, 3x more throughput

Our first model to unify the capabilities of our flagship models into a single, versatile model.
0xAzah 🇺🇦 (@azahcrypto) 's Twitter Profile Photo

Is the Strait of Hormuz an obsolete vulnerability? This conceptual map I imagined "the 'Trans-Arabian Strategic Bypass' (TASB)"could be a change in the Arabian Peninsula from its oldest geopolitical chokepoint.

Is the Strait of Hormuz an obsolete vulnerability? This conceptual map I imagined "the 'Trans-Arabian Strategic Bypass' (TASB)"could be a change in the Arabian Peninsula from its oldest geopolitical chokepoint.
Le JDD (@lejdd) 's Twitter Profile Photo

Immigration : le parlement espagnol s'oppose à la régularisation massive de migrants voulue par Sanchez ow.ly/Iq75106w6RQ

Bastion (@bastionmediaoff) 's Twitter Profile Photo

🔴🇫🇷 𝗔𝗟𝗘𝗥𝗧𝗘 𝗜𝗡𝗙𝗢 — À Nice, Christian Estrosi a été REFUSÉ à l’entrée de la prière de l’Aïd au Palais des Expositions, devant près de 7 000 fidèles. Le maire sortant, non invité malgré ses insistances auprès du président de l’Union des musulmans des Alpes-Maritimes

🔴🇫🇷 𝗔𝗟𝗘𝗥𝗧𝗘 𝗜𝗡𝗙𝗢 — À Nice, Christian Estrosi a été REFUSÉ à l’entrée de la prière de l’Aïd au Palais des Expositions, devant près de 7 000 fidèles.

Le maire sortant, non invité malgré ses insistances auprès du président de l’Union des musulmans des Alpes-Maritimes
KNDS France (@knds_france) 's Twitter Profile Photo

Innover c'est anticiper. Chez KNDS nos produits ne répondent pas seulement aux menaces d'aujourd'hui, ils anticipent celles de demain. #WeAreKNDS

David Hendrickson (@teksedge) 's Twitter Profile Photo

🎗️ "Medium-Sized" LLM Burners Coming Soon! 🔥 This Could Make Local HyperToken Generation a Reality. ⚡️ NVIDIA’s worst nightmare? 😱 ⚙️ Application-Specific Hardware Taalas new PCIe ASIC board would burn the entire medium-sized Qwen 3.5-27B LLM straight into silicon 🤯 (already

🎗️ "Medium-Sized" LLM Burners Coming Soon! 🔥

This Could Make Local HyperToken Generation a Reality. ⚡️ NVIDIA’s worst nightmare? 😱

⚙️ Application-Specific Hardware
Taalas new PCIe ASIC board would burn the entire medium-sized Qwen 3.5-27B LLM straight into silicon 🤯 (already
Restitutor Orientis 🇨🇭 (@restitutorii) 's Twitter Profile Photo

🇮🇷⚔️🇺🇸🇮🇱 Iran: Trump ne veut pas intervenir pour Ormuz: "Tous ces pays qui ne peuvent plus se procurer de carburant aviation à cause du détroit d’Ormuz, comme le Royaume-Uni, qui a refusé de s’impliquer contre l’Iran, j’ai une suggestion pour vous: premièrement, achetez aux

🇮🇷⚔️🇺🇸🇮🇱
Iran: Trump ne veut pas intervenir pour Ormuz: "Tous ces pays qui ne peuvent plus se procurer de carburant aviation à cause du détroit d’Ormuz, comme le Royaume-Uni, qui a refusé de s’impliquer contre l’Iran, j’ai une suggestion pour vous: premièrement, achetez aux
Julien | Tech & Invests (@julientechinvst) 's Twitter Profile Photo

Je quote le mec parce que je vois passer ce genre de commentaires régulièrement. L’IA locale c’est un fantasme. Globalement, pour faire de l’inférence, il faut 1) charger le modèle, 2) charger le contexte dans le KV-cache. Il y a globalement 3 précisions possibles en

Secretary Marco Rubio (@secrubio) 's Twitter Profile Photo

Until recently, Hamideh Soleimani Afshar and her daughter were green card holders living lavishly in the United States. Afshar is the niece of deceased Iranian Major General Qasem Soleimani. She is also an outspoken supporter of the Iranian regime who celebrated attacks on

Sudo su (@sudoingx) 's Twitter Profile Photo

before i touch any turbo or quant tricks on the new qwen 3.6-27b dense, i ran a full context sweep on a single 3090. same flags as the march qwen 3.5 baseline. same hardware. the architecture is inherited. exact same vram footprint as 3.5 at every context: 16 gigs at 4k, 18 at

before i touch any turbo or quant tricks on the new qwen 3.6-27b dense, i ran a full context sweep on a single 3090. same flags as the march qwen 3.5 baseline. same hardware.

the architecture is inherited. exact same vram footprint as 3.5 at every context: 16 gigs at 4k, 18 at
Visegrád 24 (@visegrad24) 's Twitter Profile Photo

The Kuban Agrarian University in Russia decided to invite a Russian soldier fighting in Ukraine to speak online to a room full of their students to encourage them to sign up with the Russian Army Instead, a Ukrainian soldier appeared on the screen