Alex Millà

My daily focus: cybersecurity and vulnerability management

[Alex Millà @AlexMilla_]

@SoyITPro ¿Qué la compre Broadcom? Todo es posible, pero no por el momento hasta que no recupere la inversión. 🤪

[Nintenderos @Nintenderos]

¡Sorteamos 3 copias de Obakeidoro 2 para Nintendo Switch o Nintendo Switch 2!

1⃣ Sigue a @Nintenderos y @Obakeidoro_en
2⃣ RT aquí

¡Suerte! - https://t.co/STI8jlDlii

[BandaAncha @bandaanchaeu]

TuLotero sufre una filtración: los DNIs de sus usuarios comprometidos

👇

https://t.co/5NVDhabLWp

[Shodan @shodanhq]

Shodan Membership sale is live until 2026-08-09 00:00 UTC (~48 hours from now): https://t.co/0iqiq2vTpn

[Quake @Quake]

Created in celebration of Quake's 30th anniversary, a new free episode, Dawn of the Machine, is available now!

Building on @MachineGames' acclaimed work on Dimension of the Machine and Dimension of the Past, it delivers ferocious combat, labyrinthine level design, and nightmarish new realms to explore.

[Hugging Models @HuggingModels]

Check out granite-4.1-8b.

[Alex Millà @AlexMilla_]

Apuesta por los drones y la IA: Ribera d’Ebre aspira a convertirse en un 'hub' tecnológico https://t.co/HD1s62V9YH

[Qwen @Alibaba_Qwen]

📢Meet Qwen3.8-Max — our most capable model to date.

Next week, the open weights of Qwen3.8-Max will be released, and Qwen3.8-27B is also going open-weights to meet you all!🎉

Qwen3.8-Max, a new bar for coding and cowork at 2.4T parameters:

- Autonomous coding: 10+ days of self-evolving development, from empty folder to production without hand-holding, complete project trace in the GitHub：https://t.co/iVHZWQoeSo
- Real work, real results: Production-quality deliverables across hundreds of professions.
- Long-horizon mastery: System-level autonomous planning with closed-loop adaptive learning, driving 500+ turns of chip design optimization and 365 days of e-commerce strategy.
- Native multimodal intelligence: Vision isn't just input — it's a continuous feedback loop for planning, execution, and self-correction.

💰Pricing:
Input: $2.0 / M tokens
Output: $6.0 / M tokens
Implicit Caching: $0.25 / M tokens

Start building with Qwen3.8-Max! 🚀

📖 Blog: https://t.co/iwjmQxLBof
✅ Qwen Studio: https://t.co/4V2pFvDovG
⚡ API: https://t.co/gAGqaLQGbN

[吉野@連邦(renpou.com) @yoshinokentarou]

スーファミの大名作「ゼルダの伝説 神々のトライフォース」が、まさかの「ニンテンドー3DS」へ2画面仕様で勝手移植!!

オリジナルの雰囲気はそのままに、3DSの特徴である、上下2画面表示、タッチ操作、ワイド画面などに対応

このプロジェクト、今後も続く予定らしいぞッッ!!
https://t.co/CbRgcocK9g

[DeepSeek @deepseek_ai]

🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!

🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇
🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!

Check out the configuration details in our official API docs: https://t.co/smCwQZMeiq

[Kimi.ai @Kimi_Moonshot]

We've open-sourced MoonEP, our high-performance communication library for distributed MoE workloads.

Built to make expert-parallel communication more efficient at scale, MoonEP helps reduce communication overhead in large MoE training and inference systems.

Explore on GitHub:
https://t.co/h2Rcwg88dQ

[Hugging Models @HuggingModels]

Finally, guys Kimi-K3 Open-weights are out on hugging face as promised

[Kimi.ai @Kimi_Moonshot]

Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B
Tech report: https://t.co/yeu6cjpMCT
Tech blog: https://t.co/YTfiMSNM1f

[Kaspersky @kaspersky]

Skip the coffee. We'll bring the tea.
SOC talk. Run it, build it, or improve it. Three models, one hour.
📅August 4, 16:00 CEST. Register and join us.
https://t.co/8ykRndLu8Y

#SOC #Cybersecurity

[Hackmanac @H4ckmanac]

🚨Cyber Alert ‼️

🇪🇸Spain - 𝗠𝗶𝗻𝗶𝘀𝘁𝗲𝗿𝗶𝗼 𝗱𝗲 𝗔𝘀𝘂𝗻𝘁𝗼𝘀 𝗘𝘅𝘁𝗲𝗿𝗶𝗼𝗿𝗲𝘀, 𝗨𝗻𝗶𝗼́𝗻 𝗘𝘂𝗿𝗼𝗽𝗲𝗮 𝘆 𝗖𝗼𝗼𝗽𝗲𝗿𝗮𝗰𝗶𝗼́𝗻

The threat actor GordonFreeman claims to have breached Spain’s Ministry of Foreign Affairs, European Union and Cooperation, allegedly exploiting an IDOR vulnerability to access and exfiltrate 1,953,721 records from its primary database.

Threat actor: GordonFreeman
Sector: Gov / Mil / LE
Data exposure (claimed): 1,953,721 records
Data type: Full names, primary and secondary email addresses, dates of birth, nationalities, phone numbers, identity document details, postal addresses, user IDs, CSRF authentication tokens, internal IP addresses, company identifiers and system metadata
Observed: Jul 25, 2026
Status: Pending verification
ESIX©: 6.97

Full details and impact assessment on https://t.co/eB7qgxKFAa

[Unsloth AI @UnslothAI]

Introducing Unsloth for AMD 🚀
You can now train & run LLMs on your AMD hardware

• We collaborated with AMD to enable you to train & run 500+ models on AMD GPUs
• Works on Windows, WSL, Linux
• Train Qwen, Gemma on 3GB VRAM

GitHub: https://t.co/2kXqhhvLsb

Works on Radeon, Instinct, Ryzen and data center GPUs with up to 2× faster with 70% less VRAM and no accuracy loss via our custom Triton kernels and math algorithms. We also support optimized ROCm builds for GGUF & Safetensors inference.

Unsloth is an open-source local UI for faster LLM training and inference, with tool-call healing, code execution, secure web search, remote APIs, and HTTPS deployment. Connect local models to Claude Code, Codex agents and run the latest Kimi, GLM, DeepSeek, Qwen3.6, and Gemma 4 models.

🔗Blog + Guide: https://t.co/U9LqyRjFdj

[Alex Millà @AlexMilla_]

@NeonZoux Imposible comprar, con ofertas como estas siempre da errores la página, que curioso.

No se puede combinarse cuando solo hay 1

[Miguel Ángel Durán @midudev]

¡Esto es histórico! NVIDIA, Microsoft, Mozilla y Linux Foundation, entre otras, han firmado una carta a favor de los modelos de pesos
abiertos.

Defienden que abrir la IA es CLAVE para competir, reducir costes, evitar depender de unas pocas empresas y mejorar la seguridad.

Y dejan otro mensaje importante: usar las respuestas de un modelo para entrenar otro mediante destilación es una práctica legítima y habitual. NO debería considerarse automáticamente un robo.

Y la frase más bestia: “Depender únicamente de modelos cerrados no es intrínsecamente seguro.”

BOOM.

[FATAL FURY: City of the Wolves @FATALFURY_PR]

／
🎊 Celebrate the Revenge Edition Launch! 🎊
＼

1 lucky winner will receive a PlayStation®5 Pro!🎮

How to enter:
✅ Follow @FATALFURY_PR
✅ Repost this post

#CotW #FatalFury

[Alex Millà @AlexMilla_]

Jensen Huang dice que las empresas estadounidenses deberían usar IA chinas: «Los modelos abiertos son necesarios»

https://t.co/ZDd6TSmnEc