For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.
AI will transform every industry, power every company, and be built by every country.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
The world needs both frontier closed models and frontier open models.
https://t.co/AUKzoQ5Ikb
🚀 SANA-Video 2.0 is here!
A full-stack optimized video model designed for efficiency — while still delivering high quality.
We combine a hybrid architecture closely related to the recent Kimi K3 design, adopt Self-Flow from FLUX 3, and further accelerate it with our own Sol-Engine.
Key technical ingredients:
🧠 Hybrid Linear–Softmax Attention
3 gated linear layers + 1 gated-softmax anchor (75% linear / 25% softmax)
🧱 Block Attention Residuals (AttnRes)
Boosts deep-layer effective rank by ~12%
🏗️ Unified 5B & 14B models
Trained from scratch under very limited resources:
• 5B → only 16 nodes of H100
• 14B → 48 nodes of B200
Results 📊
• 84.30 VBench Total
• 3.2× faster DiT forward than matched full-softmax at 720p/60s
• 720p/5s in 13.06s on a single H100 with Sol-Engine
• 120× faster than Wan 2.2-A14B under the same one-H100 setup
SANA-Video 2.0 shows that high-quality 720p video generation can be both efficient and practical on a single GPU.
🎬 Project: https://t.co/40uvReE9oE
📄 Paper: https://t.co/cmJZhU64pp
💻 Code: https://t.co/3UlHHqdv0Y
Proud of the team! 🎉 More details below 🧵
We (@lawrence_cjs, @yuyangzhao_ , @shanasaimoe) from the SANA team just posted a blog on the core of Linear Attention: how it achieves infinite context lengths with global awareness but constant memory usage! We explore state accumulation mechanics, the evolution from Softmax to Linear KV caches, applications in long video generation (like our SANA-Video), and advanced strategies for LLMs to overcome retrieval limits. Perfect for scalable AI! Read more: https://t.co/yJRtVTkt1r
I'm busy with work lately, so it has been a while, but I just published a new Magisk canary build!
The most notable change is the brand new magic mount backend, contributed by @shanasaimoe and LSPosed folks. It now properly supports overlayfs systems 🎉
The latest Google Pixel 6 pwned with a 0day in kernel! Achieved arbitrary read/write to escalate privilege and disable SELinux without hijacking control flow. The bug also affects Pixel 6 Pro, other Pixels are not affected :)
Here we share a video to present how our E2V-SDE paper (that has been accepted to #CVPR2022) largely consists of texts that are uncredited verbatim copies from more than 10 previously published papers. Enjoy!
https://t.co/BrlGpQdNLH
#CVPR22@CVPR@CSProfKGD@CVPRConf
@shawnwillden For some reason, Pixel 6 moved its stab to /system/etc. MagiskInit was no longer able to patch the mount tables at runtime, and therefore things broke 😃
The commit that fixed the issue also linked to the commit in AOSP that introduced this change.
https://t.co/ZpyufOD5xZ
Here is a new *pure Java* library to bypass Android's hidden API restrictions.
The details are written in Chinese, but the gist is: using Java's Unsafe API, it directly accesses native pointers in ART and reconstruct Java reflection objects from it.
https://t.co/LU2rFnY2Dq