Sebastian Russo
263 posts
@sebbrusso
product @googledeepmind | stanford cs
much of the nature of the world is explained by the nature of bureaucracy and yet - bureaucracy is rarely written about well beyond cliches. the great authors and people who worked in a large organization are generally disjoint. more common in east asian media
🚨 New SISL preprint: State-of-the-art language reward models are still badly biased. Past fixes overcorrect, some can be fixed with simple latent interventions, and some indicate the need for larger efforts.
RL on LLMs inefficiently uses one scalar per rollout. But users regularly give much richer feedback: "make it formal," "step 3 is wrong." Can we train LLMs on this human-AI interaction? We introduce RL from Text Feedback, with 1) Self-Distillation; 2) Feedback Modeling (1/n) 🧵
🎨 Qwen-Image-Layered is LIVE — native image decomposition, fully open-sourced! ✨ Why it stands out ✅ Photoshop-grade layering Physically isolated RGBA layers with true native editability ✅ Prompt-controlled structure Explicitly specify 3–10 layers — from coarse layouts to fine-grained details ✅ Infinite decomposition Keep drilling down: layers within layers, to any depth of detail 🤗 Hugging Face: huggingface.co/Qwen/Qwen-Imag… 🧩 ModelScope: modelscope.cn/models/Qwen/Qw… 💻 GitHub: github.com/QwenLM/Qwen-Im… 📝 Blog: qwen.ai/blog?id=qwen-i… 📄 Technical Report: arxiv.org/abs/2512.15603 🚀 Demo (HF): huggingface.co/spaces/Qwen/Qw… 🚀 Demo (ModelScope): modelscope.cn/studios/Qwen/Q…
WSJ Edit Board -- The University of California eliminated the SAT as an admissions requirement five years ago. Now arrives the dispiriting result: Many freshmen at one of its top public universities can’t do middle-school math. wsj.com/opinion/a-math… via @WSJopinion
I think it’s very very possible that cursor can surpass all frontier labs on a coding model
