"Pieter Abbeel, a Berkeley professor, is part of the team that has started Embodied Intelligence to make it possible for robots to learn on their own.

seen from T1
seen from United States

seen from Malaysia
seen from United States
seen from China
seen from United States
seen from Russia

seen from Malaysia

seen from Malaysia

seen from Malaysia

seen from Switzerland

seen from Malaysia
seen from Germany
seen from Netherlands
seen from Mexico

seen from Malaysia
seen from United States
seen from China
seen from China

seen from Germany
"Pieter Abbeel, a Berkeley professor, is part of the team that has started Embodied Intelligence to make it possible for robots to learn on their own.

Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
Free to watch • No registration required • HD streaming
Embodied intelligence automotive China: Physical AI transforms EV charging, service robots, and logistics ecosystems.
"# Unveiling the Future: Introducing SIMA 2 – The Next Era of AI Agents The future is now, and it goes by the name of SIMA 2! 🚀 We are thrilled to introduce this groundbreaking leap in the evolution of embodied AI agents, now equipped with the astonishing Gemini at its core. This isn’t just another AI; it’s a revolutionary engine of thought and innovation that transcends the limitations of mere instruction following. SIMA 2 is set to redefine the landscape of AI interactions in complex, three-dimensional environments. So, what exactly makes SIMA 2 the game changer we’ve all been waiting for? Let’s break it down. 🔵 **Advanced Reasoning**: Gone are the days when AI could only follow rudimentary commands. SIMA 2 is capable of executing high-level strategies across various games, articulating its intentions, processing visual information, and strategically outlining the steps it's undertaking. We’re talking about an AI that not only understands the “how” but the “why” behind its actions, paving the way for more profound interactions. 🔵 **Improved Generalization**: Imagine an AI that can synthesize learnings from one game to another seamlessly. SIMA 2 does just that! It can take concepts like “mining” from one context and apply it to “harvesting” in another, making it a powerhouse of conceptual connectivity. This remarkable ability allows it to efficiently adapt and solve problems across different scenarios with agility and insight. 🔵 **Self-Improvement**: Here’s where things get really exciting. With SIMA 2, self-improvement isn’t just a buzzword; it’s a reality! Through a continuous cycle of trial-and-error learning and the rich feedback from Gemini, this AI can autonomously master entirely new skills in environments it has never encountered before—all without needing a nudge from humans. The future of independent learning is upon us! 🔵 **Adaptability**: When we challenged SIMA 2 in simulated 3D worlds created using our cutting-edge Genie 3 model, the results were nothing short of astonishing. Witnessing its ability to navigate these intricate environments, follow nuanced instructions, and actively pursue objectives illustrates a level of adaptability unseen in AI until now. It’s a testament to how far we’ve come toward integrating AI into tangible, real-world applications. This is not merely academic research; this marks a significant stride toward deploying AI in robotics and brings us closer to realizing Artificial General Intelligence (AGI) in the physical realm. The implications of SIMA 2 extend far beyond gaming and simulations, reaching into the very fabric of how we envision AI aiding human endeavors in various sectors. Ready to embrace the future? Dive deeper into the remarkable journey of SIMA 2 and its transformative capabilities by visiting [our website](https://ift.tt/Uebsv95). For more updates and insights, subscribe to our channel on [YouTube](https://www.youtube.com/@googledeepmind), follow us on [Twitter](https://twitter.com/GoogleDeepMind), check out our feed on [Instagram](https://ift.tt/ZkES2Ac), and connect with us on [LinkedIn](https://ift.tt/UgTZo2i). The journey into the next era of AI collaboration has just begun—join us! 🎉"
🤖 Physical AI革命:2025年Q4の重要論文10選
AIが物理世界を理解し、行動する時代が来た。
2025年10月〜12月にかけて、Physical AI分野で画期的な研究が次々と発表されました。ここでは、その中から特に重要な10本の論文をピックアップしてお届けします。
🔥 注目ハイライト
1. NVIDIAの「Cosmos-Predict2.5」
2億本のビデオで学習した世界シミュレーター
Text・Image・VideoからリアルなWorld Simulationを生成。Sim2Real変換も高精度で実現し、ロボティクス研究の新たな基盤に。
2. VAT(Vision Action Transformer)
LIBEROベンチマーク98.15%達成
ViTの全階層特徴量を活用した革新的アーキテクチャ。ロボットの知覚-行動融合において従来手法を大幅に凌駕。
3. APEX:人間-AI協調システム
Mixed Realityで科学実験を変革
ウェアラブルデバイスを通じたCo-Embodied Intelligence。人間の精密さとAIの推論を融合し、製造現場を自律プロセスに。
💡 5つの重要トレンド
統合フレームワーク — 知覚・推論・行動の壁が消える
World Models — 物理世界のシミュレーションと予測
エッジ展開 — MERINDAでモバイルGPU比11倍のDRAM削減
アクセシビリティ — LoRA-VLAで低コストロボットにも最先端AI
評価基盤 — PAI-Benchで2,808の実世界ケースを標準化
🧠 パラダイムシフト
「知能はシンボル処理からではなく、物理的経験から創発する」 — Fundamentals of Physical AI
身体性・知覚・行動・学習・自律性・文脈感受性。この6つの原則が相互作用ループとして機能するとき、真のPhysical AIが生まれる。
🚀 今後の研究方向
物理的一貫性メトリクスの標準化
Long-horizon予測の改善
Sim2Real/Real2Simギャップの縮小
安全性・説明可能性の向上
AIは「考える」だけでなく「動く」存在へ。
Physical AIの波は、ロボティクス、製造、ヘルスケア、そして私たちの日常を変えていく。2025年はその転換点となる年だ。
詳細なレポートはarXivで各論文をチェック!
Whole Like The Whole Of An Image
In this post I want to focus on how musical composition can naturally discover OOO insights—how rhythm and sound patterns lead our bodies to philosophical understanding. The equal weight given to salt and stars demonstrates how our embodied musical intuition grasps what reason struggles to articulate. This poem came together quickly, with memory and body processes holding concept and imagery…

Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
Free to watch • No registration required • HD streaming
Week 6
While going through the Cyberfeminism index this week, I found a paper that lead me down some interesting thought processes. The paper was ‘Why are Mammals called Mammals?’ This paper is about the cultural relationship between society and science, with our view of mammals being shaped by breasts, women, and animals. While this paper was more of a feminist look at scientific classification, it made me start thinking about evolutionary connections to intelligence. This is partially because in my other course, Poli 13DR we had just read Waller’s Perpatrators of Genocide, which at one point discussed the ways evolution affects our psychology, specifically in the cases of xenophobia and the formation of in-groups and out-groups. Both of these papers made me start thinking about the ways that evolutionary instincts have translated into social tendencies, and also by extension social intelligence. There are parts of the human psyche that still seem mysterious to us all. To some, our inability to understand our own mind’s complexity is evidence and symptom of our great intelligence. However, whenever I face the human mind through the lens of evolution those mysteries seem to be more like developments of instinct into situations that no longer call for that instinct.
Our intelligence and ingenuity was a design from natural selection. The intelligent creatures that surround us were formed by the same forces. So it seems that intelligence is a natural byproduct of living long enough in the world for it to shape you for your environment. You don’t just understand the world around you inherently, you understand it as a tool for survival, that understanding shaped by the environment itself.
All of this made me consider that in our final assignment, evolution should be something strongly considered. With the ecosystem and environment provided, in what ways would a creature be forced to genetically adapt? What ‘intelligence’ would suit the creature in this space? How would it be shaped to understand its world?
Assignment #1: Defining Intelligence
Intelligence is about the expansion of knowledge to form an understanding of the world and other people. Part of this expansion means recognizing new information, and creating a network out of this information. A fundamental part of creating this network is the identification of patterns, and with these patterns forming connections about and between different data points. The creation of these networks and connections is what brings true understanding on a subject, these abilities presupposing intelligence.
3 KEYWORDS: Network, Connections, Understanding
The first image is a sketching of my ideas for images to exemplify my keywords.
NETWORK: Image 2 is a result of a p5.js code I wrote that creates a network of nodes from two starting nodes. The colorful lines are connections made between nodes of the same color.
LINK TO CODE:
A web editor for p5.js, a JavaScript library with the goal of making coding accessible to artists, designers, educators, and beginners.
CONNECTIONS: Image 3 is a digital drawing I did on procreate meant to represent the patterns humans identify in nature, creating fictional connections. These patterns may not have any actual connection (the cloud isn't really a bunny, and the stars aren't purposefully representing anything we know), but they were connections we created due to our intelligence.
UNDERSTANDING: Image 4 is a collage between an anatomical image and a digital drawing. This is meant to be a representation of emotional recognition and sympathy, which is a way we understand other human beings. I wanted the anatomical aspect as a reminder that despite the fact it is an emotional reflection, they are still based on a subconscious logical connection. We can't sympathize without the ability to look at previously catalogued examples of our own emotions, as well as make the connection between that experience and the expressions of the person in front of us.
LINK TO REFERENCE IMAGE: