Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Introducing Moto: Latent Motion Token as the Bridging Language for Robot Manipulation. Motion prior learned from videos can be seamlessly transferred to robot manipulation. Code and model released! Yi CHEN Yuying Ge Yixiao Ge Mingyu Ding Ying Shan

Xihui Liu

2,140 subscribers

15,039 views • 1 year ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Super excited about Hydra-0 from Hongyu Li and team! The key idea is to use flow as a shared visual interface across embodiments/objects for controllable video generation, allowing a single generalist world model to learn from human, handheld-gripper, and robot interaction data. My favorite result is the video below: start from a real video of a human doing the task (left), extract the desired object flow, and condition the model on that flow (right). The model then hallucinates a plausible robot motion that could produce the same object motion. Very cool glimpse of how a generalist world model can bridge human demonstrations and robot control.

Super excited about Hydra-0 from Hongyu Li and team! The key idea is to use flow as a shared visual interface across embodiments/objects for controllable video generation, allowing a single generalist world model to learn from human, handheld-gripper, and robot interaction data. My favorite result is the video below: start from a real video of a human doing the task (left), extract the desired object flow, and condition the model on that flow (right). The model then hallucinates a plausible robot motion that could produce the same object motion. Very cool glimpse of how a generalist world model can bridge human demonstrations and robot control.

Yunzhu Li

10,695 views • 29 days ago

✨ Any static 3D assets ➡️ 4D dynamic worlds. Introducing CHORD, a universal framework for generating scene-level 4D dynamic motion from any static 3D inputs. It generalizes surprisingly well across a wide range of objects 🤯 and can even be used to learn robotics manipulation policy 🤖! Project page: Dive deeper in a 🧵: 1/n

✨ Any static 3D assets ➡️ 4D dynamic worlds. Introducing CHORD, a universal framework for generating scene-level 4D dynamic motion from any static 3D inputs. It generalizes surprisingly well across a wide range of objects 🤯 and can even be used to learn robotics manipulation policy 🤖! Project page: Dive deeper in a 🧵: 1/n

Chen Geng

43,510 views • 8 months ago

30 minutes of video. Robot learns the task. Open-source, end-to-end. An open-source framework for training robot policies from only 30 minutes of human egocentric videos captured via Meta Aria glasses: Achieving zero-shot transfer to robots without any robot data collection. The method relies on Interaction-Centric Tokens that encode hand-object spatial relationships invariant to embodiment and viewpoint, supplemented by auxiliary objectives like object motion prediction and latent consistency to extract richer supervision signals from the same data. HumanEgo demonstrates strong cross-embodiment, cross-environment performance on bimanual tasks, outperforming baselines like ACT and teleop data while being trainable on a single RTX 4090 GPU. Thanks for sharing, Zhi (Leo) Wang. 📌 Website: Paper: Code: Video: ——- Weekly robotics and AI insights. Subscribe free:

30 minutes of video. Robot learns the task. Open-source, end-to-end. An open-source framework for training robot policies from only 30 minutes of human egocentric videos captured via Meta Aria glasses: Achieving zero-shot transfer to robots without any robot data collection. The method relies on Interaction-Centric Tokens that encode hand-object spatial relationships invariant to embodiment and viewpoint, supplemented by auxiliary objectives like object motion prediction and latent consistency to extract richer supervision signals from the same data. HumanEgo demonstrates strong cross-embodiment, cross-environment performance on bimanual tasks, outperforming baselines like ACT and teleop data while being trainable on a single RTX 4090 GPU. Thanks for sharing, Zhi (Leo) Wang. 📌 Website: Paper: Code: Video: ——- Weekly robotics and AI insights. Subscribe free:

Ilir Aliu

17,077 views • 3 months ago

Live from World Robot Conference 2026 in Beijing: ROBOTERA is showcasing teleoperation of the XHAND 1 Pro dexterous hand with MANUS™ gloves. Video 1: a ROBOTERA operator demonstrates high-precision teleoperation. Video 2: a visitor tries the setup without first running a user-specific calibration, yet the teleoperation remains stable, smooth, and responsive. With 21 active degrees of freedom, fully direct-drive actuation, and integrated tactile sensing, XHAND 1 Pro is built for advanced robotic manipulation and embodied intelligence research. MANUS provides high-fidelity, low-latency hand motion data as real-time input for robot control. #WRC2026 #robotics #humanoids #teleoperation #PhysicalAI

Live from World Robot Conference 2026 in Beijing: ROBOTERA is showcasing teleoperation of the XHAND 1 Pro dexterous hand with MANUS™ gloves. Video 1: a ROBOTERA operator demonstrates high-precision teleoperation. Video 2: a visitor tries the setup without first running a user-specific calibration, yet the teleoperation remains stable, smooth, and responsive. With 21 active degrees of freedom, fully direct-drive actuation, and integrated tactile sensing, XHAND 1 Pro is built for advanced robotic manipulation and embodied intelligence research. MANUS provides high-fidelity, low-latency hand motion data as real-time input for robot control. #WRC2026 #robotics #humanoids #teleoperation #PhysicalAI

MANUS™

11,998 views • 28 days ago

Building robots that can effectively operate alongside human workers is difficult. 🛠️ Advances in open-source physics, open foundation models, and frameworks are helping accelerate physical #AI deployment. ✔️ Newton Physics Engine, an open-source GPU-powered simulation built on OpenUSD, speeds up robot learning for advanced manipulation and mobility. ✔️ NVIDIA Cosmos Reason, an open reasoning vision language model, gives robots the ability to think like humans using prior knowledge, common sense and physics ✔️NVIDIA Isaac GR00T N1.6, an open robot foundation model, enables humanoids to understand ambiguous instructions Leading robotics developers including Agility Robotics, Lightwheel, Mentee Robotics, UniversalRobots, and Wandelbots are adopting simulation technologies and libraries to accelerate physical AI development and deployment. Omniverse Ambassador Dylan Tobin built an AI chatbot trained on Isaac Sim workflows, helping devs navigate Omniverse faster. Read the full blog 👉

Building robots that can effectively operate alongside human workers is difficult. 🛠️ Advances in open-source physics, open foundation models, and frameworks are helping accelerate physical #AI deployment. ✔️ Newton Physics Engine, an open-source GPU-powered simulation built on OpenUSD, speeds up robot learning for advanced manipulation and mobility. ✔️ NVIDIA Cosmos Reason, an open reasoning vision language model, gives robots the ability to think like humans using prior knowledge, common sense and physics ✔️NVIDIA Isaac GR00T N1.6, an open robot foundation model, enables humanoids to understand ambiguous instructions Leading robotics developers including Agility Robotics, Lightwheel, Mentee Robotics, UniversalRobots, and Wandelbots are adopting simulation technologies and libraries to accelerate physical AI development and deployment. Omniverse Ambassador Dylan Tobin built an AI chatbot trained on Isaac Sim workflows, helping devs navigate Omniverse faster. Read the full blog 👉

NVIDIA

48,500 views • 11 months ago

ZACHXBT FLAGS POSSIBLE INSIDER ACTIVITY ON $RAVE On-chain investigator ZachXBT has raised concerns about alleged insider manipulation of the RaveDAO token, posting that insiders appear to control more than 90% of $RAVE's supply. In his post, ZachXBT pointed to what he described as pump-and-dump activity originating from Binance, @bitgetglobal, and Gate.io, and called on Binance's He Yi and Bitget CEO Gracy Chen @Bitget to investigate and offboard those involved. He also offered a $10,000 bounty for whistleblowers who come forward with evidence related to the alleged scheme. Bitget CEO Gracy Chen publicly responded shortly after, confirming the exchange had opened an investigation into $RAVE. $RAVE is up over 10,000% in the past 30 days, with the token rallying another 44% on Saturday.

ZACHXBT FLAGS POSSIBLE INSIDER ACTIVITY ON $RAVE On-chain investigator ZachXBT has raised concerns about alleged insider manipulation of the RaveDAO token, posting that insiders appear to control more than 90% of $RAVE's supply. In his post, ZachXBT pointed to what he described as pump-and-dump activity originating from Binance, @bitgetglobal, and Gate.io, and called on Binance's He Yi and Bitget CEO Gracy Chen @Bitget to investigate and offboard those involved. He also offered a $10,000 bounty for whistleblowers who come forward with evidence related to the alleged scheme. Bitget CEO Gracy Chen publicly responded shortly after, confirming the exchange had opened an investigation into $RAVE. $RAVE is up over 10,000% in the past 30 days, with the token rallying another 44% on Saturday.

BSCN

37,739 views • 5 months ago

𝗥𝗼𝗯𝗼𝘁𝘀 𝗱𝗼𝗻’𝘁 𝗻𝗲𝗲𝗱 𝗺𝗼𝗿𝗲 𝗱𝗲𝗺𝗼𝗻𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻𝘀. 𝗧𝗵𝗲𝘆 𝗻𝗲𝗲𝗱 𝘁𝗼 𝗹𝗲𝗮𝗿𝗻 𝗳𝗿𝗼𝗺 𝗳𝗮𝗶𝗹𝘂𝗿𝗲 — 𝗮𝗳𝘁𝗲𝗿 𝘄𝗮𝘁𝗰𝗵𝗶𝗻𝗴 𝗵𝘂𝗺𝗮𝗻𝘀. Most robot learning systems assume failure is the end of learning. In our new work, we study whether robots can improve after deployment by learning from their own failures, without any human intervention, teleoperation, or corrective labels. The key idea is simple: human videos contain structure about how the world works. We use them to learn cross-embodiment representations of action, dynamics, and value, enabling a shared predictive space between human behavior and robot experience. This allows a new learning loop: 👉 pretrain on human videos 👉 deploy robot policy 👉 observe failures 👉 reinterpret failures using human priors 👉 improve autonomously We evaluate this across 7 real-world manipulation tasks, showing: 📈 40% → 81% success rate 🏆 Strong improvements over π0.6 RECAP and RISE ✔️ Zero human intervention during post-deployment improvement 🧬 Generalizes across robot embodiments and policy backbones A key finding is that explicit failure repair significantly outperforms failure reweighting, yielding substantially larger gains under identical data conditions (+25 pts vs +5 pts on the same π0.5 base policy). Overall, the results suggest a shift in how we think about robot learning: Human videos are not only for pretraining policies. They can provide the structure needed for continual self-improvement after deployment. 📄 Paper: 🌐 Project: I am grateful for working with the fantastic leads Hanzhi Chen and Anran Zhang, and our collaborators Simon Schaefer, Kejia Chen, Shi Chen, Daniel Cremers. Special thanks to Stefan Leutenegger for co-advising this project with me. ETH Zürich TU München Microsoft Check out Hanzhi's 🧵 for more details

𝗥𝗼𝗯𝗼𝘁𝘀 𝗱𝗼𝗻’𝘁 𝗻𝗲𝗲𝗱 𝗺𝗼𝗿𝗲 𝗱𝗲𝗺𝗼𝗻𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻𝘀. 𝗧𝗵𝗲𝘆 𝗻𝗲𝗲𝗱 𝘁𝗼 𝗹𝗲𝗮𝗿𝗻 𝗳𝗿𝗼𝗺 𝗳𝗮𝗶𝗹𝘂𝗿𝗲 — 𝗮𝗳𝘁𝗲𝗿 𝘄𝗮𝘁𝗰𝗵𝗶𝗻𝗴 𝗵𝘂𝗺𝗮𝗻𝘀. Most robot learning systems assume failure is the end of learning. In our new work, we study whether robots can improve after deployment by learning from their own failures, without any human intervention, teleoperation, or corrective labels. The key idea is simple: human videos contain structure about how the world works. We use them to learn cross-embodiment representations of action, dynamics, and value, enabling a shared predictive space between human behavior and robot experience. This allows a new learning loop: 👉 pretrain on human videos 👉 deploy robot policy 👉 observe failures 👉 reinterpret failures using human priors 👉 improve autonomously We evaluate this across 7 real-world manipulation tasks, showing: 📈 40% → 81% success rate 🏆 Strong improvements over π0.6 RECAP and RISE ✔️ Zero human intervention during post-deployment improvement 🧬 Generalizes across robot embodiments and policy backbones A key finding is that explicit failure repair significantly outperforms failure reweighting, yielding substantially larger gains under identical data conditions (+25 pts vs +5 pts on the same π0.5 base policy). Overall, the results suggest a shift in how we think about robot learning: Human videos are not only for pretraining policies. They can provide the structure needed for continual self-improvement after deployment. 📄 Paper: 🌐 Project: I am grateful for working with the fantastic leads Hanzhi Chen and Anran Zhang, and our collaborators Simon Schaefer, Kejia Chen, Shi Chen, Daniel Cremers. Special thanks to Stefan Leutenegger for co-advising this project with me. ETH Zürich TU München Microsoft Check out Hanzhi's 🧵 for more details

Oier Mees

12,514 views • 2 months ago

TESLA HALTED MODEL S AND MODEL X PRODUCTION TO BUILD AN ARMY OF OPTIMUS ROBOTS The Fremont assembly line was torn down in 46 days. In its place, Tesla is building a line for humanoid production, aiming for a million units a year A humanoid robot is a body shaped like a human. Physical AI is the intelligence that controls that body Walking and making coffee is often just imitation learning from a scripted routine. But once the environment shifts, the learned trick stops working Language models had the entire internet to train on. Robotics has nothing close to that scale of data, which is why one giant brain hasn't worked for anyone yet The industry is moving toward modularity instead - separate models for vision, movement, and planning, each improved on its own The real question is no longer whether a robot can move impressively. It's whether it can pull its sensors into one picture of the world and adapt to whatever wasn't scripted for it

TESLA HALTED MODEL S AND MODEL X PRODUCTION TO BUILD AN ARMY OF OPTIMUS ROBOTS The Fremont assembly line was torn down in 46 days. In its place, Tesla is building a line for humanoid production, aiming for a million units a year A humanoid robot is a body shaped like a human. Physical AI is the intelligence that controls that body Walking and making coffee is often just imitation learning from a scripted routine. But once the environment shifts, the learned trick stops working Language models had the entire internet to train on. Robotics has nothing close to that scale of data, which is why one giant brain hasn't worked for anyone yet The industry is moving toward modularity instead - separate models for vision, movement, and planning, each improved on its own The real question is no longer whether a robot can move impressively. It's whether it can pull its sensors into one picture of the world and adapt to whatever wasn't scripted for it

iamigorekk

22,205 views • 1 month ago

𝗗𝗼𝗻'𝘁 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗲 𝗿𝗼𝗯𝗼𝘁 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹𝘀. 𝗦𝘁𝗲𝗲𝗿 𝘁𝗵𝗲𝗺 𝘄𝗶𝘁𝗵 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀 𝗶𝗻𝘀𝘁𝗲𝗮𝗱, 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗯𝗮𝘀𝗲 𝗽𝗼𝗹𝗶𝗰𝘆 Modern VLAs and world-action models can perform impressive manipulation skills, but adapting them reliably to new robots and tasks remains challenging. A natural solution is DAgger-style online imitation learning: deploy the robot, collect human corrections, and update the policy. Yet foundation models are fragile in the low-data regime, fine-tuning on a handful of interventions can improve one behavior while degrading others. Online post-training or reinforcement learning can require costly data collection and exploration, making real-world learning expensive and potentially unsafe. In our new paper, 𝗙𝗹𝗼𝘄𝗗𝗔𝗴𝗴𝗲𝗿, we take a different approach: 𝗜𝗻𝘀𝘁𝗲𝗮𝗱 𝗼𝗳 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹, 𝘄𝗲 𝗹𝗲𝗮𝗿𝗻 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗲𝗲𝗿 𝗶𝘁 𝗳𝗿𝗼𝗺 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀. The key idea is 𝗮𝗰𝘁𝗶𝗼𝗻 𝗶𝗻𝘃𝗲𝗿𝘀𝗶𝗼𝗻: we map human corrective actions back into the latent noise space of the frozen generative policy. These latent targets train a lightweight controller that adapts the robot while preserving the original model's capabilities. Across simulation and real robots, FlowDAgger: 📈 Learns from only 5–20 human intervention episodes 🏆 Outperforms supervised fine-tuning and latent-space reinforcement learning 🤖 Works across VLAs, diffusion policies, and world-action models ✔️ Provides reliable improvements without modifying the pretrained policy We believe this offers a practical path toward making robot foundation models improve during deployment, learning from the way humans naturally teach: through corrections. 📄 Paper: 🌐 Project: 💻 Code: This project was led by my amazing colleague Michael Murray with help from Daphne Chen, Simran Bagaria, Dean Fortier, Tess Hellebrekers, Harshavardhan Reddy Gajarla, Galen Mullins and Andrey Kolobov at Microsoft Research and Maya Cakmak at University of Washington

𝗗𝗼𝗻'𝘁 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗲 𝗿𝗼𝗯𝗼𝘁 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹𝘀. 𝗦𝘁𝗲𝗲𝗿 𝘁𝗵𝗲𝗺 𝘄𝗶𝘁𝗵 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀 𝗶𝗻𝘀𝘁𝗲𝗮𝗱, 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗯𝗮𝘀𝗲 𝗽𝗼𝗹𝗶𝗰𝘆 Modern VLAs and world-action models can perform impressive manipulation skills, but adapting them reliably to new robots and tasks remains challenging. A natural solution is DAgger-style online imitation learning: deploy the robot, collect human corrections, and update the policy. Yet foundation models are fragile in the low-data regime, fine-tuning on a handful of interventions can improve one behavior while degrading others. Online post-training or reinforcement learning can require costly data collection and exploration, making real-world learning expensive and potentially unsafe. In our new paper, 𝗙𝗹𝗼𝘄𝗗𝗔𝗴𝗴𝗲𝗿, we take a different approach: 𝗜𝗻𝘀𝘁𝗲𝗮𝗱 𝗼𝗳 𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘁𝗵𝗲 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹, 𝘄𝗲 𝗹𝗲𝗮𝗿𝗻 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗲𝗲𝗿 𝗶𝘁 𝗳𝗿𝗼𝗺 𝗵𝘂𝗺𝗮𝗻 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝗼𝗻𝘀. The key idea is 𝗮𝗰𝘁𝗶𝗼𝗻 𝗶𝗻𝘃𝗲𝗿𝘀𝗶𝗼𝗻: we map human corrective actions back into the latent noise space of the frozen generative policy. These latent targets train a lightweight controller that adapts the robot while preserving the original model's capabilities. Across simulation and real robots, FlowDAgger: 📈 Learns from only 5–20 human intervention episodes 🏆 Outperforms supervised fine-tuning and latent-space reinforcement learning 🤖 Works across VLAs, diffusion policies, and world-action models ✔️ Provides reliable improvements without modifying the pretrained policy We believe this offers a practical path toward making robot foundation models improve during deployment, learning from the way humans naturally teach: through corrections. 📄 Paper: 🌐 Project: 💻 Code: This project was led by my amazing colleague Michael Murray with help from Daphne Chen, Simran Bagaria, Dean Fortier, Tess Hellebrekers, Harshavardhan Reddy Gajarla, Galen Mullins and Andrey Kolobov at Microsoft Research and Maya Cakmak at University of Washington

Oier Mees

13,276 views • 2 months ago

Introducing VL-JEPA: Vision-Language Joint Embedding Predictive Architecture for streaming, live action recognition, retrieval, VQA, and classification tasks with better performance and higher efficiency than large VLMs. • VL-JEPA is the first non-generative model that can perform general-domain vision-language tasks in real-time, built on a joint embedding predictive architecture. • We demonstrate in controlled experiments that VL-JEPA, trained with latent space embedding prediction, outperforms VLMs that rely on data space token prediction. • We show that VL-JEPA delivers significant efficiency gains over VLMs for online video streaming applications, thanks to its non-autoregressive design and native support for selective decoding. • We highlight that our VL-JEPA model, with an unified model architecture, can effectively handle a wide range of classification, retrieval, and VQA tasks at the same time. by Delong Chen (陈德龙) Mustafa Shukor Théo Moutakanni Willy Jade Lei Yu Tejaswi Kasarla Allen Bolourchi Yann LeCun Pascale Fung

Introducing VL-JEPA: Vision-Language Joint Embedding Predictive Architecture for streaming, live action recognition, retrieval, VQA, and classification tasks with better performance and higher efficiency than large VLMs. • VL-JEPA is the first non-generative model that can perform general-domain vision-language tasks in real-time, built on a joint embedding predictive architecture. • We demonstrate in controlled experiments that VL-JEPA, trained with latent space embedding prediction, outperforms VLMs that rely on data space token prediction. • We show that VL-JEPA delivers significant efficiency gains over VLMs for online video streaming applications, thanks to its non-autoregressive design and native support for selective decoding. • We highlight that our VL-JEPA model, with an unified model architecture, can effectively handle a wide range of classification, retrieval, and VQA tasks at the same time. by Delong Chen (陈德龙) Mustafa Shukor Théo Moutakanni Willy Jade Lei Yu Tejaswi Kasarla Allen Bolourchi Yann LeCun Pascale Fung

Pascale Fung

90,144 views • 9 months ago

🦿Xpeng showed a humanoid robot called IRON whose movement looked so human that the team literally cut it open on stage to prove it is a machine. IRON uses a bionic body with a flexible spine, synthetic muscles, and soft skin so joints and torso can twist smoothly like a person. The system has 82 degrees of freedom in total with 22 in each hand for fine finger control. Compute runs on 3 custom AI chips rated at 2,250 TOPS (Tera Operations Per Second), which is far above typical laptop neural accelerators, so it can handle vision and motion planning on the robot. The AI stack focuses on turning camera input directly into body movement without routing through text, which reduces lag and makes the gait look natural. Xpeng staged the cut-open demo at AI Day in Guangzhou this week, addressing rumors that a performer was inside by exposing internal actuators, wiring, and cooling. Company materials also mention a large physical-world model and a multi-brain control setup for dialogue, perception, and locomotion, hinting at a path from stage demos to service work. Production is targeted for 2026, so near-term tasks will be limited, but the hardware shows a serious step toward human-scale manipulation.

🦿Xpeng showed a humanoid robot called IRON whose movement looked so human that the team literally cut it open on stage to prove it is a machine. IRON uses a bionic body with a flexible spine, synthetic muscles, and soft skin so joints and torso can twist smoothly like a person. The system has 82 degrees of freedom in total with 22 in each hand for fine finger control. Compute runs on 3 custom AI chips rated at 2,250 TOPS (Tera Operations Per Second), which is far above typical laptop neural accelerators, so it can handle vision and motion planning on the robot. The AI stack focuses on turning camera input directly into body movement without routing through text, which reduces lag and makes the gait look natural. Xpeng staged the cut-open demo at AI Day in Guangzhou this week, addressing rumors that a performer was inside by exposing internal actuators, wiring, and cooling. Company materials also mention a large physical-world model and a multi-brain control setup for dialogue, perception, and locomotion, hinting at a path from stage demos to service work. Production is targeted for 2026, so near-term tasks will be limited, but the hardware shows a serious step toward human-scale manipulation.

Rohan Paul

3,802,543 views • 10 months ago

[Most robots react. This one thinks a step ahead.] Ant Group's Robbyant just published LingBot-VA 2.0 — a video-action foundation model built from scratch for robot control, not fine-tuned from a video generator. The usual approach takes a video generator made for content creation and bolts a robot policy onto it. LingBot-VA 2.0 argues that's the wrong starting point, and pretrains the whole causal stack natively instead. What stands out: → Foresight Reasoning — the robot predicts the next action chunk while executing the current one, then overwrites the imagined frame with the real observation. Prediction and execution stop waiting on each other. → 927 ms → 142 ms per chunk, across four cumulative optimizations. That lifts asynchronous control from 35 Hz to 225 Hz — a 6.5× speedup. → One shared latent space. A semantic visual-action tokenizer puts world states and actions in the same coordinates, so unlabeled web video carries action-relevant signal. → Sparse MoE video stream — 128 experts, top-8 routing. Roughly 2.5B of ~15.3B parameters fire per token. → Few-shot by design — adapts from 10–15 demonstrations, and a human demo video can replace the text instruction entirely. Full breakdown: Paper: Project Page: Robbyant Ant Group

[Most robots react. This one thinks a step ahead.] Ant Group's Robbyant just published LingBot-VA 2.0 — a video-action foundation model built from scratch for robot control, not fine-tuned from a video generator. The usual approach takes a video generator made for content creation and bolts a robot policy onto it. LingBot-VA 2.0 argues that's the wrong starting point, and pretrains the whole causal stack natively instead. What stands out: → Foresight Reasoning — the robot predicts the next action chunk while executing the current one, then overwrites the imagined frame with the real observation. Prediction and execution stop waiting on each other. → 927 ms → 142 ms per chunk, across four cumulative optimizations. That lifts asynchronous control from 35 Hz to 225 Hz — a 6.5× speedup. → One shared latent space. A semantic visual-action tokenizer puts world states and actions in the same coordinates, so unlabeled web video carries action-relevant signal. → Sparse MoE video stream — 128 experts, top-8 routing. Roughly 2.5B of ~15.3B parameters fire per token. → Few-shot by design — adapts from 10–15 demonstrations, and a human demo video can replace the text instruction entirely. Full breakdown: Paper: Project Page: Robbyant Ant Group

Marktechpost AI

196,499 views • 2 months ago

The Hidden Language of Diffusion Models paper page: tackle the challenge of understanding concept representations in text-to-image models by decomposing an input text prompt into a small set of interpretable elements. This is achieved by learning a pseudo-token that is a sparse weighted combination of tokens from the model's vocabulary, with the objective of reconstructing the images generated for the given concept. Applied over the state-of-the-art Stable Diffusion model, this decomposition reveals non-trivial and surprising structures in the representations of concepts. For example, we find that some concepts such as "a president" or "a composer" are dominated by specific instances (e.g., "Obama", "Biden") and their interpolations. Other concepts, such as "happiness" combine associated terms that can be concrete ("family", "laughter") or abstract ("friendship", "emotion"). In addition to peering into the inner workings of Stable Diffusion, our method also enables applications such as single-image decomposition to tokens, bias detection and mitigation, and semantic image manipulation

The Hidden Language of Diffusion Models paper page: tackle the challenge of understanding concept representations in text-to-image models by decomposing an input text prompt into a small set of interpretable elements. This is achieved by learning a pseudo-token that is a sparse weighted combination of tokens from the model's vocabulary, with the objective of reconstructing the images generated for the given concept. Applied over the state-of-the-art Stable Diffusion model, this decomposition reveals non-trivial and surprising structures in the representations of concepts. For example, we find that some concepts such as "a president" or "a composer" are dominated by specific instances (e.g., "Obama", "Biden") and their interpolations. Other concepts, such as "happiness" combine associated terms that can be concrete ("family", "laughter") or abstract ("friendship", "emotion"). In addition to peering into the inner workings of Stable Diffusion, our method also enables applications such as single-image decomposition to tokens, bias detection and mitigation, and semantic image manipulation

AK

41,830 views • 3 years ago

Kling 2.6 Motion Control is absolutely insane 🤯 Take any reference video and transfer the exact motion onto an AI character: full-body sync, facial expressions, hand gestures, everything. All with just a few clicks. Perfect for e-comm brands and agencies creating AI video ads that don't look like AI. Here's the problem: AI-generated video ads still look robotic. The movements are stiff, the expressions are flat. Your audience clocks it as AI instantly and keeps scrolling. Kling 2.6 Motion Control fixes it: → Start with any reference clip (stock footage, existing UGC, motion reference) → Upload to Kling → Map the exact movement onto any AI character → Full-body motion, hand gestures, facial expressions—all transferred → Generate up to 30 seconds of video No stiff AI movements, no uncanny valley, no instant "skip this ad" reaction. What this unlocks: - Use one winning UGC motion → swap in different AI creators - Pull reference clips from anywhere → generate branded variations - Create dynamic AI video ads with real human movement - Test multiple "creators" without filming anyone new I recorded a quick walkthrough showing how to do this step-by-step. Want access? > Comment "KLING" > Like this post And I'll send it over (must be following so I can DM)

Kling 2.6 Motion Control is absolutely insane 🤯 Take any reference video and transfer the exact motion onto an AI character: full-body sync, facial expressions, hand gestures, everything. All with just a few clicks. Perfect for e-comm brands and agencies creating AI video ads that don't look like AI. Here's the problem: AI-generated video ads still look robotic. The movements are stiff, the expressions are flat. Your audience clocks it as AI instantly and keeps scrolling. Kling 2.6 Motion Control fixes it: → Start with any reference clip (stock footage, existing UGC, motion reference) → Upload to Kling → Map the exact movement onto any AI character → Full-body motion, hand gestures, facial expressions—all transferred → Generate up to 30 seconds of video No stiff AI movements, no uncanny valley, no instant "skip this ad" reaction. What this unlocks: - Use one winning UGC motion → swap in different AI creators - Pull reference clips from anywhere → generate branded variations - Create dynamic AI video ads with real human movement - Test multiple "creators" without filming anyone new I recorded a quick walkthrough showing how to do this step-by-step. Want access? > Comment "KLING" > Like this post And I'll send it over (must be following so I can DM)

Mike Futia

25,247 views • 7 months ago

You pretrained a robot policy on millions of camera frames. Now you want to add a force sensor. Do you really have to retrain on everything from scratch? MuSe adds a force-torque sensor to a frozen vision-only policy using a tiny amount of contact data. - lifts contact-rich task success (peg insertion 60% to 87%, vase wiping 33% to 77%) - fuse the new sensor both early (shared token space) and late (cross-attention), train the policy as a world model that predicts future video, future force, and actions together, and replay old vision-only data with the force input masked to prevent forgetting. - adding new sensor modalities to pre-trained World Model can be cheap & improve performance significantly

You pretrained a robot policy on millions of camera frames. Now you want to add a force sensor. Do you really have to retrain on everything from scratch? MuSe adds a force-torque sensor to a frozen vision-only policy using a tiny amount of contact data. - lifts contact-rich task success (peg insertion 60% to 87%, vase wiping 33% to 77%) - fuse the new sensor both early (shared token space) and late (cross-attention), train the policy as a world model that predicts future video, future force, and actions together, and replay old vision-only data with the force input masked to prevent forgetting. - adding new sensor modalities to pre-trained World Model can be cheap & improve performance significantly

Vai Viswanathan

36,059 views • 2 months ago

🛠️ What if a robot could invent its own tools. And teach itself how to use them? That’s exactly what VLMgineer does: a new framework that lets Vision Language Models (VLMs) design physical tools and the actions to use them, entirely on their own. No templates. No human demonstrations. Just raw, AI-driven creativity. Why it matters ✅ Co-designs tools and actions together using VLMs, ensuring tight coupling between form and function ✅ Uses VLM-guided evolution (not random search) to refine designs intelligently ✅ Outperforms human-designed tools by +64.7% in task success across 12 RoboToolBench challenges ✅ Produces better-than-everyday tools for real manipulation tasks—measured in success rate and elegance It builds on the emerging trend of large-model-guided evolutionary design (like Eureka and AlphaEvolve) and brings it into physical robotics. It opens the door to general-purpose, automated hardware design, no strong priors needed. Code & paper: —- Weekly robotics and AI insights. Subscribe free:

🛠️ What if a robot could invent its own tools. And teach itself how to use them? That’s exactly what VLMgineer does: a new framework that lets Vision Language Models (VLMs) design physical tools and the actions to use them, entirely on their own. No templates. No human demonstrations. Just raw, AI-driven creativity. Why it matters ✅ Co-designs tools and actions together using VLMs, ensuring tight coupling between form and function ✅ Uses VLM-guided evolution (not random search) to refine designs intelligently ✅ Outperforms human-designed tools by +64.7% in task success across 12 RoboToolBench challenges ✅ Produces better-than-everyday tools for real manipulation tasks—measured in success rate and elegance It builds on the emerging trend of large-model-guided evolutionary design (like Eureka and AlphaEvolve) and brings it into physical robotics. It opens the door to general-purpose, automated hardware design, no strong priors needed. Code & paper: —- Weekly robotics and AI insights. Subscribe free:

Ilir Aliu

14,294 views • 9 months ago

Today may be the ImageNet moment for robotics. RT-X: the largest open-source robot dataset ever compiled, across 33 institutes, 22 robot hardware, 527 skills, and 1M episodes. Why is robotics lagging so far behind NLP, vision, and other AI domains? Data scarcity is the main culprit to blame, among other difficulties. Unlike text, images, and videos, you cannot download mass amounts of onboard robot control data from the internet. They simply don't exist in the wild. 11 yrs ago, ImageNet kicked off the deep learning revolution. 3-4 yrs ago, internet-scale data fueled the first GPTs and Diffusions that define this era of foundation models. I think 2023 is finally the year for robotics to scale up. Robot foundation models like VIMA ( my team's work at NVIDIA) and RT-1/2 ( Google DeepMind's effort) are extremely data hungry. While massively parallel simulations like NVIDIA IsaacGym & Omniverse can alleviate the problem to some extent, it's still not quite enough to bridge the gap to the messy, physical world. This new dataset is not just a technical contribution. I also see it as a commendable effort to overcome institutional bureaucracies and unite researchers from around the world to tackle a grand challenge together. Robotics will be the final holy grail that we capture in AI. We are not there yet, but ascending in the right gradient direction. RT-X website: Launch blog:

Today may be the ImageNet moment for robotics. RT-X: the largest open-source robot dataset ever compiled, across 33 institutes, 22 robot hardware, 527 skills, and 1M episodes. Why is robotics lagging so far behind NLP, vision, and other AI domains? Data scarcity is the main culprit to blame, among other difficulties. Unlike text, images, and videos, you cannot download mass amounts of onboard robot control data from the internet. They simply don't exist in the wild. 11 yrs ago, ImageNet kicked off the deep learning revolution. 3-4 yrs ago, internet-scale data fueled the first GPTs and Diffusions that define this era of foundation models. I think 2023 is finally the year for robotics to scale up. Robot foundation models like VIMA ( my team's work at NVIDIA) and RT-1/2 ( Google DeepMind's effort) are extremely data hungry. While massively parallel simulations like NVIDIA IsaacGym & Omniverse can alleviate the problem to some extent, it's still not quite enough to bridge the gap to the messy, physical world. This new dataset is not just a technical contribution. I also see it as a commendable effort to overcome institutional bureaucracies and unite researchers from around the world to tackle a grand challenge together. Robotics will be the final holy grail that we capture in AI. We are not there yet, but ascending in the right gradient direction. RT-X website: Launch blog:

Jim Fan

265,061 views • 3 years ago

Robot Utility Models (RUMs) enable basic tasks – door opening, drawer opening, object reorientation, etc. – at ~90% accuracy without ANY finetuning (i.e. zero-shot) in unseen new environments. Fully open source!!! models, data, code & hw. We think this is super exciting, why?👇 1. Unlocks many practical home utility tasks that often involve these basic tasks as part of an action chain. “Go get me a fork” involves opening the kitchen door and then opening the cutlery drawer. 2. This works well **zero-shot in unseen and new** environments, which is practically a huge deal. Turn the robot on, and get going. 3. The recipe for building a new model is fairly generic, and we think with a bit more refinement this can be a general recipe to build many more Utility models. More details and access 👇

Robot Utility Models (RUMs) enable basic tasks – door opening, drawer opening, object reorientation, etc. – at ~90% accuracy without ANY finetuning (i.e. zero-shot) in unseen new environments. Fully open source!!! models, data, code & hw. We think this is super exciting, why?👇 1. Unlocks many practical home utility tasks that often involve these basic tasks as part of an action chain. “Go get me a fork” involves opening the kitchen door and then opening the cutlery drawer. 2. This works well zero-shot in unseen and new environments, which is practically a huge deal. Turn the robot on, and get going. 3. The recipe for building a new model is fairly generic, and we think with a bit more refinement this can be a general recipe to build many more Utility models. More details and access 👇

Mahi Shafiullah 🏠🤖

89,535 views • 2 years ago

South Korea's local governments are deploying around 7,000 AI-robot dolls to seniors and dementia patients. The $1,800 robot doll by Hyodal can hold full conversations to tackle loneliness and remind users to take medication. Dystopian, yes, but the data is fascinating: 1. Studies (with over 9,000 users) found that depression levels reduced from 5.73 to 3.14, and medicine intake improved from 2.69 to 2.87. 2. The doll comes with a companion app and web monitoring platform for caretakers to monitor remotely. 3. Safety features are installed to alert when no movement has been detected for a certain period, essentially always watching the user. 4. The doll also offers touch interaction, 24-hour voice reminders, check-ins, voice messages, a health coach, quizzes, exercise, music, and more. 5. Caregivers have access to the app, allowing them to send/receive voice messages, make group announcements, and monitor motion detection. I'd definitely have some privacy and data collection concerns here before handing this off to my family, but the product actually seems really cool. Will be interesting to watch the data to see if this idea has legs. Keep in mind, SK has a rapidly aging population and one of the world's lowest birth rates, so it makes sense for the local governments to be early adopters here.

South Korea's local governments are deploying around 7,000 AI-robot dolls to seniors and dementia patients. The $1,800 robot doll by Hyodal can hold full conversations to tackle loneliness and remind users to take medication. Dystopian, yes, but the data is fascinating: 1. Studies (with over 9,000 users) found that depression levels reduced from 5.73 to 3.14, and medicine intake improved from 2.69 to 2.87. 2. The doll comes with a companion app and web monitoring platform for caretakers to monitor remotely. 3. Safety features are installed to alert when no movement has been detected for a certain period, essentially always watching the user. 4. The doll also offers touch interaction, 24-hour voice reminders, check-ins, voice messages, a health coach, quizzes, exercise, music, and more. 5. Caregivers have access to the app, allowing them to send/receive voice messages, make group announcements, and monitor motion detection. I'd definitely have some privacy and data collection concerns here before handing this off to my family, but the product actually seems really cool. Will be interesting to watch the data to see if this idea has legs. Keep in mind, SK has a rapidly aging population and one of the world's lowest birth rates, so it makes sense for the local governments to be early adopters here.

Rowan Cheung

420,862 views • 2 years ago

Created with seadance 2.5 When the city falls, she becomes the weapon. A cinematic transformation from human to colossal combat machine, built for one final stand against the monsters. Prompt: 15-second ultra-photorealistic cinematic sci-fi action sequence. A powerful young woman stands alone in the middle of a destroyed futuristic city at night. Broken buildings, burning vehicles, smoke, sparks, rain-soaked streets and massive alien monsters surrounding her. She wears a sleek futuristic black tactical suit. The entire scene feels physically real, gritty and cinematic. 0–3 sec — Transformation Begins: Extreme close-up of her eyes as a mysterious blue-white energy suddenly activates inside her body. Energy pulses across her skin like glowing circuits. Strong wind surrounds her, dust and debris lifting from the ground. She looks directly at the approaching monsters with a determined expression. 3–6 sec — Full Transformation: The energy explodes outward. Massive mechanical armor rapidly forms around her body piece by piece, with realistic metallic components locking together, pistons moving, cables connecting and glowing energy cores powering up. Her body transforms seamlessly into a gigantic 20-meter-tall humanoid combat robot. The transformation must feel physically believable, detailed and mechanical—not magical or cartoonish. 6–9 sec — Power Activation: Low-angle shot of the completed giant robot standing above the ruined city. Its energy core ignites, eyes glow intensely, steam and smoke release from mechanical joints. It clenches its enormous metal fists and takes one powerful step forward, causing the wet ground to crack and debris to jump. 9–12 sec — First Attack: Three enormous alien monsters charge toward the robot. The robot launches forward with surprising speed, grabs one monster and violently throws it through several abandoned vehicles. Another monster attacks from behind; the robot instantly turns and blocks the strike with its armored arm. 12–15 sec — Final Action Shot: Dynamic tracking shot as the giant robot punches the largest monster with massive force, sending it crashing through a damaged building. A huge shockwave pushes smoke and debris outward. The robot stands in the middle of the destruction as rain falls over its glowing armor, then looks toward the remaining monsters. Ultra-photorealistic CGI/VFX, realistic metallic materials, physically accurate reflections, detailed mechanical engineering, realistic weight and momentum, natural motion blur, realistic debris physics, cinematic explosions, volumetric smoke, rain interaction, dramatic lighting, high-detail textures, dynamic camera movement, low-angle hero shots, handheld impact moments, seamless transformation, consistent character design throughout, epic Hollywood sci-fi action movie quality, 4K, no cartoon look, no anime, no text, no watermark.

Created with seadance 2.5 When the city falls, she becomes the weapon. A cinematic transformation from human to colossal combat machine, built for one final stand against the monsters. Prompt: 15-second ultra-photorealistic cinematic sci-fi action sequence. A powerful young woman stands alone in the middle of a destroyed futuristic city at night. Broken buildings, burning vehicles, smoke, sparks, rain-soaked streets and massive alien monsters surrounding her. She wears a sleek futuristic black tactical suit. The entire scene feels physically real, gritty and cinematic. 0–3 sec — Transformation Begins: Extreme close-up of her eyes as a mysterious blue-white energy suddenly activates inside her body. Energy pulses across her skin like glowing circuits. Strong wind surrounds her, dust and debris lifting from the ground. She looks directly at the approaching monsters with a determined expression. 3–6 sec — Full Transformation: The energy explodes outward. Massive mechanical armor rapidly forms around her body piece by piece, with realistic metallic components locking together, pistons moving, cables connecting and glowing energy cores powering up. Her body transforms seamlessly into a gigantic 20-meter-tall humanoid combat robot. The transformation must feel physically believable, detailed and mechanical—not magical or cartoonish. 6–9 sec — Power Activation: Low-angle shot of the completed giant robot standing above the ruined city. Its energy core ignites, eyes glow intensely, steam and smoke release from mechanical joints. It clenches its enormous metal fists and takes one powerful step forward, causing the wet ground to crack and debris to jump. 9–12 sec — First Attack: Three enormous alien monsters charge toward the robot. The robot launches forward with surprising speed, grabs one monster and violently throws it through several abandoned vehicles. Another monster attacks from behind; the robot instantly turns and blocks the strike with its armored arm. 12–15 sec — Final Action Shot: Dynamic tracking shot as the giant robot punches the largest monster with massive force, sending it crashing through a damaged building. A huge shockwave pushes smoke and debris outward. The robot stands in the middle of the destruction as rain falls over its glowing armor, then looks toward the remaining monsters. Ultra-photorealistic CGI/VFX, realistic metallic materials, physically accurate reflections, detailed mechanical engineering, realistic weight and momentum, natural motion blur, realistic debris physics, cinematic explosions, volumetric smoke, rain interaction, dramatic lighting, high-detail textures, dynamic camera movement, low-angle hero shots, handheld impact moments, seamless transformation, consistent character design throughout, epic Hollywood sci-fi action movie quality, 4K, no cartoon look, no anime, no text, no watermark.

Ruzaina

12,113 views • 18 days ago