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.

These days, it feels like a new robotic hand is developed every week. But how can we give them all vision-based general grasping ability? Meet AnyDexGrasp: it enables robust grasping across different robot hands—with just 40 training objects and 4–8 hours of trial and error!

Hao-Shu Fang

1,719 subscribers

25,409 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

The full video of hand and glove co-design in action. For hand 🖐️ we try to make sure it's able to repeat what a human hand can do; hyperextension in the index finger is a great example here. Most robotic hands lack that, and it is very important for grasping objects securely. Let’s try to close the embodiment gap. For glove 🧤 we try to make sure it’s comfortable to wear, while also has same kinematics, contacts with the robot hand.

The full video of hand and glove co-design in action. For hand 🖐️ we try to make sure it's able to repeat what a human hand can do; hyperextension in the index finger is a great example here. Most robotic hands lack that, and it is very important for grasping objects securely. Let’s try to close the embodiment gap. For glove 🧤 we try to make sure it’s comfortable to wear, while also has same kinematics, contacts with the robot hand.

Quanting Xie

20,989 views • 5 months ago

A robot hand grasp over 500 totally new objects without fail? Zero-shot, single-view & super reliable ⬇️ + Paper Grasping random objects is hard for robots, especially when shapes, weights, and materials vary. RobustDexGrasp solves this with a smart new way of seeing and controlling the hand, leading to near-perfect grasping, even in noisy or cluttered scenes. Thank you for sharing, Hui Zhang 🙏 Follow him!! What makes it special ✅ Grabs 500+ unseen objects with 94.6% success using only single-view input ✅ Learns local shapes, not full geometry, for better generalization ✅ Trained with just 35 objects in sim but works in the real world with hundreds more ✅ Adapts to noise, unexpected forces, and even plays chess with VLM planning It shows that smart sensing and adaptive control can take dexterous grasping to the next level. Project: Paper:

A robot hand grasp over 500 totally new objects without fail? Zero-shot, single-view & super reliable ⬇️ + Paper Grasping random objects is hard for robots, especially when shapes, weights, and materials vary. RobustDexGrasp solves this with a smart new way of seeing and controlling the hand, leading to near-perfect grasping, even in noisy or cluttered scenes. Thank you for sharing, Hui Zhang 🙏 Follow him!! What makes it special ✅ Grabs 500+ unseen objects with 94.6% success using only single-view input ✅ Learns local shapes, not full geometry, for better generalization ✅ Trained with just 35 objects in sim but works in the real world with hundreds more ✅ Adapts to noise, unexpected forces, and even plays chess with VLM planning It shows that smart sensing and adaptive control can take dexterous grasping to the next level. Project: Paper:

Ilir Aliu

37,980 views • 1 year ago

A policy that teaches robot hands to touch things the way humans do... not just grab and move, but feel and adjust in real time. Robot manipulation research often stops at picking up objects and placing them. CGP goes further: it handles tasks like opening jars, flipping objects in-hand, wiping dishes, and grasping fragile eggs, the kind of dexterous, contact-rich skills that require constant micro-adjustments based on what the fingers are actually feeling. The robot doesn't just see what it's doing; it predicts what contact should feel like at each step, then checks whether reality matches the prediction. If a finger is slipping, the policy knows before the object drops. Works on real robot hands (both 4-finger and 5-finger designs) with tactile sensors embedded in the fingertips Robust to visual distractions! The robot keeps flipping a box correctly even when the camera view is disrupted, because it's grounding decisions in touch, not just vision. Baseline policies without contact grounding fail in predictable ways: slipping mid-task, incomplete motions, loss of grasp, CGP avoids these This is a meaningful step toward robots that can handle the physical world with the kind of reliable, adaptive grip that humans take for granted. Relevant for manufacturing, logistics, assistive robotics, and anywhere fragile or irregular objects need to be handled carefully. Published at RSS 2026, developed with Meta Reality Labs Research. Thanks for sharing, Zhengtong Xu / Zhengtong Xu ——- Weekly robotics and AI insights. Subscribe free:

A policy that teaches robot hands to touch things the way humans do... not just grab and move, but feel and adjust in real time. Robot manipulation research often stops at picking up objects and placing them. CGP goes further: it handles tasks like opening jars, flipping objects in-hand, wiping dishes, and grasping fragile eggs, the kind of dexterous, contact-rich skills that require constant micro-adjustments based on what the fingers are actually feeling. The robot doesn't just see what it's doing; it predicts what contact should feel like at each step, then checks whether reality matches the prediction. If a finger is slipping, the policy knows before the object drops. Works on real robot hands (both 4-finger and 5-finger designs) with tactile sensors embedded in the fingertips Robust to visual distractions! The robot keeps flipping a box correctly even when the camera view is disrupted, because it's grounding decisions in touch, not just vision. Baseline policies without contact grounding fail in predictable ways: slipping mid-task, incomplete motions, loss of grasp, CGP avoids these This is a meaningful step toward robots that can handle the physical world with the kind of reliable, adaptive grip that humans take for granted. Relevant for manufacturing, logistics, assistive robotics, and anywhere fragile or irregular objects need to be handled carefully. Published at RSS 2026, developed with Meta Reality Labs Research. Thanks for sharing, Zhengtong Xu / Zhengtong Xu ——- Weekly robotics and AI insights. Subscribe free:

Ilir Aliu

12,769 views • 3 months ago

Xynova just launched Flex 2, its second-gen hybrid dexterous hand — and Xiaomi is already on the cap table. 🤖✋ The spec sheet is sharp: 23 DOF, 400g palm, ±0.1mm repeatability, 0.05N force control, 12kg single-hand grasp load, and multimodal sensing for adaptive grasping and slip detection. Xynova is not only building a robotic hand. It is building the manipulation stack around it: micro linear actuators, joint modules, sensing, control, and developer-facing integration. The Xiaomi link is confirmed. Xiaomi Strategic Investment joined Xynova’s angel round and continued adding in the Pre-A round. What is not confirmed: whether Xiaomi’s own robot hand uses Xynova’s technology. That detail still has no public proof. But the direction is clear: Xiaomi is putting money into the component layer that decides whether humanoids can actually handle objects, tools, and daily tasks. This is the part of robotics that rarely looks flashy on stage. But it decides what a robot can really do with its hands.

Xynova just launched Flex 2, its second-gen hybrid dexterous hand — and Xiaomi is already on the cap table. 🤖✋ The spec sheet is sharp: 23 DOF, 400g palm, ±0.1mm repeatability, 0.05N force control, 12kg single-hand grasp load, and multimodal sensing for adaptive grasping and slip detection. Xynova is not only building a robotic hand. It is building the manipulation stack around it: micro linear actuators, joint modules, sensing, control, and developer-facing integration. The Xiaomi link is confirmed. Xiaomi Strategic Investment joined Xynova’s angel round and continued adding in the Pre-A round. What is not confirmed: whether Xiaomi’s own robot hand uses Xynova’s technology. That detail still has no public proof. But the direction is clear: Xiaomi is putting money into the component layer that decides whether humanoids can actually handle objects, tools, and daily tasks. This is the part of robotics that rarely looks flashy on stage. But it decides what a robot can really do with its hands.

RoboHub🤖

71,176 views • 3 months ago

Gemini + π0 = actually useful robots! (Similar to what Physical Intelligence did with "Hi Robot") I can now verbally tell the robot that I'm building a red Lego wall or wooden tower, and it will infer the next steps by itself and pass me the necessary pieces, tools, or materials, ha! You can also just ask it to bring you things! The pipeline works as follows: - OpenAI Whisper (local) → speech to text - Gemini → makes sense of user requests, converts to robot tasks, bounding boxes, grasping points, etc. (System 2 thinking FTW!) - π0 → robotic actions The π0 was finetuned just for pick-and-place Lego bricks only, and it generalizes beautifully to all kinds of tasks. However, there's lots of room for improvement when it comes to grasping & accuracy. Things that could help: - Conditioning on grasping points - Better data collection (I'm not that great at teleop) - Lots more synthetic data and simulations

Gemini + π0 = actually useful robots! (Similar to what Physical Intelligence did with "Hi Robot") I can now verbally tell the robot that I'm building a red Lego wall or wooden tower, and it will infer the next steps by itself and pass me the necessary pieces, tools, or materials, ha! You can also just ask it to bring you things! The pipeline works as follows: - OpenAI Whisper (local) → speech to text - Gemini → makes sense of user requests, converts to robot tasks, bounding boxes, grasping points, etc. (System 2 thinking FTW!) - π0 → robotic actions The π0 was finetuned just for pick-and-place Lego bricks only, and it generalizes beautifully to all kinds of tasks. However, there's lots of room for improvement when it comes to grasping & accuracy. Things that could help: - Conditioning on grasping points - Better data collection (I'm not that great at teleop) - Lots more synthetic data and simulations

Shreyas Gite

25,485 views • 1 year ago

Trump can whine all day long about how “hard it is” to sit down and shut up a couple of hours a day, a few days a week. I don’t give a fuck.

Trump can whine all day long about how “hard it is” to sit down and shut up a couple of hours a day, a few days a week. I don’t give a fuck.

Jo

78,680 views • 2 years ago

It is really hard to get robot policies that are both precise (small margins for error) and general (robust to env variations). We just released ViSk, where skin sensing is used to train fine-grained policies with ~1 hour of data. Below is a single-take video.

It is really hard to get robot policies that are both precise (small margins for error) and general (robust to env variations). We just released ViSk, where skin sensing is used to train fine-grained policies with ~1 hour of data. Below is a single-take video.

Lerrel Pinto

17,169 views • 1 year ago

Open-Source Robotic Hand That Costs Just $300. California-based company TetherIA.ai has officially launched its open-source robotic hand, the Aero Hand Open. Designed to be the most capable at the lowest cost, this hand is a new tool for embodied AI research. In the clip, it demonstrates impressive dexterity by grasping an M5 screw, picking up an iPhone, and opening a soda can. Key Features: ► Price: $314.00 ► Open Source: CAD, electronics, and firmware are all open source. ► Actuation: 7 Active Degrees of Freedom (16 joints) with a tendon-driven design. ► Weight: Under 400 grams. ► Purpose: Built for embodied AI research to make dexterous manipulation accessible to all. For more details and to purchase:

Open-Source Robotic Hand That Costs Just $300. California-based company TetherIA.ai has officially launched its open-source robotic hand, the Aero Hand Open. Designed to be the most capable at the lowest cost, this hand is a new tool for embodied AI research. In the clip, it demonstrates impressive dexterity by grasping an M5 screw, picking up an iPhone, and opening a soda can. Key Features: ► Price: $314.00 ► Open Source: CAD, electronics, and firmware are all open source. ► Actuation: 7 Active Degrees of Freedom (16 joints) with a tendon-driven design. ► Weight: Under 400 grams. ► Purpose: Built for embodied AI research to make dexterous manipulation accessible to all. For more details and to purchase:

RoboHub🤖

45,830 views • 11 months ago

How can we build robotic hands with truly superhuman dexterity? Daxo Robotics is developing a unique tendon-driven soft robot hand, which aims to be tougher and more capable than a traditional humanoid hand. Each finger consists of many different tendons, which act in concert to move or manipulate. This is a special episode of RoboPapers where, instead of talking about a scientific paper, we talk to Tom Zhang, founder of Daxo Robotics, in order to learn both about his background and about how this one-of-a-kind robot hand design works. Watch Episode #40 of RoboPapers with Michael Cho - Rbt/Acc and Chris Paxton now!

How can we build robotic hands with truly superhuman dexterity? Daxo Robotics is developing a unique tendon-driven soft robot hand, which aims to be tougher and more capable than a traditional humanoid hand. Each finger consists of many different tendons, which act in concert to move or manipulate. This is a special episode of RoboPapers where, instead of talking about a scientific paper, we talk to Tom Zhang, founder of Daxo Robotics, in order to learn both about his background and about how this one-of-a-kind robot hand design works. Watch Episode #40 of RoboPapers with Michael Cho - Rbt/Acc and Chris Paxton now!

RoboPapers

16,500 views • 10 months ago

Westworld-like hand Suzhou-bassed Rochu Robotics just tested a biomimetic hand that is completely different from the rigid motor-driven dexterous hands on the market: Pneumatic/Fluid-driven artificial muscles + 1:1 human skeletal structure. These artificial muscles contract when pressurized,just like biological ones --pulling on biomimetic tendons to move the joints with smooth, natural motion. Just 4 biomimetic muscles can easily lift a 3kg kettlebell, and demonstrate the versatility of flexible grasping…while also acing fast finger challenges (peace sign, rock on, counting, etc.) almost matching a real human hand. True compliance and safety in human-robot interaction,no more cold, stiff mechanics. This one is developing real life-like movement logic.

Westworld-like hand Suzhou-bassed Rochu Robotics just tested a biomimetic hand that is completely different from the rigid motor-driven dexterous hands on the market: Pneumatic/Fluid-driven artificial muscles + 1:1 human skeletal structure. These artificial muscles contract when pressurized,just like biological ones --pulling on biomimetic tendons to move the joints with smooth, natural motion. Just 4 biomimetic muscles can easily lift a 3kg kettlebell, and demonstrate the versatility of flexible grasping…while also acing fast finger challenges (peace sign, rock on, counting, etc.) almost matching a real human hand. True compliance and safety in human-robot interaction,no more cold, stiff mechanics. This one is developing real life-like movement logic.

CyberRobo

28,240 views • 2 months ago

It's a 3D printer, and 3D assembly station! 🖨️ The Functgraph developed at Meiji University starts as a regular 3D printer but upgrades itself into a mini factory. It can print parts for its own tools, pick them up, clean them, and put them together, all by itself. Think of it like a robot that can 3D print a spatula, assemble it, and then use it to flip pancakes. 🥞 Instead of just printing objects, the Functgraph can actually perform physical tasks, like folding laundry or slicing vegetables, by using printed and assembled tools. It’s a step toward the idea of a robot that can download "apps" as physical skills, much like your phone downloads software! 👀

It's a 3D printer, and 3D assembly station! 🖨️ The Functgraph developed at Meiji University starts as a regular 3D printer but upgrades itself into a mini factory. It can print parts for its own tools, pick them up, clean them, and put them together, all by itself. Think of it like a robot that can 3D print a spatula, assemble it, and then use it to flip pancakes. 🥞 Instead of just printing objects, the Functgraph can actually perform physical tasks, like folding laundry or slicing vegetables, by using printed and assembled tools. It’s a step toward the idea of a robot that can download "apps" as physical skills, much like your phone downloads software! 👀

Lukas Ziegler

149,956 views • 1 year ago

This looks like another step toward removing the data bottleneck in robotics. The bottleneck exists because recording demonstrations is easy, but labeling every grasp, hold, move, and release is slow and expensive. Perceptron Egocentric dropped a new robotics video annotation system. You give it raw robot-camera or first-person/egocentric video. It gives back structured labels that look more like training data: where each small manipulation starts and ends, what the subtask is, what each hand is doing, where the hands are frame by frame, and which hand is left or right. So, Input: raw video of a person or robot doing a task. Output: a machine-readable breakdown of the physical actions inside that video. Basically, it is a machine that converts robot video into policy-training supervision. That means timestamps, subtask boundaries, per-hand actions, and left-right hand grounding. So this will turn large piles of robot or egocentric video into supervision for training robot policies.

This looks like another step toward removing the data bottleneck in robotics. The bottleneck exists because recording demonstrations is easy, but labeling every grasp, hold, move, and release is slow and expensive. Perceptron Egocentric dropped a new robotics video annotation system. You give it raw robot-camera or first-person/egocentric video. It gives back structured labels that look more like training data: where each small manipulation starts and ends, what the subtask is, what each hand is doing, where the hands are frame by frame, and which hand is left or right. So, Input: raw video of a person or robot doing a task. Output: a machine-readable breakdown of the physical actions inside that video. Basically, it is a machine that converts robot video into policy-training supervision. That means timestamps, subtask boundaries, per-hand actions, and left-right hand grounding. So this will turn large piles of robot or egocentric video into supervision for training robot policies.

Rohan Paul

18,903 views • 1 month ago

Today, we publicly released RoboCasa365, a large-scale simulation benchmark for training and systematically evaluating generalist robot models. Built upon our original RoboCasa framework, it offers: • 2,500 realistic kitchen environments; • 365 everyday tasks (basic skills + long-horizon mobile manipulation); • Over 3,200 objects with many articulated fixtures/appliances. All are designed for fully controlled, reproducible benchmarking of robotic policies. Progress in robotic foundation models is real. But it’s still hard to answer basic questions like: How close are we to general-purpose autonomy? What factors drive generalization? What are the model/data scaling curves like? Real-world eval is slow and noisy, and existing sims (like LIBERO, which we built 3 years ago) often lack sufficient task and scene diversity. This benchmark comes with 2,200+ hours of demonstrations and 500K+ trajectories to support studies of multi-task training, pretraining, and continual learning at scale. Check it out at

Today, we publicly released RoboCasa365, a large-scale simulation benchmark for training and systematically evaluating generalist robot models. Built upon our original RoboCasa framework, it offers: • 2,500 realistic kitchen environments; • 365 everyday tasks (basic skills + long-horizon mobile manipulation); • Over 3,200 objects with many articulated fixtures/appliances. All are designed for fully controlled, reproducible benchmarking of robotic policies. Progress in robotic foundation models is real. But it’s still hard to answer basic questions like: How close are we to general-purpose autonomy? What factors drive generalization? What are the model/data scaling curves like? Real-world eval is slow and noisy, and existing sims (like LIBERO, which we built 3 years ago) often lack sufficient task and scene diversity. This benchmark comes with 2,200+ hours of demonstrations and 500K+ trajectories to support studies of multi-task training, pretraining, and continual learning at scale. Check it out at

Yuke Zhu

24,027 views • 6 months ago

There once was a time that you didn’t yet realize you had hands. It’s a funny thought, but 100% true. In fact it’s not until somewhere between 2-4 months of age that babies fully “discover” their hands. Stranger still, thanks in part to the Palmar (or grasping) reflex, infants actually use their hands (involuntarily) well before they realize they exist. Once the discovery is made, scenes like this one - shared to IG by epicpsychology - become common. Your baby may stare at their hands in wonder for prolonged periods, soaking in not only their presence but their ability to control their movement (clumsy though these earliest attempts may be). This discovery marks the dawn of a whole new set of usages. What began as involuntary and haphazard, becomes increasingly deliberate and controlled - as your baby begins purposefully grasping and reaching for objects in their field of vision. You can support the development of these skills by providing safe toys and objects within arm’s length for them to reach for, grasp, and manipulate. Just don’t forget that almost any such object will be headed straight toward your baby’s mouth (as a primary means of exploration), so be particularly mindful of both their cleanliness and potential choking hazards during this exploratory phase.

There once was a time that you didn’t yet realize you had hands. It’s a funny thought, but 100% true. In fact it’s not until somewhere between 2-4 months of age that babies fully “discover” their hands. Stranger still, thanks in part to the Palmar (or grasping) reflex, infants actually use their hands (involuntarily) well before they realize they exist. Once the discovery is made, scenes like this one - shared to IG by epicpsychology - become common. Your baby may stare at their hands in wonder for prolonged periods, soaking in not only their presence but their ability to control their movement (clumsy though these earliest attempts may be). This discovery marks the dawn of a whole new set of usages. What began as involuntary and haphazard, becomes increasingly deliberate and controlled - as your baby begins purposefully grasping and reaching for objects in their field of vision. You can support the development of these skills by providing safe toys and objects within arm’s length for them to reach for, grasp, and manipulate. Just don’t forget that almost any such object will be headed straight toward your baby’s mouth (as a primary means of exploration), so be particularly mindful of both their cleanliness and potential choking hazards during this exploratory phase.

Dan Wuori

86,855 views • 1 year ago

In my experience, robot 'generalists' are often jacks of all trades but masters of none. In training across multiple tasks and environments, robot policies fail to generalize robustly and effectively to each particular test setting. What if at test time, we non-parametrically *retrieved* “relevant” data from the training set and used it to significantly improve the performance of few-shot imitation learning to be robust to various test time scenes. Notably, we are *not* collecting lots of new data, just training more on sub-components of the same training data! Now, we’re certainly not the first to suggest retrieval, but in our new work - STRAP, we show how retrieving relevant *sub-trajectories* from offline datasets can significantly increase data reuse across tasks, when paired with an appropriate metric space. A 🧵 (1/7)

In my experience, robot 'generalists' are often jacks of all trades but masters of none. In training across multiple tasks and environments, robot policies fail to generalize robustly and effectively to each particular test setting. What if at test time, we non-parametrically retrieved “relevant” data from the training set and used it to significantly improve the performance of few-shot imitation learning to be robust to various test time scenes. Notably, we are not collecting lots of new data, just training more on sub-components of the same training data! Now, we’re certainly not the first to suggest retrieval, but in our new work - STRAP, we show how retrieving relevant sub-trajectories from offline datasets can significantly increase data reuse across tasks, when paired with an appropriate metric space. A 🧵 (1/7)

Abhishek Gupta

12,045 views • 1 year ago

Most of today's AI can see the world, but it doesn’t **feel** it. Capturing the sense of touch is crucial for dexterous robotic manipulation, user modeling, and understanding physical interactions. Introducing OpenTouch: bringing full-hand tactile sensing into real-world AI🖐️ OpenTouch is collected in-the-wild using tactile sensing gloves, hand pose tracking gloves, and egocentric glasses. It includes: • 5 hours of real-world data, • 3 hours densely annotated contact-rich interactions, • 2,900 curated interaction clips, • across 800 objects, 14 environments, and 29 grasp types. all open at:

Most of today's AI can see the world, but it doesn’t feel it. Capturing the sense of touch is crucial for dexterous robotic manipulation, user modeling, and understanding physical interactions. Introducing OpenTouch: bringing full-hand tactile sensing into real-world AI🖐️ OpenTouch is collected in-the-wild using tactile sensing gloves, hand pose tracking gloves, and egocentric glasses. It includes: • 5 hours of real-world data, • 3 hours densely annotated contact-rich interactions, • 2,900 curated interaction clips, • across 800 objects, 14 environments, and 29 grasp types. all open at:

Paul Liang

47,158 views • 4 months ago