Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

Sharing some exciting DYNA-1 result: zero-shot environment generalization We put DYNA-1 under test in a completely different environment from our training distribution – with an entirely different background (Dyna Robotics banner) and metal table. The table has a reflective and smooth surface, creating a wildly different visual appearance as... show more

Jason Ma

20,654 subscribers

49,008 görüntüleme • 1 yıl önce •via X (Twitter)

Bilim & Teknoloji Eğitim

Anya Rossi• Live Now

Private livecam show

10 Yorum

Jason Ma profil fotoğrafı

Jason Ma1 yıl önce

Original thread:

MrRobotics profil fotoğrafı

MrRobotics1 yıl önce

@DynaRobotics why is it 10X and doesn't look impressive as the first demo?

Hiro Protagonist profil fotoğrafı

Hiro Protagonist1 yıl önce

@DynaRobotics 10x speed so it's still basically useless for real life applications. Robotics amateur hour.

pfung profil fotoğrafı

pfung1 yıl önce

@DynaRobotics nice robustness!

Ming Qin profil fotoğrafı

Ming Qin1 yıl önce

@DynaRobotics The table can even reflect the robot itself, and it doesn’t seem surprised at all

atharva profil fotoğrafı

atharva1 yıl önce

@DynaRobotics insane

Max von Wolff profil fotoğrafı

Max von Wolff1 yıl önce

@DynaRobotics Impressive!

Humanoid Pulse profil fotoğrafı

Humanoid Pulse1 yıl önce

@DynaRobotics wow- there is something sacinating in that - can watch for hours🤓

Peter Christie profil fotoğrafı

Peter Christie1 yıl önce

@DynaRobotics Nope….

Michael Cho - Rbt/Acc profil fotoğrafı

Michael Cho - Rbt/Acc1 yıl önce

@DynaRobotics Impressive stuff!

Benzer Videolar

Introducing Dynamism v1 (DYNA-1) by Dyna Robotics – the first robot foundation model built for round-the-clock, high-throughput dexterous autonomy. Here is a time-lapse video of our model autonomously folding 850+ napkins in a span of 24 hours with • 99.4% success rate — zero human intervention • 60% human throughput speed • 4.3/5 quality ratings (set by the client) A thread on our motivation, insights and results:

Introducing Dynamism v1 (DYNA-1) by Dyna Robotics – the first robot foundation model built for round-the-clock, high-throughput dexterous autonomy. Here is a time-lapse video of our model autonomously folding 850+ napkins in a span of 24 hours with • 99.4% success rate — zero human intervention • 60% human throughput speed • 4.3/5 quality ratings (set by the client) A thread on our motivation, insights and results:

Jason Ma

518,131 görüntüleme • 1 yıl önce

If insanity is defined as doing the same thing over and over, expecting a different result, we have to stop being insane in our approach to fixing the environment and try something new.

If insanity is defined as doing the same thing over and over, expecting a different result, we have to stop being insane in our approach to fixing the environment and try something new.

Bill Maher

718,578 görüntüleme • 3 yıl önce

🎃 Halloween is coming. Our hardworking team is lining up for sweet treats, of course, served by Dynasaur! DYNA VLA model now has robust agentic reasoning capability, allowing it to serve arbitrary combinations and counts of candies! Pure imitation learning can’t work given the combinatorially many possibilities. No video edits. Uninterrupted, real-life, as always 🤖 Happy Halloween from DYNA!🍬

🎃 Halloween is coming. Our hardworking team is lining up for sweet treats, of course, served by Dynasaur! DYNA VLA model now has robust agentic reasoning capability, allowing it to serve arbitrary combinations and counts of candies! Pure imitation learning can’t work given the combinatorially many possibilities. No video edits. Uninterrupted, real-life, as always 🤖 Happy Halloween from DYNA!🍬

Dyna Robotics

25,000 görüntüleme • 7 ay önce

We are really impressed with Dyna Robotics robot folding demo at Foxglove's Actuate conference. We've come a long way since the early PR2 folding research!

We are really impressed with Dyna Robotics robot folding demo at Foxglove's Actuate conference. We've come a long way since the early PR2 folding research!

Open Robotics

29,394 görüntüleme • 8 ay önce

We use our warmup to MOVE with our 3D Movement Prep system. This means training the brain as well as the body with a variety of patterns and positions in different planes.

We use our warmup to MOVE with our 3D Movement Prep system. This means training the brain as well as the body with a variety of patterns and positions in different planes.

Zach Dechant

46,705 görüntüleme • 1 yıl önce

🚨BREAKTHROUGH ALERT: Robotics hits its "GPT moment" with Dyna Robotics. One unified model mastering language, vision, and control. This could redefine the future of robotics—here’s how:

🚨BREAKTHROUGH ALERT: Robotics hits its "GPT moment" with Dyna Robotics. One unified model mastering language, vision, and control. This could redefine the future of robotics—here’s how:

Tulsi Soni

41,717 görüntüleme • 1 yıl önce

A beautiful day with my daughter Rafif in the neighborhoods of Al-Mawasi, Khan Younis. We try, as much as we can, to hold on to moments of joy and to keep going, waiting for our simple right to life, safety, and stability. A different environment from the harsh reality we live in here. Rafif is a child of this world, and her right—and the right of all children in Gaza—must return one day.

A beautiful day with my daughter Rafif in the neighborhoods of Al-Mawasi, Khan Younis. We try, as much as we can, to hold on to moments of joy and to keep going, waiting for our simple right to life, safety, and stability. A different environment from the harsh reality we live in here. Rafif is a child of this world, and her right—and the right of all children in Gaza—must return one day.

Mohammed jawad 🇵🇸

13,636 görüntüleme • 4 ay önce

The whole point of decentralization is to have the infrastructure as diversified as possible to ensure maximum security for the project. As we were starting out and were still small, we had to test multiple different providers, including smaller ones, to find the best options for the project based on three different criteria: price, security level, and scalability. As we moved forward and got introduced to different connections, we transitioned to enterprise-level providers. We are still in talks with many of them and actively searching for more, as this is a never-ending process. Below is a video from one of our smaller data centers (one of the contractors we used earlier). This showcases how small and diversified our infrastructure is and also must be in the future, as this is the key to decentralization.

The whole point of decentralization is to have the infrastructure as diversified as possible to ensure maximum security for the project. As we were starting out and were still small, we had to test multiple different providers, including smaller ones, to find the best options for the project based on three different criteria: price, security level, and scalability. As we moved forward and got introduced to different connections, we transitioned to enterprise-level providers. We are still in talks with many of them and actively searching for more, as this is a never-ending process. Below is a video from one of our smaller data centers (one of the contractors we used earlier). This showcases how small and diversified our infrastructure is and also must be in the future, as this is the key to decentralization.

OpSec

357,420 görüntüleme • 2 yıl önce

When training ACT-1, we treated data from diverse, long-horizon tasks in the wild as a first-class citizen. This makes generalization the default, not an exception. The capability envelope expands. More to come.

When training ACT-1, we treated data from diverse, long-horizon tasks in the wild as a first-class citizen. This makes generalization the default, not an exception. The capability envelope expands. More to come.

Alper Canberk

390,138 görüntüleme • 6 ay önce

We pushed MolmoAct 2 into a completely out-of-distribution studio setting to test its robustness under extreme environmental changes. Despite only 10 minutes of task-specific fine-tuning data collected elsewhere, it adapted surprisingly well to the new environment.

We pushed MolmoAct 2 into a completely out-of-distribution studio setting to test its robustness under extreme environmental changes. Despite only 10 minutes of task-specific fine-tuning data collected elsewhere, it adapted surprisingly well to the new environment.

Jiafei Duan

11,381 görüntüleme • 1 ay önce

We just did World’s first on-stage autonomous demo of long-horizon dexterous VLA 🚨 No training. No setup. Performance out of the box. Live demo is hard and unpredictable, but we felt great about our model’s generalization, and it went pretty well! 💯 Zero-shot. 100% success.

We just did World’s first on-stage autonomous demo of long-horizon dexterous VLA 🚨 No training. No setup. Performance out of the box. Live demo is hard and unpredictable, but we felt great about our model’s generalization, and it went pretty well! 💯 Zero-shot. 100% success.

Jason Ma

63,454 görüntüleme • 8 ay önce

🚨Regarding the incident with the smashing of the Jesus statue, we wanted to clarify something important: We published a video showing a statue being destroyed by Muslims in Lebanon. The video is from December 2025, in which Muslims smashed a statue of Jesus because Christmas songs were playing from the church. In the photo from the last few days, there is a soldier whose case is currently under heavy investigation and has drawn strong condemnations. These are two entirely separate incidents - different locations and different times. After we posted the video, many people mistakenly assumed it was the same statue from the soldier’s photo. Some even began using the video as if it were a continuation of that recent story, which is not true. As stated, the video actually shows an incident from five months ago in which Muslims destroyed a statue of Jesus. At the end of the video, we added a screenshot from a Lebanese news channel from five months ago as proof that the video depicts a different event - not the viral incident from the past few days.

🚨Regarding the incident with the smashing of the Jesus statue, we wanted to clarify something important: We published a video showing a statue being destroyed by Muslims in Lebanon. The video is from December 2025, in which Muslims smashed a statue of Jesus because Christmas songs were playing from the church. In the photo from the last few days, there is a soldier whose case is currently under heavy investigation and has drawn strong condemnations. These are two entirely separate incidents - different locations and different times. After we posted the video, many people mistakenly assumed it was the same statue from the soldier’s photo. Some even began using the video as if it were a continuation of that recent story, which is not true. As stated, the video actually shows an incident from five months ago in which Muslims destroyed a statue of Jesus. At the end of the video, we added a screenshot from a Lebanese news channel from five months ago as proof that the video depicts a different event - not the viral incident from the past few days.

GAZAWOOD - the PALLYWOOD saga

42,003 görüntüleme • 1 ay önce

Here’s How 1 McDonald’s Hamburger Contains 1300 Strands Of Different DNA. This Is The Result Of Going From Small Farms To Mega Corporations “We're pooling the bacteria from a thousand different animals” “Thousands of animals” contribute “to a single hamburger.” “This is a combo bin of meat. The meat in these boxes comes from many different animals and many different places. Some is even from overseas. All of it is ground together to make the juicy hamburger Americans love. Hamburger used to be the scraps left over from butchering just one cow.” But a significant change has occurred. Now parts of many cows are blended together by the ton. ‌ If we take the meat from one animal, grind it up and make ground beef from just one animal, we're including only the bacteria from one animal. But if we take the meat from a thousand different animals and grind that together, we're pooling the bacteria from a thousand different animals as well. So explain this to me, how many different cows would be in one burger? Well, I suspect that there are hundreds or even thousands of animals that have contributed to a single hamburger.”

Here’s How 1 McDonald’s Hamburger Contains 1300 Strands Of Different DNA. This Is The Result Of Going From Small Farms To Mega Corporations “We're pooling the bacteria from a thousand different animals” “Thousands of animals” contribute “to a single hamburger.” “This is a combo bin of meat. The meat in these boxes comes from many different animals and many different places. Some is even from overseas. All of it is ground together to make the juicy hamburger Americans love. Hamburger used to be the scraps left over from butchering just one cow.” But a significant change has occurred. Now parts of many cows are blended together by the ton. ‌ If we take the meat from one animal, grind it up and make ground beef from just one animal, we're including only the bacteria from one animal. But if we take the meat from a thousand different animals and grind that together, we're pooling the bacteria from a thousand different animals as well. So explain this to me, how many different cows would be in one burger? Well, I suspect that there are hundreds or even thousands of animals that have contributed to a single hamburger.”

Wall Street Apes

3,114,000 görüntüleme • 1 yıl önce

A question I kept returning to was how soon robots will be ubiquitous in our daily lives. Sergey uses self-driving as an analogy, but explains why robotics may not follow the same path because the cost of mistakes is very different. Self-driving took over a decade and is still in its nascency, in part because even rare mistakes are unacceptable. Robots may deploy earlier, and improve faster, because some mistakes are tolerable. Breaking a dish is different from causing an accident.

A question I kept returning to was how soon robots will be ubiquitous in our daily lives. Sergey uses self-driving as an analogy, but explains why robotics may not follow the same path because the cost of mistakes is very different. Self-driving took over a decade and is still in its nascency, in part because even rare mistakes are unacceptable. Robots may deploy earlier, and improve faster, because some mistakes are tolerable. Breaking a dish is different from causing an accident.

Patrick OShaughnessy

57,780 görüntüleme • 2 ay önce

q. RM said that the members have different lifestyles now and have different boundaries. maybe you can expand on that a little bit? What does that mean? 🐰: well, all of the members come from very different backgrounds, and even since we met, everyone is moving at their own speed, taking things at their own tempo, on their own timeline! 🐰: beyond that we like different kinds of music and prefer different kinds of food, too! the seven of us are so different yet we work together, sing together and perform together and when we collaborate, of course, we’re different there as well, that’s probably why he said that!

q. RM said that the members have different lifestyles now and have different boundaries. maybe you can expand on that a little bit? What does that mean? 🐰: well, all of the members come from very different backgrounds, and even since we met, everyone is moving at their own speed, taking things at their own tempo, on their own timeline! 🐰: beyond that we like different kinds of music and prefer different kinds of food, too! the seven of us are so different yet we work together, sing together and perform together and when we collaborate, of course, we’re different there as well, that’s probably why he said that!

ajk ⋆

154,832 görüntüleme • 1 ay önce

Curious whether video generation models (like #SORA) qualify as world models? We conduct a systematic study to answer this question by investigating whether a video gen model is able to learn physical laws. Three are three key messages to take home: 1⃣The model generalises perfectly for in-distribution data, but fails to do out-of-distribution generalization. For combinatorial scenarios, scaling law is observed. 2⃣The models fail to abstract general rules and instead tries to mimic the closest training example. 3⃣The model prioritizes different attributes when referencing training data: color > size > velocity > shape. This work is a joint effort with our outstanding intern Yang Yue. Paper: Webpage:

Curious whether video generation models (like #SORA) qualify as world models? We conduct a systematic study to answer this question by investigating whether a video gen model is able to learn physical laws. Three are three key messages to take home: 1⃣The model generalises perfectly for in-distribution data, but fails to do out-of-distribution generalization. For combinatorial scenarios, scaling law is observed. 2⃣The models fail to abstract general rules and instead tries to mimic the closest training example. 3⃣The model prioritizes different attributes when referencing training data: color > size > velocity > shape. This work is a joint effort with our outstanding intern Yang Yue. Paper: Webpage:

Bingyi Kang

606,519 görüntüleme • 1 yıl önce

Physics feels stable, predictable, and well-behaved largely because we learned it in 3D. That comfort hides a trap. In 1907, Paul Ehrenfest pointed out something unsettling. If you take the laws we treat as fundamental and transplant them into a different number of spatial dimensions, they often stop working the way we expect. Not just numerically different but qualitatively different. The issue isn’t the force law by itself but it's geometry. Gauss’s law ties inverse-square forces to the surface area of spheres, and sphere geometry depends on dimension. Change the dimension, and the same-looking force produces a different potential, a different balance of attraction and inertia, and a different fate for motion.

Physics feels stable, predictable, and well-behaved largely because we learned it in 3D. That comfort hides a trap. In 1907, Paul Ehrenfest pointed out something unsettling. If you take the laws we treat as fundamental and transplant them into a different number of spatial dimensions, they often stop working the way we expect. Not just numerically different but qualitatively different. The issue isn’t the force law by itself but it's geometry. Gauss’s law ties inverse-square forces to the surface area of spheres, and sphere geometry depends on dimension. Change the dimension, and the same-looking force produces a different potential, a different balance of attraction and inertia, and a different fate for motion.

Mathelirium

30,393 görüntüleme • 2 ay önce

#WATCH | Greater Noida, UP | Philosopher and Author Acharya Prashant says, "...As long as the principle of success remains the same as it is now, we will not be able to save the environment. All our success comes at the cost of the environment...Unless the human heart is purified, we will not be able to clean the environment..."

#WATCH | Greater Noida, UP | Philosopher and Author Acharya Prashant says, "...As long as the principle of success remains the same as it is now, we will not be able to save the environment. All our success comes at the cost of the environment...Unless the human heart is purified, we will not be able to clean the environment..."

ANI

22,763 görüntüleme • 1 yıl önce

"There's a different vibe in here, a different culture being brought into the White Sox camp..." Greg and DeRo are coming to you from Glendale as we turn our attention to the White Sox! 📺

"There's a different vibe in here, a different culture being brought into the White Sox camp..." Greg and DeRo are coming to you from Glendale as we turn our attention to the White Sox! 📺

MLB Network

55,569 görüntüleme • 1 yıl önce