Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

We trained a robotic foundation model that can drive mobile robots in six different countries, and navigate Sproul Plaza in midday on the UC Berkeley campus! Some cool new work w/ noriaki_hirose, Lydia Ignatova, Kyle Stachowicz, Catherine Glossop, Dhruv Shah

Sergey Levine

131,721 subscribers

52,979 Aufrufe • vor 1 Jahr •via X (Twitter)

Wissenschaft & Technologie Bildung

Anya Rossi• Live Now

Private livecam show

5 Kommentare

Profilbild von Michael Cho - Rbt/Acc

Michael Cho - Rbt/Accvor 1 Jahr

@NoriakiHirose @KyleStachowicz @CatGlossop @shahdhruv_ Thanks for giving meaning to what we do @frodobots . Means a lot to us that our work help enable research like this; onwards!

Profilbild von Winnie

Winnievor 1 Jahr

@NoriakiHirose @KyleStachowicz @CatGlossop @shahdhruv_ Impressive! It took a lot of courage to write a PDF on failed methods.

Profilbild von paras

parasvor 1 Jahr

@NoriakiHirose @KyleStachowicz @CatGlossop @shahdhruv_ Awesome work!!

Profilbild von Felipe Chávez

Felipe Chávezvor 1 Jahr

@NoriakiHirose @KyleStachowicz @CatGlossop @shahdhruv_ Promising !

Profilbild von zhao

zhaovor 1 Jahr

@NoriakiHirose @KyleStachowicz @CatGlossop @shahdhruv_ dm

Ähnliche Videos

We can learn a model that provides shaped "process rewards" for robotic RL, that evolves automatically as the policy gets better. This improves performance on benchmarks, and works in the real world! Some fun new work with Raymond Tsao & Andrew Wagenmaker

We can learn a model that provides shaped "process rewards" for robotic RL, that evolves automatically as the policy gets better. This improves performance on benchmarks, and works in the real world! Some fun new work with Raymond Tsao & Andrew Wagenmaker

Sergey Levine

37,474 Aufrufe • vor 2 Monaten

Many of you asked for code & weights for π₀, we are happy to announce that we are releasing π₀ and pre-trained checkpoints in our new openpi repository! We tested the model on a few public robots, and we include code for you to fine-tune it yourself.

Many of you asked for code & weights for π₀, we are happy to announce that we are releasing π₀ and pre-trained checkpoints in our new openpi repository! We tested the model on a few public robots, and we include code for you to fine-tune it yourself.

Physical Intelligence

441,382 Aufrufe • vor 1 Jahr

In just ~3 months, as a solo founder with no prior robotics experience, General Trajectory trained a foundation model for dexterous manipulation that lets humanoid robots pick up unseen objects and perform real-world work. It generalizes to novel objects and scenes, including cases where prior SoTA models achieve 0% success. Congrats on the launch Joshua!

In just ~3 months, as a solo founder with no prior robotics experience, General Trajectory trained a foundation model for dexterous manipulation that lets humanoid robots pick up unseen objects and perform real-world work. It generalizes to novel objects and scenes, including cases where prior SoTA models achieve 0% success. Congrats on the launch Joshua!

Y Combinator

68,639 Aufrufe • vor 7 Monaten

This is NASA’s Astrobee, a cube-shaped robot that flies freely inside the ISS to help astronauts with daily tasks. It can move using built-in fans, navigate with vision sensors, and work either on its own or be controlled from Earth. Some versions have robotic arms that can detect and grab objects, and this tech could be used in the future to remove space debris or service satellites.

This is NASA’s Astrobee, a cube-shaped robot that flies freely inside the ISS to help astronauts with daily tasks. It can move using built-in fans, navigate with vision sensors, and work either on its own or be controlled from Earth. Some versions have robotic arms that can detect and grab objects, and this tech could be used in the future to remove space debris or service satellites.

Space and Technology

51,832 Aufrufe • vor 4 Monaten

Our vision is for AI that uses world models to adapt in new and dynamic environments and efficiently learn new skills. We’re sharing V-JEPA 2, a new world model with state-of-the-art performance in visual understanding and prediction. V-JEPA 2 is a 1.2 billion-parameter model, trained on video, that can enable zero-shot planning in robots—allowing them to plan and execute tasks in unfamiliar environments. Learn more about V-JEPA 2 ➡️ As we continue working toward our goal of achieving advanced machine intelligence (AMI), we’re also releasing three new benchmarks for evaluating how well existing models can reason about the physical world from video. Learn more and download the new benchmarks ➡️

Our vision is for AI that uses world models to adapt in new and dynamic environments and efficiently learn new skills. We’re sharing V-JEPA 2, a new world model with state-of-the-art performance in visual understanding and prediction. V-JEPA 2 is a 1.2 billion-parameter model, trained on video, that can enable zero-shot planning in robots—allowing them to plan and execute tasks in unfamiliar environments. Learn more about V-JEPA 2 ➡️ As we continue working toward our goal of achieving advanced machine intelligence (AMI), we’re also releasing three new benchmarks for evaluating how well existing models can reason about the physical world from video. Learn more and download the new benchmarks ➡️

AI at Meta

310,435 Aufrufe • vor 1 Jahr

Elon Musk in new interview on how Tesla’s Optimus robot will learn more quickly: "What we're going to need to do is build a lot of robots and put them in a kind of like a "Optimus Academy" so they can do self-play in reality. We're actually building that out. We're gonna have at least 10,000 Optimus robots, maybe 20,000 or 30,000, that are doing self-play and testing different tasks. Tesla has quite a good reality generator. A physics-accurate reality generator that we made for the cars. We'll do the same things for the robots, and actually have done that for the robots. You can do millions of simulated robots in a simulated world, and you use the tens of thousands in the real world to close the simulation to reality gap.”

Elon Musk in new interview on how Tesla’s Optimus robot will learn more quickly: "What we're going to need to do is build a lot of robots and put them in a kind of like a "Optimus Academy" so they can do self-play in reality. We're actually building that out. We're gonna have at least 10,000 Optimus robots, maybe 20,000 or 30,000, that are doing self-play and testing different tasks. Tesla has quite a good reality generator. A physics-accurate reality generator that we made for the cars. We'll do the same things for the robots, and actually have done that for the robots. You can do millions of simulated robots in a simulated world, and you use the tens of thousands in the real world to close the simulation to reality gap.”

Sawyer Merritt

250,956 Aufrufe • vor 6 Monaten

you can now control things with your brain. literally. we're building the most wearable BCI on the planet, with , backed by Khosla Ventures Accel Initialized Capital & Kevin Weil 🇺🇸. we collected the world’s largest neural dataset and trained the most capable Brain Foundation Model. then we invented a new class of biosensors powered by custom ASICs. type without typing. click without clicking. a cap that lets your brain do the work. we’re sabi.

you can now control things with your brain. literally. we're building the most wearable BCI on the planet, with , backed by Khosla Ventures Accel Initialized Capital & Kevin Weil 🇺🇸. we collected the world’s largest neural dataset and trained the most capable Brain Foundation Model. then we invented a new class of biosensors powered by custom ASICs. type without typing. click without clicking. a cap that lets your brain do the work. we’re sabi.

Rahul Chhabra

6,095,787 Aufrufe • vor 4 Monaten

"Everyone keeps talking about reindustrializing the country, but nobody's talking about the reality that nobody wants to work off some random freeway exit in the middle of Texas." Jan Sramek 🇺🇲 🌁 ⛰️ (Founder & CEO California Forever) on how he’s thinking about building a new kind of manufacturing hub, one that people actually want to work in. Here’s what he told us: "You drive an hour and a half to a factory next to a Burger King and a McDonald’s. Who wants to work there?" "If we're serious about getting people to work in manufacturing again, we have to offer the quality of life our generation expects – where you can bike to work, grab lunch with friends in an outdoor plaza, and hang out after." "There’s basically nowhere in America you can do that, except here."

"Everyone keeps talking about reindustrializing the country, but nobody's talking about the reality that nobody wants to work off some random freeway exit in the middle of Texas." Jan Sramek 🇺🇲 🌁 ⛰️ (Founder & CEO California Forever) on how he’s thinking about building a new kind of manufacturing hub, one that people actually want to work in. Here’s what he told us: "You drive an hour and a half to a factory next to a Burger King and a McDonald’s. Who wants to work there?" "If we're serious about getting people to work in manufacturing again, we have to offer the quality of life our generation expects – where you can bike to work, grab lunch with friends in an outdoor plaza, and hang out after." "There’s basically nowhere in America you can do that, except here."

TBPN

76,968 Aufrufe • vor 1 Jahr

Robot trained entirely in simulation! 🤯 Sudo AI just released #sudo R1. This could be the moment the data bottleneck in robotics finally breaks, a major step forward for getting simulation-trained robots to actually work in the real world. #sudo R1 is a manipulation foundation model that reliably picks objects it has never seen: glass, fabric, metal, irregular shapes. All with zero real-world training data. No tele-op. No demos. No manual labeling. Pure simulation. And picking isn't just any skill, it's the foundation. Almost everything we want robots to do starts with being able to reliably pick something up. Data collection is the bottleneck holding the whole field back. Every new task means hours of tele-op and armies of labelers. If simulation alone can actually deliver production-grade picking, that bottleneck starts to disappear. Really exciting to see! If the results hold, the path to general-purpose robots just got a whole lot shorter. 👇 Read more here: ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Robot trained entirely in simulation! 🤯 Sudo AI just released #sudo R1. This could be the moment the data bottleneck in robotics finally breaks, a major step forward for getting simulation-trained robots to actually work in the real world. #sudo R1 is a manipulation foundation model that reliably picks objects it has never seen: glass, fabric, metal, irregular shapes. All with zero real-world training data. No tele-op. No demos. No manual labeling. Pure simulation. And picking isn't just any skill, it's the foundation. Almost everything we want robots to do starts with being able to reliably pick something up. Data collection is the bottleneck holding the whole field back. Every new task means hours of tele-op and armies of labelers. If simulation alone can actually deliver production-grade picking, that bottleneck starts to disappear. Really exciting to see! If the results hold, the path to general-purpose robots just got a whole lot shorter. 👇 Read more here: ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Lukas Ziegler

25,901 Aufrufe • vor 4 Monaten

We are releasing 4M-21 with a permissive license, including its source code and trained models. It's a pretty effective multimodal model that solves 10s of tasks & modalities. See the demo code, sample results, and the tokenizers of diverse modalities on the website. IMO, the multitask learning aspect of multimodal models has really taken a step forward. We can train a single model on many diverse tasks with ~SOTA accuracy. But a long way to go in terms of transfer/emergence. 🌐 ⌨️ Joint work w/ EPFL Apple.

We are releasing 4M-21 with a permissive license, including its source code and trained models. It's a pretty effective multimodal model that solves 10s of tasks & modalities. See the demo code, sample results, and the tokenizers of diverse modalities on the website. IMO, the multitask learning aspect of multimodal models has really taken a step forward. We can train a single model on many diverse tasks with ~SOTA accuracy. But a long way to go in terms of transfer/emergence. 🌐 ⌨️ Joint work w/ EPFL Apple.

Amir Zamir

69,564 Aufrufe • vor 2 Jahren

Been experimenting with Apple's new Core AI framework over the past few days Built a mobile app that runs local AI models directly on device, exploring Core AI model deployment (.aimodel), model specialization, caching, and inference across Apple Silicon Also spent some time integrating with Apple's Foundation Models framework and understanding how custom models can coexist with Apple's on-device AI stack Still early, but the combination of Core AI, Foundation Models, and the Apple Neural Engine feels like a pretty compelling foundation for local first AI applications Still in alpha and working towards a more stable version, but it's been a fun way to explore where Apple's on device AI ecosystem is headed !!

Been experimenting with Apple's new Core AI framework over the past few days Built a mobile app that runs local AI models directly on device, exploring Core AI model deployment (.aimodel), model specialization, caching, and inference across Apple Silicon Also spent some time integrating with Apple's Foundation Models framework and understanding how custom models can coexist with Apple's on-device AI stack Still early, but the combination of Core AI, Foundation Models, and the Apple Neural Engine feels like a pretty compelling foundation for local first AI applications Still in alpha and working towards a more stable version, but it's been a fun way to explore where Apple's on device AI ecosystem is headed !!

Sagar

13,221 Aufrufe • vor 2 Monaten

On Friday, we presented new, fully-furnished mobile homes to two families who had lost theirs in Hurricane Helene. We had prayer with the families, presented a Bible to them, and dedicated these homes to the glory of God. We’re so thankful that Samaritan's Purse can bring help in Jesus’ Name! Watch some highlights…

On Friday, we presented new, fully-furnished mobile homes to two families who had lost theirs in Hurricane Helene. We had prayer with the families, presented a Bible to them, and dedicated these homes to the glory of God. We’re so thankful that Samaritan's Purse can bring help in Jesus’ Name! Watch some highlights…

Franklin Graham

107,723 Aufrufe • vor 1 Jahr

We are back. After one year of quiet building. Introducing GENE-26.5, our first robotic brain that takes a major step toward human-level capability. For years, robotics has struggled to learn from the world’s largest and valuable data source: Humans. Solving it means rethinking the whole stack from the ground up: - A robotics-native foundation model. - A 1:1 human-like robotic hand. - A noninvasive data collection glove for motion, force, and touch. - A simulator that turns weeks of experiments into minutes. GENE-26.5 is trained across language, vision, proprioception, tactile, and action. We designed a set of tasks to test how far we can go with this new paradigm. Fully autonomous, 1x speed, one model, same weights. (Enjoy with sound on) We are approaching the endgame for robotics. And this is just a beginning.

We are back. After one year of quiet building. Introducing GENE-26.5, our first robotic brain that takes a major step toward human-level capability. For years, robotics has struggled to learn from the world’s largest and valuable data source: Humans. Solving it means rethinking the whole stack from the ground up: - A robotics-native foundation model. - A 1:1 human-like robotic hand. - A noninvasive data collection glove for motion, force, and touch. - A simulator that turns weeks of experiments into minutes. GENE-26.5 is trained across language, vision, proprioception, tactile, and action. We designed a set of tasks to test how far we can go with this new paradigm. Fully autonomous, 1x speed, one model, same weights. (Enjoy with sound on) We are approaching the endgame for robotics. And this is just a beginning.

Genesis AI

2,733,307 Aufrufe • vor 3 Monaten

Running legged robots with AI. Today we met @swiss_mile, the team connecting AI with the physical world using autonomous machines. An ideal mobile robot needs to move quickly, overcoming obstacles while carrying heavy loads—something current robots still struggle with. Many robots haven't been widely adopted due to issues with real-world problem-solving. To address this, the Swiss-mile team have created a solution that combines robots with wheels and legs, powered by advanced AI. These robots can walk, drive, stand, open doors, and even handle packages. By integrating these features, the team are on a mission to push automation forward beyond stationary robotic arms and have raised $22m in Seed Funding from Jeff Bezos, HongShan, Amazon Industrial Innovation Fund and Armada Investment to do so.

Running legged robots with AI. Today we met @swiss_mile, the team connecting AI with the physical world using autonomous machines. An ideal mobile robot needs to move quickly, overcoming obstacles while carrying heavy loads—something current robots still struggle with. Many robots haven't been widely adopted due to issues with real-world problem-solving. To address this, the Swiss-mile team have created a solution that combines robots with wheels and legs, powered by advanced AI. These robots can walk, drive, stand, open doors, and even handle packages. By integrating these features, the team are on a mission to push automation forward beyond stationary robotic arms and have raised $22m in Seed Funding from Jeff Bezos, HongShan, Amazon Industrial Innovation Fund and Armada Investment to do so.

Circuit

17,876 Aufrufe • vor 1 Jahr

I wrote some new code to drive X-Plane directly rather than using the internal FDR playback, this lets me add in the Helicopter. Given the lack of precise data on the helicopter's route I made a rough estimate on speed and altitude. And you can see it's hidden against city lights for a long time. Be aware that the cockpit model is likely to be different and there's no data being sent to the gauges.

I wrote some new code to drive X-Plane directly rather than using the internal FDR playback, this lets me add in the Helicopter. Given the lack of precise data on the helicopter's route I made a rough estimate on speed and altitude. And you can see it's hidden against city lights for a long time. Be aware that the cockpit model is likely to be different and there's no data being sent to the gauges.

Scott Manley

144,051 Aufrufe • vor 1 Jahr

Robots might learn better from video than from language! 📼 Most Vision-Language-Action (VLA) models learn what to do from text, but still struggle with how things move in the real world. That makes them data-hungry and slow to train. mimic video takes a different route. Instead of grounding robot control in text, it grounds it in video, using large pre-trained video models that already capture physical motion and dynamics. The idea is straightforward: let the video model handle “what will happen next,” and let a smaller control model focus only on turning that visual plan into robot actions. The result is big gains in practice. Robots trained this way need 10× less data, converge twice as fast, and perform better on both simulated benchmarks and real bimanual manipulation tasks. If robots can “imagine” motion using video, control becomes a much simpler problem. Shoutout to Jonas Pai, Liam Achenbach, Oier Mees, Elvis Nava and the rest of the team! Here's the project page: ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Robots might learn better from video than from language! 📼 Most Vision-Language-Action (VLA) models learn what to do from text, but still struggle with how things move in the real world. That makes them data-hungry and slow to train. mimic video takes a different route. Instead of grounding robot control in text, it grounds it in video, using large pre-trained video models that already capture physical motion and dynamics. The idea is straightforward: let the video model handle “what will happen next,” and let a smaller control model focus only on turning that visual plan into robot actions. The result is big gains in practice. Robots trained this way need 10× less data, converge twice as fast, and perform better on both simulated benchmarks and real bimanual manipulation tasks. If robots can “imagine” motion using video, control becomes a much simpler problem. Shoutout to Jonas Pai, Liam Achenbach, Oier Mees, Elvis Nava and the rest of the team! Here's the project page: ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Lukas Ziegler

50,090 Aufrufe • vor 8 Monaten

This nice older couple that runs a YouTube channel called “Bronco Mustang Lifestyle” celebrated their 45th anniversary by buying their first EV: a 2026 Tesla Model Y RWD with FSD (Supervised). "We thought this would be a cool car that we can drive around and let it drive us. We are newbys to the EV world and are watching lots of videos to get us up to speed. In a way, it’s like getting a new computer or phone. I was surprised at how big and roomy the Model Y is." The full video of them picking up the car is linked below:

This nice older couple that runs a YouTube channel called “Bronco Mustang Lifestyle” celebrated their 45th anniversary by buying their first EV: a 2026 Tesla Model Y RWD with FSD (Supervised). "We thought this would be a cool car that we can drive around and let it drive us. We are newbys to the EV world and are watching lots of videos to get us up to speed. In a way, it’s like getting a new computer or phone. I was surprised at how big and roomy the Model Y is." The full video of them picking up the car is linked below:

Sawyer Merritt

237,830 Aufrufe • vor 6 Monaten