正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Physical Intelligence co-founder, Sergey Levine: The best synthetic experience now comes from strong models, not hand-coded physics Models can capture finer details than humans, but still need real-world experience "information in, capability out" Whether through simulation or direct data, what matters is what the model can do

Haider.

50,369 subscribers

14,197 次观看 • 9 个月前 •via X (Twitter)

健康养生科学技术教育

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

What is the best data for training humanoid & robotics foundation models? Pete Florence Pete Florence (CEO Eric Nehrlich, ex-Google DeepMind) dropped his live data tier list in this 7-minute clip on TBPN: - S-tier: Real-world robot experience (especially glove/sensor high-dexterity data) - A/B-tier: Internet/YouTube videos. Surprisingly powerful for transfer learning (the “web data” moment for physical AI) - B-tier: Text/common crawl (Reddit, books, etc.). Useful priors, but not enough alone - C-tier: Motion Capture. Great for whole-body motion, weak on finger dexterity - C or lower: Simulation / synthetic / world models. High potential, still waiting for strong real-world proof Generalist has collected 270,000+ hours of real-world manipulation data (scaling ~10k hours/week). And Pete stressed one key point: “The quality of data is incredibly important.” It’s not just about volume. It’s developing intuition for what actually drives performance through hands-on work. As Physical AI scales, curated real-world and high-quality internet video looks like a winning combo. h/t YUJI | AI Robotics

What is the best data for training humanoid & robotics foundation models? Pete Florence Pete Florence (CEO Eric Nehrlich, ex-Google DeepMind) dropped his live data tier list in this 7-minute clip on TBPN: - S-tier: Real-world robot experience (especially glove/sensor high-dexterity data) - A/B-tier: Internet/YouTube videos. Surprisingly powerful for transfer learning (the “web data” moment for physical AI) - B-tier: Text/common crawl (Reddit, books, etc.). Useful priors, but not enough alone - C-tier: Motion Capture. Great for whole-body motion, weak on finger dexterity - C or lower: Simulation / synthetic / world models. High potential, still waiting for strong real-world proof Generalist has collected 270,000+ hours of real-world manipulation data (scaling ~10k hours/week). And Pete stressed one key point: “The quality of data is incredibly important.” It’s not just about volume. It’s developing intuition for what actually drives performance through hands-on work. As Physical AI scales, curated real-world and high-quality internet video looks like a winning combo. h/t YUJI | AI Robotics

Michelle

15,195 次观看 • 28 天前

Nexar’s real-world data just built the best incident prediction model — Meet BADAS 1.0. It beat state-of-the-art models by learning from 10B+ real miles and 60M+ real events, not simulations. 🚀 Discover what it can do for you

Nexar’s real-world data just built the best incident prediction model — Meet BADAS 1.0. It beat state-of-the-art models by learning from 10B+ real miles and 60M+ real events, not simulations. 🚀 Discover what it can do for you

Nexar

709,509 次观看 • 8 个月前

The internet was the training set for intelligence Nobody has built the equivalent for the physical world Previously, Lucas Ngoo co-founded Carousell, scaled it $1B+ Now at CortexAI he's collecting the data robotics labs need to train foundation models Cameras, VR headsets, glasses on factory workers, retail workers, everyday people. Recording real-world manipulation work Not building the robot. Not training the model. Collecting what goes in 2026: scale data 2027+: start rolling robots out Initially 90% teleoperation, 10% autonomy While humans step in, the system keeps learning That's the flywheel toward full autonomy 🎙️ Lucas Ngoo, Co-Founder & CEO, CortexAI on Fondo.com Start (full ep in comments)

The internet was the training set for intelligence Nobody has built the equivalent for the physical world Previously, Lucas Ngoo co-founded Carousell, scaled it $1B+ Now at CortexAI he's collecting the data robotics labs need to train foundation models Cameras, VR headsets, glasses on factory workers, retail workers, everyday people. Recording real-world manipulation work Not building the robot. Not training the model. Collecting what goes in 2026: scale data 2027+: start rolling robots out Initially 90% teleoperation, 10% autonomy While humans step in, the system keeps learning That's the flywheel toward full autonomy 🎙️ Lucas Ngoo, Co-Founder & CEO, CortexAI on Fondo.com Start (full ep in comments)

David J Phillips

110,268 次观看 • 3 个月前

better real-world physics are coming to video models thanks to synthetic video data

better real-world physics are coming to video models thanks to synthetic video data

Dreaming Tulpa 🥓👑

16,135 次观看 • 1 年前

Demis Hassabis on the limit in today’s AI: language can describe the world, but it cannot contain it - and why "World Models" are his "longest standing passion". Language models absorbed far more structure about reality from text than many researchers expected, because human language quietly carries physics, psychology, culture, tools, plans, and cause-and-effect. But text is still a compressed residue of experience, not experience itself. A sentence can say a cup falls from a table, yet it does not fully encode weight, grip, balance, friction, timing, sound, surprise, or the tiny motor corrections a body makes before it even notices them. The world is not only made of facts that can be named; it is made of constraints that have to be lived through, touched, predicted, violated, and repaired. That is why world models matter. They aim to learn the hidden grammar of physical reality: how objects persist, how forces unfold, how space changes when an agent moves, and how action creates feedback. Language models can often reason about the world because people have written so much about it. World models try to learn what the world is like before it becomes words. The difference is exactly what matters because intelligence is not just answering well; it is knowing what would happen next if you moved, reached, pushed, smelled, slipped, or failed. A mind trained only on descriptions may become brilliant at explanation. A mind trained on experience may become better at consequence. --- Full video from "Google DeepMind" and "Hannah Fry" YT channel (link in comment)

Demis Hassabis on the limit in today’s AI: language can describe the world, but it cannot contain it - and why "World Models" are his "longest standing passion". Language models absorbed far more structure about reality from text than many researchers expected, because human language quietly carries physics, psychology, culture, tools, plans, and cause-and-effect. But text is still a compressed residue of experience, not experience itself. A sentence can say a cup falls from a table, yet it does not fully encode weight, grip, balance, friction, timing, sound, surprise, or the tiny motor corrections a body makes before it even notices them. The world is not only made of facts that can be named; it is made of constraints that have to be lived through, touched, predicted, violated, and repaired. That is why world models matter. They aim to learn the hidden grammar of physical reality: how objects persist, how forces unfold, how space changes when an agent moves, and how action creates feedback. Language models can often reason about the world because people have written so much about it. World models try to learn what the world is like before it becomes words. The difference is exactly what matters because intelligence is not just answering well; it is knowing what would happen next if you moved, reached, pushed, smelled, slipped, or failed. A mind trained only on descriptions may become brilliant at explanation. A mind trained on experience may become better at consequence. --- Full video from "Google DeepMind" and "Hannah Fry" YT channel (link in comment)

Rohan Paul

49,938 次观看 • 1 个月前

Sergey Brin says we're asking the wrong question about AI It's not just "can it do what humans do?", it's "what can it do that humans cannot do?" We don't know if there's a ceiling to intelligence "humans evolved over millions of years, but AI is iterating in months"

Sergey Brin says we're asking the wrong question about AI It's not just "can it do what humans do?", it's "what can it do that humans cannot do?" We don't know if there's a ceiling to intelligence "humans evolved over millions of years, but AI is iterating in months"

Haider.

90,630 次观看 • 6 个月前

Nexar’s real-world data just built the best incident prediction model — Meet BADAS 1.0. 🔥 It beat state-of-the-art models by learning from 10B+ real miles and 60M+ real events, not simulations. 🚀 Discover what it can do for you: #BADAS #AI #ADAS #AutonomousVehicles #RoadSafety

Nexar’s real-world data just built the best incident prediction model — Meet BADAS 1.0. 🔥 It beat state-of-the-art models by learning from 10B+ real miles and 60M+ real events, not simulations. 🚀 Discover what it can do for you: #BADAS #AI #ADAS #AutonomousVehicles #RoadSafety

Nexar

172,209 次观看 • 8 个月前

Dario Amodei says static data is no longer the most central thing in AI development Models are now relying on 'dynamic data', learning through trial and error by generating their own synthetic data in reinforcement environments "as AI creates its own experience, the open web matters less"

Dario Amodei says static data is no longer the most central thing in AI development Models are now relying on 'dynamic data', learning through trial and error by generating their own synthetic data in reinforcement environments "as AI creates its own experience, the open web matters less"

Haider.

16,326 次观看 • 4 个月前

Without World Models, There Is No AGI. Google Just Proved It. If AGI ever happens, it will not come from bigger chatbots alone. From the very start of this interview, one thing is crystal clear: without world models, we will never reach AGI. And right now, Google is leading with its world simulator Genie 3. Here is the core of what Demis Hassabis explains in this conversation: • World models are the missing core of AGI Hassabis says his deepest long term focus has always been world models and simulations. Not just language. Not just prediction. Actual internal simulations of reality. • LLMs are impressive, but incomplete Language models understand more about the world than expected because human language encodes a lot of reality. Still, language is only a shadow of the real thing. • What text can never fully teach Reality includes things text struggles to express: •3D space and spatial dynamics •Physical causality and mechanics •Sensorimotor experience like movement, force, smell, or balance • Experience beats description To close the gap, AI must learn from interaction and experience, not just static text. That is how you build an internal world simulator. • Why Genie 3 matters With Google DeepMind pushing systems like Genie 3, AI starts to model reality itself, not just talk about it. • Robots and real world assistants depend on this True robotics, smart glasses, and universal assistants require AI that understands the physical world you live in, not just your screen. Bottom line: AGI will not emerge from better text prediction. It will emerge from systems that can simulate, predict, and understand reality itself. Right now, Google is clearly ahead on that path. Curious what you think. Are world models the real AGI unlock, or just another stepping stone?

Without World Models, There Is No AGI. Google Just Proved It. If AGI ever happens, it will not come from bigger chatbots alone. From the very start of this interview, one thing is crystal clear: without world models, we will never reach AGI. And right now, Google is leading with its world simulator Genie 3. Here is the core of what Demis Hassabis explains in this conversation: • World models are the missing core of AGI Hassabis says his deepest long term focus has always been world models and simulations. Not just language. Not just prediction. Actual internal simulations of reality. • LLMs are impressive, but incomplete Language models understand more about the world than expected because human language encodes a lot of reality. Still, language is only a shadow of the real thing. • What text can never fully teach Reality includes things text struggles to express: •3D space and spatial dynamics •Physical causality and mechanics •Sensorimotor experience like movement, force, smell, or balance • Experience beats description To close the gap, AI must learn from interaction and experience, not just static text. That is how you build an internal world simulator. • Why Genie 3 matters With Google DeepMind pushing systems like Genie 3, AI starts to model reality itself, not just talk about it. • Robots and real world assistants depend on this True robotics, smart glasses, and universal assistants require AI that understands the physical world you live in, not just your screen. Bottom line: AGI will not emerge from better text prediction. It will emerge from systems that can simulate, predict, and understand reality itself. Right now, Google is clearly ahead on that path. Curious what you think. Are world models the real AGI unlock, or just another stepping stone?

VraserX e/acc

23,784 次观看 • 6 个月前

My conversation with Sergey Levine (Sergey Levine). Sergey is the co-founder of Physical Intelligence -- a company building foundation models that can control any robot to do any task in any environment. The company's thesis is that generality is more scalable than specialization, meaning that a model trained across many different robots and tasks will ultimately outperform any system built to do one thing well (eg, just wash dishes). Sergey is a researcher by background, but I think you will appreciate how practical and commercially grounded this conversation is. We discuss: - Why changing a diaper will be the last task a robot masters - The simulation v. real-world data debate - How multimodal LLMs give robots common sense - Moravec's Paradox + Robot Olympics - Why robots can do long-horizon tasks now - A realistic timeline for robots in our homes I should note that I am an investor in Physical Intelligence -- I made the investment because I believe it is one of the most important companies tackling the problem of robotics. Enjoy! Timestamps: 0:00 Intro 2:39 Defining Physical Intelligence 5:19 The Challenge of Building General Models 6:34 The Stakes and Future of General Purpose Robotics 8:15 Pros and Cons of Humanoid Robots 10:12 Historical Milestones in Robotics Research 15:31 Combining Generative AI and Deep RL 21:24 Moravec's Paradox 25:33 Kitchen Robots 29:30 Simulation vs. Real-World Data 30:48 The Robot Olympics 36:31 The Physiological Reality of Embodiment 38:56 Controversies in the Robotics Community 44:18 What Makes a Great Researcher 48:27 How Businesses Should Prepare for Robotics 54:09 Tracking Progress Through Research Papers 57:02 The Next Step: Mid-Level Reasoning 1:02:00 The Kindest Thing

My conversation with Sergey Levine (Sergey Levine). Sergey is the co-founder of Physical Intelligence -- a company building foundation models that can control any robot to do any task in any environment. The company's thesis is that generality is more scalable than specialization, meaning that a model trained across many different robots and tasks will ultimately outperform any system built to do one thing well (eg, just wash dishes). Sergey is a researcher by background, but I think you will appreciate how practical and commercially grounded this conversation is. We discuss: - Why changing a diaper will be the last task a robot masters - The simulation v. real-world data debate - How multimodal LLMs give robots common sense - Moravec's Paradox + Robot Olympics - Why robots can do long-horizon tasks now - A realistic timeline for robots in our homes I should note that I am an investor in Physical Intelligence -- I made the investment because I believe it is one of the most important companies tackling the problem of robotics. Enjoy! Timestamps: 0:00 Intro 2:39 Defining Physical Intelligence 5:19 The Challenge of Building General Models 6:34 The Stakes and Future of General Purpose Robotics 8:15 Pros and Cons of Humanoid Robots 10:12 Historical Milestones in Robotics Research 15:31 Combining Generative AI and Deep RL 21:24 Moravec's Paradox 25:33 Kitchen Robots 29:30 Simulation vs. Real-World Data 30:48 The Robot Olympics 36:31 The Physiological Reality of Embodiment 38:56 Controversies in the Robotics Community 44:18 What Makes a Great Researcher 48:27 How Businesses Should Prepare for Robotics 54:09 Tracking Progress Through Research Papers 57:02 The Next Step: Mid-Level Reasoning 1:02:00 The Kindest Thing

Patrick OShaughnessy

133,833 次观看 • 3 个月前

Over the last few months, we’ve been thinking about how to learn from “off-domain” data - data from non-robot sources like video or simulation. These data sources are not quite good enough to learn policies (even monolithic VLA models) directly, but they still contain lots of information that can be useful for generalizable robot control. How can we develop robot learning models that are able to make use of this type of data for generalizable control? In new work, that we call HAMSTER, we show that VLMs can be useful for enabling robotic learning from off-domain data, but specifically when used through hierarchical VLA architectures. We show that this class of models can learn generalizable robot policies for the real world from large-scale, off-domain data. A 🧵 (1/10)

Over the last few months, we’ve been thinking about how to learn from “off-domain” data - data from non-robot sources like video or simulation. These data sources are not quite good enough to learn policies (even monolithic VLA models) directly, but they still contain lots of information that can be useful for generalizable robot control. How can we develop robot learning models that are able to make use of this type of data for generalizable control? In new work, that we call HAMSTER, we show that VLMs can be useful for enabling robotic learning from off-domain data, but specifically when used through hierarchical VLA architectures. We show that this class of models can learn generalizable robot policies for the real world from large-scale, off-domain data. A 🧵 (1/10)

Abhishek Gupta

11,994 次观看 • 1 年前

Unify the #developer experience with Custom Model Import for #AmazonBedrock. ☁️✨ Developers can now access both base models & imported custom models through a unified API. #AWS 🔗

Unify the #developer experience with Custom Model Import for #AmazonBedrock. ☁️✨ Developers can now access both base models & imported custom models through a unified API. #AWS 🔗

Amazon Web Services

21,516 次观看 • 2 年前

Yann LeCun says language isn’t intelligence. Predicting text doesn’t mean understanding reality. The real world is messy, physical, and causal and today’s LLMs barely touch that. The next leap is Physical AI: world models, cause and effect, real planning. Do you think LLMs can evolve into this, or do we need a completely new architecture?

Yann LeCun says language isn’t intelligence. Predicting text doesn’t mean understanding reality. The real world is messy, physical, and causal and today’s LLMs barely touch that. The next leap is Physical AI: world models, cause and effect, real planning. Do you think LLMs can evolve into this, or do we need a completely new architecture?

VraserX e/acc

76,154 次观看 • 5 个月前

The goal isn't to tell a robot to fold your t-shirt. You should be able to prompt it “run my household” once, and have it cook, clean, shop, and adapt for months. Physical Intelligence co-founder Sergey Levine says once these robots start collecting real-world experience, the learning flywheel begins. Single digit years looks realistic.

The goal isn't to tell a robot to fold your t-shirt. You should be able to prompt it “run my household” once, and have it cook, clean, shop, and adapt for months. Physical Intelligence co-founder Sergey Levine says once these robots start collecting real-world experience, the learning flywheel begins. Single digit years looks realistic.

vitrupo

68,513 次观看 • 9 个月前

We pre-train LLMs on the whole of the internet. You might think this explains how they learn so many emergent capabilities: the knowledge is implicit in the training data. But in fact models can do things that were never demonstrated anywhere in training! Sergey Levine argues that the real source of emergent capabilities is compositionality:

We pre-train LLMs on the whole of the internet. You might think this explains how they learn so many emergent capabilities: the knowledge is implicit in the training data. But in fact models can do things that were never demonstrated anywhere in training! Sergey Levine argues that the real source of emergent capabilities is compositionality:

Dwarkesh Patel

151,680 次观看 • 1 个月前

Given the strong community adoption and real-world deployment of MolmoAct2 on YAM, we're introducing zero-shot evaluation of MolmoAct2 Bimanual YAM in simulation. Now you can test out our models without a real-world robot and build on them! Code: Simulation built on Maniskill!

Given the strong community adoption and real-world deployment of MolmoAct2 on YAM, we're introducing zero-shot evaluation of MolmoAct2 Bimanual YAM in simulation. Now you can test out our models without a real-world robot and build on them! Code: Simulation built on Maniskill!

Jiafei Duan

36,290 次观看 • 29 天前

Pay very close attention to what the NIO World Model can do, this is next level autonomous driving experience. $NIO

Pay very close attention to what the NIO World Model can do, this is next level autonomous driving experience. $NIO

Jason

23,650 次观看 • 9 个月前

NEWS: Nvidia has just introduced "Cosmos," a world foundation model created to understand the physical world. The model can generate synthetic data to train robotics.

NEWS: Nvidia has just introduced "Cosmos," a world foundation model created to understand the physical world. The model can generate synthetic data to train robotics.

Sawyer Merritt

254,296 次观看 • 1 年前

📁 Fei-Fei Li founder of World Labs, says the next leap in AI is not language. Human intelligence does not just speak, it moves, perceives, and acts in the physical world. Spatial intelligence is the real core of intelligence. From text to space, from models to 3D and 4D worlds, from understanding words to interacting with reality. The next chapter is not read, it is inhabited.

📁 Fei-Fei Li founder of World Labs, says the next leap in AI is not language. Human intelligence does not just speak, it moves, perceives, and acts in the physical world. Spatial intelligence is the real core of intelligence. From text to space, from models to 3D and 4D worlds, from understanding words to interacting with reality. The next chapter is not read, it is inhabited.

Jon Hernandez

20,592 次观看 • 5 个月前

Terence Tao says intelligence is not a straight line from subhuman to superhuman, where humans become irrelevant Instead, a 'Copernican view of intelligence' fits better — just as Earth is not the center of the universe, human intelligence is not the center of all cognition "we need to think beyond what humans or computers can do alone"

Terence Tao says intelligence is not a straight line from subhuman to superhuman, where humans become irrelevant Instead, a 'Copernican view of intelligence' fits better — just as Earth is not the center of the universe, human intelligence is not the center of all cognition "we need to think beyond what humans or computers can do alone"

Haider.

32,570 次观看 • 2 个月前