Video yükleniyor...

Video Yüklenemedi

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

Ana Sayfaya Dön

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... show more

David J Phillips

20,461 subscribers

110,268 görüntüleme • 5 ay önce •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

the micro1 robotics lab: real world data for intelligent models that co-exist in the physical world. we’re in-the-wild across 75 countries in 6,000+ unique environments collecting data. diverse movements, objects, and settings. the future of AI is as human as you can imagine. join us to start training robots today (link in comments).

the micro1 robotics lab: real world data for intelligent models that co-exist in the physical world. we’re in-the-wild across 75 countries in 6,000+ unique environments collecting data. diverse movements, objects, and settings. the future of AI is as human as you can imagine. join us to start training robots today (link in comments).

Ali Ansari

208,500 görüntüleme • 5 ay önce

Axis is officially LIVE on Base. 🔵 Axis is scaling Physical AI for the real world, contributed by everyone. You can control robots in a virtual world, generate training data at scale, and help build the brain behind tomorrow's robots. All from browser. No hardware needed. Start building robotics intelligence today:

Axis is officially LIVE on Base. 🔵 Axis is scaling Physical AI for the real world, contributed by everyone. You can control robots in a virtual world, generate training data at scale, and help build the brain behind tomorrow's robots. All from browser. No hardware needed. Start building robotics intelligence today:

Axis Robotics

393,954 görüntüleme • 5 ay önce

World Labs CEO Dr. Fei-Fei Li & SceniX Co-Founder Yunzhu Li on the data bottleneck in robotics and how world models help: "The lack of data in training, the lack of data in evaluation, this is very, very different from language models, where data is abundant on the internet." "And we know that in order for robotics to work, we have to somehow unlock the power of scaling law. But where does that come from?" "It's a profound problem that everybody's battling with in robotics." "We see a lot of unlock in being able to do this whole process through the modeling of the environments." "Being able to create these digital worlds... that's just going to unlock so much more potential for being able to replace all the costly and the unsafe data in the real environments with the data generated from the worlds for robots to be able to do scalable learning and evaluations." Fei-Fei Li Yunzhu Li martin_casado

World Labs CEO Dr. Fei-Fei Li & SceniX Co-Founder Yunzhu Li on the data bottleneck in robotics and how world models help: "The lack of data in training, the lack of data in evaluation, this is very, very different from language models, where data is abundant on the internet." "And we know that in order for robotics to work, we have to somehow unlock the power of scaling law. But where does that come from?" "It's a profound problem that everybody's battling with in robotics." "We see a lot of unlock in being able to do this whole process through the modeling of the environments." "Being able to create these digital worlds... that's just going to unlock so much more potential for being able to replace all the costly and the unsafe data in the real environments with the data generated from the worlds for robots to be able to do scalable learning and evaluations." Fei-Fei Li Yunzhu Li martin_casado

a16z

76,473 görüntüleme • 1 ay önce

1/7 Physical AI has a data problem. More compute alone won’t teach a robot how to move, grasp, react, or recover from mistakes. LLMs had the entire internet to learn from. Robots need something different: huge amounts of real action data. That’s where Axis Robotics gets interesting. $AXIS is building a data engine for Physical AI, turning robot training into something people can contribute to directly through a browser. The loop is straightforward: Generate → Collect → Process → Deploy → Correct → Repeat. What looks simple on the surface could become something much bigger underneath: a scalable way to produce the training data robots need to become genuinely useful in the physical world. The real unlock for Physical AI might not be another bigger model. It might be better data.

1/7 Physical AI has a data problem. More compute alone won’t teach a robot how to move, grasp, react, or recover from mistakes. LLMs had the entire internet to learn from. Robots need something different: huge amounts of real action data. That’s where Axis Robotics gets interesting. $AXIS is building a data engine for Physical AI, turning robot training into something people can contribute to directly through a browser. The loop is straightforward: Generate → Collect → Process → Deploy → Correct → Repeat. What looks simple on the surface could become something much bigger underneath: a scalable way to produce the training data robots need to become genuinely useful in the physical world. The real unlock for Physical AI might not be another bigger model. It might be better data.

_victorad

18,903 görüntüleme • 9 gün önce

We might be solving the wrong problem in robotics. That’s what this makes clear. UMI → Universal Manipulation Interface A simple $400 gripper that lets you teach robots by demonstration. You hold it like a tool. Show the task. The robot learns. No teleoperation. No expensive hardware. No robot-specific data. Stanford open-sourced everything → hardware, code, datasets. What stands out to me is the bottleneck. Not algorithms. Data. Teleoperation → ~35 demos/hour UMI → ~111 demos/hour And the data transfers across robots → UR5, Franka, others. The design is surprisingly practical: → GoPro fisheye lens (155° FOV) + mirrors for depth → SLAM + IMU for precise 6DoF tracking → latency matching for dynamic tasks → diffusion policies for multimodal actions Then it scales. Cheng Chi takes this further with Sunday Robotics (with Tony Zhao). A $200 glove → deployed in 500+ homes → ~10 million real-world interactions. Not lab data. Real human behavior. Their robot learns dishes, laundry, espresso → with zero robot-specific data. This is where the shift becomes obvious. From training robots in controlled environments → to learning directly from humans at scale So here’s the real question: Will robotics be unlocked by better models… or by unlocking data? #ArtificialIntelligence #Robotics #AI #Innovation #FutureOfWork

We might be solving the wrong problem in robotics. That’s what this makes clear. UMI → Universal Manipulation Interface A simple $400 gripper that lets you teach robots by demonstration. You hold it like a tool. Show the task. The robot learns. No teleoperation. No expensive hardware. No robot-specific data. Stanford open-sourced everything → hardware, code, datasets. What stands out to me is the bottleneck. Not algorithms. Data. Teleoperation → ~35 demos/hour UMI → ~111 demos/hour And the data transfers across robots → UR5, Franka, others. The design is surprisingly practical: → GoPro fisheye lens (155° FOV) + mirrors for depth → SLAM + IMU for precise 6DoF tracking → latency matching for dynamic tasks → diffusion policies for multimodal actions Then it scales. Cheng Chi takes this further with Sunday Robotics (with Tony Zhao). A $200 glove → deployed in 500+ homes → ~10 million real-world interactions. Not lab data. Real human behavior. Their robot learns dishes, laundry, espresso → with zero robot-specific data. This is where the shift becomes obvious. From training robots in controlled environments → to learning directly from humans at scale So here’s the real question: Will robotics be unlocked by better models… or by unlocking data? #ArtificialIntelligence #Robotics #AI #Innovation #FutureOfWork

Pascal Bornet

186,712 görüntüleme • 5 ay önce

It's easy for robots to nail a demo. It's much harder to get them to work reliably in the real world. The bottleneck is data. At Axis Robotics we're building a compounding data engine that combines simulation, ego-centric data and DAgger correction data at scale. Watch our CEO ChrisF(✱,✱) break down the data bottleneck, why messy data can still train good models, and what it takes to get robots out of the lab with blocmates. ⬇️

It's easy for robots to nail a demo. It's much harder to get them to work reliably in the real world. The bottleneck is data. At Axis Robotics we're building a compounding data engine that combines simulation, ego-centric data and DAgger correction data at scale. Watch our CEO ChrisF(✱,✱) break down the data bottleneck, why messy data can still train good models, and what it takes to get robots out of the lab with blocmates. ⬇️

Axis Robotics

22,634 görüntüleme • 23 gün önce

🦔Workers in India are wearing head-mounted cameras for 12 cents an hour to collect training data for humanoid robots. The footage of them doing everyday tasks like cooking, cleaning, sorting, and walking through public spaces gets sold to robotics companies building the models meant to replace those same kinds of jobs in higher-wage countries. The arrangement has been running for roughly two years. Workers do not own the data, do not get residuals, and in many cases are not told what their footage is being used to train. My Take The workers wearing the cameras live in a country where robotics automation will hit decades later, so they are training their own future replacements at a delay that hides the consequence from them personally. The companies buying the data are mostly US and Chinese, building humanoid robots aimed at warehouses, retail, and service jobs in countries paying $15 to $25 an hour rather than 12 cents. Robotics companies need motion data that mimics how humans actually move through real environments, and synthetic data has not been good enough yet. Paying 12 cents an hour in Bengaluru is cheaper than running motion capture studios in Boston, and it works at scale because the worker absorbs the cost of the camera, the discomfort of wearing it, and the long-term loss of any rights to their own movement data. The robotics labor market that eventually emerges from this footage will displace far more wages than the data collection cost to gather. That is the trade investors funding humanoid robotics startups are betting will pay off, and the workers in the videos are the ones paying the tab up front. Hedgie🤗

🦔Workers in India are wearing head-mounted cameras for 12 cents an hour to collect training data for humanoid robots. The footage of them doing everyday tasks like cooking, cleaning, sorting, and walking through public spaces gets sold to robotics companies building the models meant to replace those same kinds of jobs in higher-wage countries. The arrangement has been running for roughly two years. Workers do not own the data, do not get residuals, and in many cases are not told what their footage is being used to train. My Take The workers wearing the cameras live in a country where robotics automation will hit decades later, so they are training their own future replacements at a delay that hides the consequence from them personally. The companies buying the data are mostly US and Chinese, building humanoid robots aimed at warehouses, retail, and service jobs in countries paying $15 to $25 an hour rather than 12 cents. Robotics companies need motion data that mimics how humans actually move through real environments, and synthetic data has not been good enough yet. Paying 12 cents an hour in Bengaluru is cheaper than running motion capture studios in Boston, and it works at scale because the worker absorbs the cost of the camera, the discomfort of wearing it, and the long-term loss of any rights to their own movement data. The robotics labor market that eventually emerges from this footage will displace far more wages than the data collection cost to gather. That is the trade investors funding humanoid robotics startups are betting will pay off, and the workers in the videos are the ones paying the tab up front. Hedgie🤗

Hedgie

162,060 görüntüleme • 3 ay önce

Karol Hausman is the co-founder and CEO of Physical Intelligence, a robotics company building a general-purpose “AI brain for the physical world.” The company has raised more than $1 billion in funding to develop foundation models that allow robots to operate across many machines, environments, and tasks rather than being programmed for a single purpose. In our conversation, we explore: • The moment a lecture from Sergey Levine convinced him to abandon his PhD research direction and pivot fully to deep learning • The case for building a general “AI brain” for the physical world rather than a single specialized robot • The role of real-world data in training robots, the limits of simulation, and how deployment could create a powerful data flywheel • The unique challenges of physical intelligence and why robots must operate with far higher reliability than language models Thank you to the partners who make this possible - Brex: The intelligent finance platform: - Granola: The app that might actually make you love meetings: Timestamps (00:00) Intro (04:05) Karol’s early fascination with robots (18:21) Karol’s entry point to robotics and PhD program (25:49) Combining robotics with LLMs: The Taylor Swift demo (30:48) The 1970s SHRDLU AI experiment (39:40) How research shapes what Physical Intelligence builds (49:07) The return of reinforcement learning in robotics (1:00:00) NVIDIA’s simulation engines (1:07:31) Compensating for missing senses

Karol Hausman is the co-founder and CEO of Physical Intelligence, a robotics company building a general-purpose “AI brain for the physical world.” The company has raised more than $1 billion in funding to develop foundation models that allow robots to operate across many machines, environments, and tasks rather than being programmed for a single purpose. In our conversation, we explore: • The moment a lecture from Sergey Levine convinced him to abandon his PhD research direction and pivot fully to deep learning • The case for building a general “AI brain” for the physical world rather than a single specialized robot • The role of real-world data in training robots, the limits of simulation, and how deployment could create a powerful data flywheel • The unique challenges of physical intelligence and why robots must operate with far higher reliability than language models Thank you to the partners who make this possible - Brex: The intelligent finance platform: - Granola: The app that might actually make you love meetings: Timestamps (00:00) Intro (04:05) Karol’s early fascination with robots (18:21) Karol’s entry point to robotics and PhD program (25:49) Combining robotics with LLMs: The Taylor Swift demo (30:48) The 1970s SHRDLU AI experiment (39:40) How research shapes what Physical Intelligence builds (49:07) The return of reinforcement learning in robotics (1:00:00) NVIDIA’s simulation engines (1:07:31) Compensating for missing senses

Mario Gabriele 🦊

27,871 görüntüleme • 5 ay önce

Chelsea Finn (Chelsea Finn) on building general-purpose robotics, and bringing intelligence into the physical world. At AI Startup School in San Francisco. 00:00 - General Purpose Robots 00:11 - Challenges in Robotics Applications 00:57 - Physical Intelligence: A New Approach 01:47 - Learning from Language Models 02:08 - Data Sources for Training Robots 03:32 - Training with Real-World Data 04:39 - Initial Successes and Challenges 09:10 - Breakthrough in Robot Training 11:03 - Improving Performance 15:43 - Expanding Capabilities 17:34 - Robots in Unseen Environments 25:54 - Handling Open-Ended Prompts 29:36 - Evaluating Robot Performance 30:03 - Future Directions and Challenges 31:27 - Audience Q&A

Chelsea Finn (Chelsea Finn) on building general-purpose robotics, and bringing intelligence into the physical world. At AI Startup School in San Francisco. 00:00 - General Purpose Robots 00:11 - Challenges in Robotics Applications 00:57 - Physical Intelligence: A New Approach 01:47 - Learning from Language Models 02:08 - Data Sources for Training Robots 03:32 - Training with Real-World Data 04:39 - Initial Successes and Challenges 09:10 - Breakthrough in Robot Training 11:03 - Improving Performance 15:43 - Expanding Capabilities 17:34 - Robots in Unseen Environments 25:54 - Handling Open-Ended Prompts 29:36 - Evaluating Robot Performance 30:03 - Future Directions and Challenges 31:27 - Audience Q&A

Y Combinator

129,834 görüntüleme • 1 yıl önce

You asked for it, we delivered. The full keynote is live. In this keynote, we unveil our enhanced vision to become the World’s Ears for AI & Robotics. We are creating the world’s first decentralized infrastructure for real-world audio intelligence. Check-out our new website: Start training Voice AI now & get rewarded: From noise-level mapping to multilingual voice data and the future of live audio streams, Silencio is building the critical audio data layer that will power foundation models, embodied AI, autonomous vehicles, and robotics.

You asked for it, we delivered. The full keynote is live. In this keynote, we unveil our enhanced vision to become the World’s Ears for AI & Robotics. We are creating the world’s first decentralized infrastructure for real-world audio intelligence. Check-out our new website: Start training Voice AI now & get rewarded: From noise-level mapping to multilingual voice data and the future of live audio streams, Silencio is building the critical audio data layer that will power foundation models, embodied AI, autonomous vehicles, and robotics.

Silencio | Voice Data for AI

176,407 görüntüleme • 11 ay önce

AI is only as good as its data. Everyone talks about building smarter AI models, but fewer people ask where the training data comes from. That’s what makes Humyn Labs interesting. They’re focused on collecting high-quality, real-world human data across different environments, behaviors, and sensory inputs. Because if AI is going to understand the physical world, it first needs to learn from the people who already live and work in it. I shared my thoughts in this short video. Let me know what stood out to you. 👇🎥 Humyn Labs KGeN Community #KGeN

AI is only as good as its data. Everyone talks about building smarter AI models, but fewer people ask where the training data comes from. That’s what makes Humyn Labs interesting. They’re focused on collecting high-quality, real-world human data across different environments, behaviors, and sensory inputs. Because if AI is going to understand the physical world, it first needs to learn from the people who already live and work in it. I shared my thoughts in this short video. Let me know what stood out to you. 👇🎥 Humyn Labs KGeN Community #KGeN

Bahdfella™ ❤️‍🔥

18,156 görüntüleme • 1 ay önce

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 Sun

15,351 görüntüleme • 2 ay önce

Robots can already fold laundry, make espresso, clean kitchens, and assemble things. The harder problem is getting them to do those tasks reliably, for long periods of time, without a human babysitting them. At Startup School 2026, Physical Intelligence cofounder Chelsea Finn explains what it takes to build general-purpose robots that work in the real world. She shares how reinforcement learning pushed robot throughput up 2x, how their systems can run autonomously for hours, and why she believes robotics is entering its GPT era: moving from specialized models toward general-purpose systems that can work across tasks, robots, and environments. 00:00 — The State of Physical Intelligence 01:23 — What It Takes to Make Robots Useful 05:11 — The Reliability Problem 07:43 — Reinforcement Learning for Robotics 09:35 — Learning From Failures 12:43 — Training Robots to Improve Themselves 14:21 — Can a Robot Work for 13 Hours Straight? 17:36 — Why Robots Need Memory 21:22 — Building a General-Purpose Robot 25:02 — From Fine-Tuning to Out-of-the-Box Models 27:35 — Training on All the Data 30:20 — One Model That Beats the Specialists 31:21 — Compositional Generalization 37:49 — The GPT Era of Robotics 39:49 — Q&A

Robots can already fold laundry, make espresso, clean kitchens, and assemble things. The harder problem is getting them to do those tasks reliably, for long periods of time, without a human babysitting them. At Startup School 2026, Physical Intelligence cofounder Chelsea Finn explains what it takes to build general-purpose robots that work in the real world. She shares how reinforcement learning pushed robot throughput up 2x, how their systems can run autonomously for hours, and why she believes robotics is entering its GPT era: moving from specialized models toward general-purpose systems that can work across tasks, robots, and environments. 00:00 — The State of Physical Intelligence 01:23 — What It Takes to Make Robots Useful 05:11 — The Reliability Problem 07:43 — Reinforcement Learning for Robotics 09:35 — Learning From Failures 12:43 — Training Robots to Improve Themselves 14:21 — Can a Robot Work for 13 Hours Straight? 17:36 — Why Robots Need Memory 21:22 — Building a General-Purpose Robot 25:02 — From Fine-Tuning to Out-of-the-Box Models 27:35 — Training on All the Data 30:20 — One Model That Beats the Specialists 31:21 — Compositional Generalization 37:49 — The GPT Era of Robotics 39:49 — Q&A

Y Combinator

109,863 görüntüleme • 22 gün önce

A simple idea. Let robots collect the data that current foundation models are missing. A robot that gets better by doing real work in the real world. For two weeks in the Stanford East Asia Library, Scanford scanned shelves, helped librarians, and improved the vision language model it depends on. The idea is very simple: Robots do useful work. They gather the real world data foundation models never see online. They fine tune their own model They go back out stronger A full loop. What they found in deployment: ✅ 2103 shelves scanned with multilingual, faded, occluded book spines ✅ 18.7 hours of librarian time saved ✅ Book ID accuracy jumped from 32.0 percent to 71.8 percent ✅ English OCR improved from 24.8 percent to 46.6 percent ✅ Chinese OCR improved from 30.8 percent to 38.0 percent The most interesting part is the shift. Robots do not only consume foundation models. They create the data these models are missing. A clean robot powered data flywheel. Work. Collect. Fine tune. Repeat. Thanks for sharing, Jenn Grannen! If you want the full write up: 📍Website: Paper: —- Weekly robotics and AI insights. Subscribe free:

A simple idea. Let robots collect the data that current foundation models are missing. A robot that gets better by doing real work in the real world. For two weeks in the Stanford East Asia Library, Scanford scanned shelves, helped librarians, and improved the vision language model it depends on. The idea is very simple: Robots do useful work. They gather the real world data foundation models never see online. They fine tune their own model They go back out stronger A full loop. What they found in deployment: ✅ 2103 shelves scanned with multilingual, faded, occluded book spines ✅ 18.7 hours of librarian time saved ✅ Book ID accuracy jumped from 32.0 percent to 71.8 percent ✅ English OCR improved from 24.8 percent to 46.6 percent ✅ Chinese OCR improved from 30.8 percent to 38.0 percent The most interesting part is the shift. Robots do not only consume foundation models. They create the data these models are missing. A clean robot powered data flywheel. Work. Collect. Fine tune. Repeat. Thanks for sharing, Jenn Grannen! If you want the full write up: 📍Website: Paper: —- Weekly robotics and AI insights. Subscribe free:

Ilir Aliu

44,660 görüntüleme • 9 ay önce

Today I assembled the first batch of reBot DevArm in the U.S! First one on the planet! 🇺🇸⚙️ But this isn’t about a robot arm. This is about what happens when open source hits robotics. For years: •Hardware was locked •Robotics was expensive •Embodied AI was out of reach Now? Anyone can get their hands on: •6 DOF manipulation •Real precision (<0.2 mm repeatability) •Full stack access (ROS, Isaac Sim, LeRobot) This is how movements start. Not in labs. Not behind paywalls. But in the hands of builders. Massive respect to Seeed Studio for pushing true open-source robotics forward! We’re about to see: → Thousands of robot arms trained like athletes → Real-world datasets explode → Physical AI built in garages, not just research labs At SoloTech , this is the unlock we’ve been waiting for, training and deploying Physical AI skill models on real hardware, anywhere. No sims. No vaporware. No excuses. Production Grade Physical AI. Want to get access? Check out the Physical AI Gym in San Francisco or DM me.

Today I assembled the first batch of reBot DevArm in the U.S! First one on the planet! 🇺🇸⚙️ But this isn’t about a robot arm. This is about what happens when open source hits robotics. For years: •Hardware was locked •Robotics was expensive •Embodied AI was out of reach Now? Anyone can get their hands on: •6 DOF manipulation •Real precision (<0.2 mm repeatability) •Full stack access (ROS, Isaac Sim, LeRobot) This is how movements start. Not in labs. Not behind paywalls. But in the hands of builders. Massive respect to Seeed Studio for pushing true open-source robotics forward! We’re about to see: → Thousands of robot arms trained like athletes → Real-world datasets explode → Physical AI built in garages, not just research labs At SoloTech , this is the unlock we’ve been waiting for, training and deploying Physical AI skill models on real hardware, anywhere. No sims. No vaporware. No excuses. Production Grade Physical AI. Want to get access? Check out the Physical AI Gym in San Francisco or DM me.

Dhruv Diddi

31,751 görüntüleme • 4 ay önce

What if a dexterous robot could learn a 20+ step chemistry experiment without a single on-robot training demo? Meet TwinDEX: a pair of co-designed, three-finger, nine-DoF dexterous manipulation interface: one wearable for data collection, one for robot deployment. The twinned design shares identical kinematics, contact surfaces, visual appearance, and sensors across collection and deployment — keeping observations and actions aligned end to end. Trained from scratch on only a few hundred wearable demonstrations - with zero on-robot training or intervention data - TwinDEX completed a standardized chemistry experiment involving tool switches, fine force control, and bimanual coordination. Robot-free data showed comparable learning efficiency on the multi-task evaluation, TwinDEX delivered 5.3 times effective throughput than on-robot teleoperation. TwinDEX demonstrates that high-quality robot-free data can fully substitute for on-robot teleoperation data on challenging dexterous tasks — removing the dependency on real-robot hardware that has been the central bottleneck to scaling dexterous manipulation data. This was the proof-it phase. Now comes scale: what emerges at tens of thousands, or millions, of episodes? Watch the demo and read the technical blog: #TwinDEX #Robotics #EmbodiedAI #DexterousManipulation

What if a dexterous robot could learn a 20+ step chemistry experiment without a single on-robot training demo? Meet TwinDEX: a pair of co-designed, three-finger, nine-DoF dexterous manipulation interface: one wearable for data collection, one for robot deployment. The twinned design shares identical kinematics, contact surfaces, visual appearance, and sensors across collection and deployment — keeping observations and actions aligned end to end. Trained from scratch on only a few hundred wearable demonstrations - with zero on-robot training or intervention data - TwinDEX completed a standardized chemistry experiment involving tool switches, fine force control, and bimanual coordination. Robot-free data showed comparable learning efficiency on the multi-task evaluation, TwinDEX delivered 5.3 times effective throughput than on-robot teleoperation. TwinDEX demonstrates that high-quality robot-free data can fully substitute for on-robot teleoperation data on challenging dexterous tasks — removing the dependency on real-robot hardware that has been the central bottleneck to scaling dexterous manipulation data. This was the proof-it phase. Now comes scale: what emerges at tens of thousands, or millions, of episodes? Watch the demo and read the technical blog: #TwinDEX #Robotics #EmbodiedAI #DexterousManipulation

X Square Robot

1,320,678 görüntüleme • 1 gün önce

I am stocked to announce that I won the OpenAI Developers Codex x Mollie Hacka Worldwide Hackathon in Paris. 60+ builders, every one of us working solo, one day to ship. I built mine around a single question: who gets to own intelligence? The default answer is scary. You hand your data to a handful of labs, they train the model, they own it, and you rent back a thin slice of what your own data made possible. That is the bargain on the table today. I do not accept it. So I built Lensemble: a Tapestry like distributed training platform for JEPA based World Models. What does it enable: World Models that a community improves together, keeps sovereign, and co-owns. Two bets sit underneath it. First, the paradigm. Language models predict the next token. Powerful for text, a dead end for the physical world. A robot does not need to autocomplete sentences, it needs to predict what happens next in the world. That is what JEPA does: it learns by predicting representations instead of pixels or tokens. I am convinced world models are the most underrated paradigm in AI right now, and the closest thing we have to a ChatGPT moment for robotics. Second, the politics. Your raw trajectories never leave your machine. Each participant trains locally against a shared protocol and ships only an update, never the data. A federated round folds those updates into one shared world model, a LeWorldModel based model, and the gain is measured, not claimed: a 12k-parameter adapter on a frozen backbone, held-out prediction error down about 12 percent, the model measurably less surprised by the world. Then the upside is split by contribution weight, so the people who improved the model own a share of what it earns. This is the thesis behind Project Tapestry, the AI Alliance and Yann LeCun's push for federated, sovereign frontier AI, carried into world models and robotics. Call it Tapestry for the physical world. All of it built solo, in a single day, with Codex as my pair the whole way. Thank you to OpenAI Codex and Mollie for backing builders who ship real things, and to Boris and the organizing crew for the room and the standard you set. Intelligence the world improves, and the world owns. That is the future I want for my kids, and the one I will keep building.

I am stocked to announce that I won the OpenAI Developers Codex x Mollie Hacka Worldwide Hackathon in Paris. 60+ builders, every one of us working solo, one day to ship. I built mine around a single question: who gets to own intelligence? The default answer is scary. You hand your data to a handful of labs, they train the model, they own it, and you rent back a thin slice of what your own data made possible. That is the bargain on the table today. I do not accept it. So I built Lensemble: a Tapestry like distributed training platform for JEPA based World Models. What does it enable: World Models that a community improves together, keeps sovereign, and co-owns. Two bets sit underneath it. First, the paradigm. Language models predict the next token. Powerful for text, a dead end for the physical world. A robot does not need to autocomplete sentences, it needs to predict what happens next in the world. That is what JEPA does: it learns by predicting representations instead of pixels or tokens. I am convinced world models are the most underrated paradigm in AI right now, and the closest thing we have to a ChatGPT moment for robotics. Second, the politics. Your raw trajectories never leave your machine. Each participant trains locally against a shared protocol and ships only an update, never the data. A federated round folds those updates into one shared world model, a LeWorldModel based model, and the gain is measured, not claimed: a 12k-parameter adapter on a frozen backbone, held-out prediction error down about 12 percent, the model measurably less surprised by the world. Then the upside is split by contribution weight, so the people who improved the model own a share of what it earns. This is the thesis behind Project Tapestry, the AI Alliance and Yann LeCun's push for federated, sovereign frontier AI, carried into world models and robotics. Call it Tapestry for the physical world. All of it built solo, in a single day, with Codex as my pair the whole way. Thank you to OpenAI Codex and Mollie for backing builders who ship real things, and to Boris and the organizing crew for the room and the standard you set. Intelligence the world improves, and the world owns. That is the future I want for my kids, and the one I will keep building.

abdel

20,037 görüntüleme • 2 ay önce

Every voice AI has the same problem The moment you stop talking, it starts talking But humans don't work that way In real conversations, pauses don't always mean interruption. Sometimes they mean: keep going Amit Yadav spent years building real-time AI systems at Meta Reality Labs for Ray-Ban smart glasses Now he's building Fern: a sales copilot that assists teams live during meetings Not after the call ...During it The hard problem isn't the answers anymore. It's timing. Turns out the hardest thing to teach an AI isn't what to say It's when to stay quiet. 🎙️ Amit Yadav (Amit Yadav), CEO & Co-Founder of Fern (ishikilabs), on Fondo.com START Huge thanks to Rho for making this episode possible!

Every voice AI has the same problem The moment you stop talking, it starts talking But humans don't work that way In real conversations, pauses don't always mean interruption. Sometimes they mean: keep going Amit Yadav spent years building real-time AI systems at Meta Reality Labs for Ray-Ban smart glasses Now he's building Fern: a sales copilot that assists teams live during meetings Not after the call ...During it The hard problem isn't the answers anymore. It's timing. Turns out the hardest thing to teach an AI isn't what to say It's when to stay quiet. 🎙️ Amit Yadav (Amit Yadav), CEO & Co-Founder of Fern (ishikilabs), on Fondo.com START Huge thanks to Rho for making this episode possible!

David J Phillips

34,155 görüntüleme • 3 ay önce

An interactive world model developed by NVIDIA in collaboration with academic partners. - DreamDojo turns egocentric human video data into physical intelligence. - Human data is more scalable than robotics data but lacks action labels. - To solve this, a dedicated action model extracts latent actions by identifying physics and motion deltas between frames. Training - A massive 44k hours of video data are used for pre-training. - Post-training on small-scale robot datasets maps human physics to specific robot embodiments. - An additional distillation stage converts the model into an autoregressive, few-step diffusion model, enabling real-time, action-controllable simulation. Primary Use Cases - Live Teleoperation: Controlling a robot inside a world simulation in real-time. - Model-based Planning: Previewing and curating the best actions for improved success. - Policy Evaluation: Testing robot policies in realistic, out-of-distribution scenarios. Everything that's open-sourced: weights, code, post-training dataset, eval set, and details to reproduce.

An interactive world model developed by NVIDIA in collaboration with academic partners. - DreamDojo turns egocentric human video data into physical intelligence. - Human data is more scalable than robotics data but lacks action labels. - To solve this, a dedicated action model extracts latent actions by identifying physics and motion deltas between frames. Training - A massive 44k hours of video data are used for pre-training. - Post-training on small-scale robot datasets maps human physics to specific robot embodiments. - An additional distillation stage converts the model into an autoregressive, few-step diffusion model, enabling real-time, action-controllable simulation. Primary Use Cases - Live Teleoperation: Controlling a robot inside a world simulation in real-time. - Model-based Planning: Previewing and curating the best actions for improved success. - Policy Evaluation: Testing robot policies in realistic, out-of-distribution scenarios. Everything that's open-sourced: weights, code, post-training dataset, eval set, and details to reproduce.

The Humanoid Hub

11,575 görüntüleme • 6 ay önce