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。

🚨 UNCANNY VALLEY: “FOLDING LAUNDRY IS STILL UNSOLVED” Rajat Bhageria explains why specialized, high-use robots are more practical than general-purpose bots trying to do everything. Founder of Chef Robotics, Rajat Bhageria: “Already, just to do chef, a robot needs to handle any ingredient, any cut, any cooking state. That’s... show more

Mario Nawfal

2,277,686 subscribers

32,264 次观看 • 1 年前 •via X (Twitter)

科学技术教育

Anya Rossi• Live Now

Private livecam show

5 条评论

max acs 的头像

max acs1 年前

@RajatBhageria @ChefRobotics He has a point tho

Eddy 💹🧲 的头像

Eddy 💹🧲1 年前

@RajatBhageria @ChefRobotics It's a very good point of view

Degens Journal 的头像

Degens Journal1 年前

@RajatBhageria @ChefRobotics Good video brother

Anonymous 的头像

Anonymous1 年前

@RajatBhageria @ChefRobotics It is true. When robots are faced with some difficult actions, they still need human assistance. This not only does not save time and manpower, but is also a waste of resources.

Nairebis - e/max-acc 的头像

Nairebis - e/max-acc1 年前

@RajatBhageria @ChefRobotics What? Optimus was folding laundry a year and a half ago. Presumably it's even better now.

相关视频

🚨 UNCANNY VALLEY: WHY SPECIALIZED ROBOTS SCALE FASTER Rajat Bhageria explains why focus, not general purpose, is driving real impact in robotics today. Founder of Chef Robotics, Rajat Bhageria: “You don't need Teslas to make food. And the robot that makes your food doesn't need to drive your Tesla. Our robots assemble food sixteen hours a day, every day…and that's all they do.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: WHY SPECIALIZED ROBOTS SCALE FASTER Rajat Bhageria explains why focus, not general purpose, is driving real impact in robotics today. Founder of Chef Robotics, Rajat Bhageria: “You don't need Teslas to make food. And the robot that makes your food doesn't need to drive your Tesla. Our robots assemble food sixteen hours a day, every day…and that's all they do.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

95,784 次观看 • 1 年前

🚨 UNCANNY VALLEY: “YOUTUBE VIDEOS WON’T TRAIN ROBOTS” Rajat Bhageria explains why learning from YouTube demos fails when it comes to building robots that actually work in the real world. Founder of Chef Robotics, Rajat Bhageria: “My hands can do weird motions a robot physically can’t. There’s a grounding problem, robots have constraints, and most demo data isn’t even from the right perspective. This isn’t LLMs. You can’t just download the internet and expect it to work.” Uncanny Valley with Danish only on 𝕏.

🚨 UNCANNY VALLEY: “YOUTUBE VIDEOS WON’T TRAIN ROBOTS” Rajat Bhageria explains why learning from YouTube demos fails when it comes to building robots that actually work in the real world. Founder of Chef Robotics, Rajat Bhageria: “My hands can do weird motions a robot physically can’t. There’s a grounding problem, robots have constraints, and most demo data isn’t even from the right perspective. This isn’t LLMs. You can’t just download the internet and expect it to work.” Uncanny Valley with Danish only on 𝕏.

Mario Nawfal

257,503 次观看 • 1 年前

🚨 UNCANNY VALLEY: “YOU NEED THE KILLER APP BEFORE THE PLATFORM” Rajat Bhageria explains why vertical success must come before platform scale in robotics. Founder of Chef Robotics, Rajat Bhageria: “Before AWS, you needed Amazon to exist. Before iOS tools took off, you needed the iPhone. With robotics, you need real customer value first, before platforms can thrive. If Chef can ship tens of thousands of robots, we show the world this industry matters.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: “YOU NEED THE KILLER APP BEFORE THE PLATFORM” Rajat Bhageria explains why vertical success must come before platform scale in robotics. Founder of Chef Robotics, Rajat Bhageria: “Before AWS, you needed Amazon to exist. Before iOS tools took off, you needed the iPhone. With robotics, you need real customer value first, before platforms can thrive. If Chef can ship tens of thousands of robots, we show the world this industry matters.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

86,499 次观看 • 1 年前

🚨 UNCANNY VALLEY: “NOTHING FUNDAMENTAL CHANGED” Rajat Bhageria questions the hype behind humanoid robots and why the excitement might be premature. Founder of Chef Robotics, Rajat Bhageria: “What actually changed that led to all this hype? CPUs and GPUs got better, cameras got a little cheaper, but nothing fundamental. It’s not like the transistor was invented and suddenly humanoids are possible.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: “NOTHING FUNDAMENTAL CHANGED” Rajat Bhageria questions the hype behind humanoid robots and why the excitement might be premature. Founder of Chef Robotics, Rajat Bhageria: “What actually changed that led to all this hype? CPUs and GPUs got better, cameras got a little cheaper, but nothing fundamental. It’s not like the transistor was invented and suddenly humanoids are possible.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

38,475 次观看 • 1 年前

🚨 UNCANNY VALLEY: “TENS OF BILLIONS OF WALL-Es ARE COMING” Rajat Bhageria shares why the future of robotics isn’t just humanoids – it’s also task-specific bots built for real jobs. Founder of Chef Robotics, Rajat Bhageria: “There’s gonna be tens of billions of Wall-E type robots for focused tasks like assembling food or building homes. And then billions of humanoids for more complex service work in places like homes or hospitals.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: “TENS OF BILLIONS OF WALL-Es ARE COMING” Rajat Bhageria shares why the future of robotics isn’t just humanoids – it’s also task-specific bots built for real jobs. Founder of Chef Robotics, Rajat Bhageria: “There’s gonna be tens of billions of Wall-E type robots for focused tasks like assembling food or building homes. And then billions of humanoids for more complex service work in places like homes or hospitals.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

46,230 次观看 • 1 年前

🚨UNCANNY VALLEY: “ROBOTICS ISN’T UNLOCKED YET…BUT IT’S COMING” Chef Robotics founder Rajat Bhageria joins Danish on Uncanny Valley to share why real-world robotics is still hard and why now might finally be the moment. Rajat Bhageria: “LLMs are good for task-level planning. But that’s not the hard part. The hard part is fine motor control. There’s no 10x unlock. No ChatGPT moment. But we’re getting closer. If we can become the success story, we can help drive the whole industry forward.” Full episode of Uncanny Valley coming soon – only on 𝕏

🚨UNCANNY VALLEY: “ROBOTICS ISN’T UNLOCKED YET…BUT IT’S COMING” Chef Robotics founder Rajat Bhageria joins Danish on Uncanny Valley to share why real-world robotics is still hard and why now might finally be the moment. Rajat Bhageria: “LLMs are good for task-level planning. But that’s not the hard part. The hard part is fine motor control. There’s no 10x unlock. No ChatGPT moment. But we’re getting closer. If we can become the success story, we can help drive the whole industry forward.” Full episode of Uncanny Valley coming soon – only on 𝕏

Mario Nawfal

99,668 次观看 • 1 年前

🚨 UNCANNY VALLEY: “10X IS OUR WORST CASE AND THAT KEEPS US GROUNDED” Rajat Bhageria shares how he thinks about valuation and why realism beats hype. Founder of Chef Robotics, Rajat Bhageria: “I hope we can get 40x, 50x, even 100x – whatever investors are willing to give. But I assume a worst-case 10x multiple on enterprise value to revenue. Even some SaaS companies are struggling to get 6x. Being grounded in the worst case helps us build smarter.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: “10X IS OUR WORST CASE AND THAT KEEPS US GROUNDED” Rajat Bhageria shares how he thinks about valuation and why realism beats hype. Founder of Chef Robotics, Rajat Bhageria: “I hope we can get 40x, 50x, even 100x – whatever investors are willing to give. But I assume a worst-case 10x multiple on enterprise value to revenue. Even some SaaS companies are struggling to get 6x. Being grounded in the worst case helps us build smarter.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

46,238 次观看 • 1 年前

🚨 UNCANNY VALLEY: “ROBOTICS IS HOT AGAIN – WE’RE FOCUSED ON BUILDING A REAL BUSINESS” Rajat Bhageria reflects on the cycles of hype and hardship and why staying focused on customers is what really matters. Founder of Chef Robotics, Rajat Bhageria: “Our last round was the fastest we’ve ever closed, but we’ve seen the hard years too. That’s why we’re not building for fundraising hype. We’re here to build a great business, with happy customers and strong repeat orders. That’s what matters.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: “ROBOTICS IS HOT AGAIN – WE’RE FOCUSED ON BUILDING A REAL BUSINESS” Rajat Bhageria reflects on the cycles of hype and hardship and why staying focused on customers is what really matters. Founder of Chef Robotics, Rajat Bhageria: “Our last round was the fastest we’ve ever closed, but we’ve seen the hard years too. That’s why we’re not building for fundraising hype. We’re here to build a great business, with happy customers and strong repeat orders. That’s what matters.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

119,624 次观看 • 1 年前

🚨 UNCANNY VALLEY: “ROBOTICS DOESN’T HAVE TO BE A RISKY BET” Rajat Bhageria shares how Chef Robotics helps hesitant customers adopt robotics by removing technical risk and showing real ROI. Founder of Chef Robotics, Rajat Bhageria: “If it’s not a good time for you, that’s okay – we’re not here to push risky deployments. We show clear ROI case studies, simulate robot performance on your exact meal data, and structure contracts so you only pay once it works on your line. It takes vision, but it’s not half a million upfront. When the numbers add up, the tech sells itself.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: “ROBOTICS DOESN’T HAVE TO BE A RISKY BET” Rajat Bhageria shares how Chef Robotics helps hesitant customers adopt robotics by removing technical risk and showing real ROI. Founder of Chef Robotics, Rajat Bhageria: “If it’s not a good time for you, that’s okay – we’re not here to push risky deployments. We show clear ROI case studies, simulate robot performance on your exact meal data, and structure contracts so you only pay once it works on your line. It takes vision, but it’s not half a million upfront. When the numbers add up, the tech sells itself.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

107,679 次观看 • 1 年前

🚨 UNCANNY VALLEY: “WE CAN MAKE SALMON THAT’S 0.001% TO 100% REAL FISH” Wildtype CEO Justin Kolbeck joins Danish to explain how cultivated seafood works – and why it doesn’t need constant FDA reapproval. Justin Kolbeck: “Our plant-based ingredients are already in the food system. The FDA reviewed the novel ingredient, which is how the salmon cells are grown. We can make a product that’s 0.001% salmon or 100% salmon. And we can change things like swapping canola for olive oil without needing a new review.” Uncanny Valley with Danish – only on 𝕏.

🚨 UNCANNY VALLEY: “WE CAN MAKE SALMON THAT’S 0.001% TO 100% REAL FISH” Wildtype CEO Justin Kolbeck joins Danish to explain how cultivated seafood works – and why it doesn’t need constant FDA reapproval. Justin Kolbeck: “Our plant-based ingredients are already in the food system. The FDA reviewed the novel ingredient, which is how the salmon cells are grown. We can make a product that’s 0.001% salmon or 100% salmon. And we can change things like swapping canola for olive oil without needing a new review.” Uncanny Valley with Danish – only on 𝕏.

Mario Nawfal

217,491 次观看 • 11 个月前

.Arielle "AZ" Zuckerberg on the new frontier of robotics: value shifts as training costs collapse. “The cost of training robots to do something that’s actually creating value has just gone down immensely.” “I agree that having a robot in my house seems a little Black Mirror-esque. But people are receptive to personal relationships with AI, like ChatGPT as a friend and coach.” “I’m much more into the idea of ambient, single-purpose objects. For instance, a robot that lives in the laundry room, moving and folding your laundry.”

.Arielle "AZ" Zuckerberg on the new frontier of robotics: value shifts as training costs collapse. “The cost of training robots to do something that’s actually creating value has just gone down immensely.” “I agree that having a robot in my house seems a little Black Mirror-esque. But people are receptive to personal relationships with AI, like ChatGPT as a friend and coach.” “I’m much more into the idea of ambient, single-purpose objects. For instance, a robot that lives in the laundry room, moving and folding your laundry.”

TBPN

42,713 次观看 • 10 个月前

Physical Intelligence (Physical Intelligence) is building a foundation model that can control any robot to do any task — what the team describes as the GPT moment for robotics. The company's cross-embodiment approach trains across many different robot platforms, and recent results show tasks being performed zero-shot that last year required hundreds of hours of data collection. In this episode of the Lightcone Podcast , co-founder Quan Vuong (Quan Vuong) sat down with Garry Tan, Jared Friedman, Diana, and Harj Taggar to talk about why robotics is finally ready for its scaling moment, how PI runs its models in the cloud rather than on-device, and the playbook for what Quan sees as a Cambrian explosion of vertical robotics companies. 00:00 — Robotics just got cheaper 00:41 — The GPT moment for robotics 02:24 — Why robots didn’t work before 05:30 — The breakthrough that changed everything 09:12 — The data problem 13:33 — Robots learning without data 15:05 — Robots folding laundry (for real) 22:18 — From engineering problem → ops problem 29:12 — The startup playbook 38:46 — Thousands of robotics startups are coming

Physical Intelligence (Physical Intelligence) is building a foundation model that can control any robot to do any task — what the team describes as the GPT moment for robotics. The company's cross-embodiment approach trains across many different robot platforms, and recent results show tasks being performed zero-shot that last year required hundreds of hours of data collection. In this episode of the Lightcone Podcast , co-founder Quan Vuong (Quan Vuong) sat down with Garry Tan, Jared Friedman, Diana, and Harj Taggar to talk about why robotics is finally ready for its scaling moment, how PI runs its models in the cloud rather than on-device, and the playbook for what Quan sees as a Cambrian explosion of vertical robotics companies. 00:00 — Robotics just got cheaper 00:41 — The GPT moment for robotics 02:24 — Why robots didn’t work before 05:30 — The breakthrough that changed everything 09:12 — The data problem 13:33 — Robots learning without data 15:05 — Robots folding laundry (for real) 22:18 — From engineering problem → ops problem 29:12 — The startup playbook 38:46 — Thousands of robotics startups are coming

Y Combinator

115,590 次观看 • 3 个月前

John: "Over in China, they’re doing robot Olympics. They’re doing marathons and all sorts of stuff." Figure founder Brett Adcock: “We look at the distribution of what humans do that is useful, and we [designed our humanoid] to do as much of that as possible.” “These things where we run marathons, or we do backflips, or we do karate moves—they’re not in the main part of the distribution. They don’t matter. And if you really want to size for those, you’re going to build a really expensive and unsafe robot that’s hard to manufacture.” “People want the $10,000 humanoid that can do general-purpose work. We want it to do laundry and dishes and be a companion. I want to ship robots at scale into the workforce to do logistics and healthcare, and build buildings and data centers.” "I don't need [our humanoid] to do backflips to do any of that work. I want robots building other robots."

John: "Over in China, they’re doing robot Olympics. They’re doing marathons and all sorts of stuff." Figure founder Brett Adcock: “We look at the distribution of what humans do that is useful, and we [designed our humanoid] to do as much of that as possible.” “These things where we run marathons, or we do backflips, or we do karate moves—they’re not in the main part of the distribution. They don’t matter. And if you really want to size for those, you’re going to build a really expensive and unsafe robot that’s hard to manufacture.” “People want the $10,000 humanoid that can do general-purpose work. We want it to do laundry and dishes and be a companion. I want to ship robots at scale into the workforce to do logistics and healthcare, and build buildings and data centers.” "I don't need [our humanoid] to do backflips to do any of that work. I want robots building other robots."

TBPN

15,419 次观看 • 5 个月前

We are thinking of robots all wrong. Why a $1,500 robot is far more important to buy than a $20,000 one. And why it will pay for itself within a year. First of all, for the last eight years I've been a Silicon Valley housewife. Picking the kids up from school. Doing a variety of tasks taking care of them from feeding them to laundry. And I've already bought a $20,000 Neo from 1X and have built the most complete list on X of the robotics industry. Just to set the tone for this conversation. We must ask ourselves "what is the goal of a robot?" before we go into why typical American homeowners might want one, and shell out quite a bit of money for one, like the Neo. I grew up in Silicon Valley back when it was all orchards and the farmers taught me "pick the low hanging fruit first." What is the low hanging fruit in the American home? Laundry? Cleaning the toilets or your home? Watering the plants? Bringing you a beer? Nope. It is the preparation of food. Yesterday I got a new Posha robot, and the attached video with founder Raghav Gupta gets into depth about what the $1,500 robot does. Cooks meals. Far more time in the home is spent cooking meals than the other tasks and is far more complex than, say, folding laundry. But there is something I think everyone is missing in the discussion of robots: "what is the goal?" I've been doing consumer research talking with many around the world about these things. People tend to have a few goals: 1. Improve their lives. 2. Save them time. 3. Save them money. 4. Enable a new business. What is the best way to improve your life? Upgrade your food. This is very hard to do when both parents in a family are working their butts off to try to improve their careers. It gets worse when a single parent is trying to keep everything going. How many times have you decided to go out to eat rather than spend an hour cooking food? Doing that for a family of four in Silicon Valley costs $100+. And guarantees your family will overeat. I've done that many times while raising my kids, and often I can't say no when desert comes around. It gets worse if you take the easy route out at home. Put a pre-processed meal into the microwave, or heat up a frozen pizza. Horrible for everyone's health. But what if you could have a robot at home that cooks your meals? Then costs go down to less than $20 and ingredients get way way better. It gets worse when you consider a $20,000 humanoid. They aren't safe enough to trust around stoves yet. And their hands aren't yet dexterous enough to do that. I doubt my Neo will be allowed to cook meals over an open flame, and if so I will have to watch it to make sure it doesn't do anything wrong. (The Neo that arrives next year will be teleoperated by a human remotely and the risks that person does something, or misses oil catching on fire is just way too high). While neither robot will be able to do all food preparation (cutting chicken up into cubes, or cutting carrots or other fruits, for instance) this robot dramatically reduces the time needed for a human to make a meal and dramatically reduces the costs to do so. And, as we discuss in the video, when the Neo does arrive the Neo will be able to use this machine too, reducing time even more (and will be able to set the table and wash the dishes, saving even more time so you can answer more emails or learn more AI programs or, even, pay attention to your kids and give them a few more minutes of quality time). The robot industry should focus on the low hanging fruit first. Cooking meals is the biggest one to improve your life, save you time, and make your family healthier. It's why I bought one. And they actually make two: Your money is way better spent getting one of these than buying a humanoid. And if you do get a humanoid, like I am, they go together like peanut butter and jelly. ⚪️ sierra catalina has been saying this for years that our focus on humanoids is overblown and that specialized robots (you see my Matic Robots in the background to prove this point) are way better for most families.

We are thinking of robots all wrong. Why a $1,500 robot is far more important to buy than a $20,000 one. And why it will pay for itself within a year. First of all, for the last eight years I've been a Silicon Valley housewife. Picking the kids up from school. Doing a variety of tasks taking care of them from feeding them to laundry. And I've already bought a $20,000 Neo from 1X and have built the most complete list on X of the robotics industry. Just to set the tone for this conversation. We must ask ourselves "what is the goal of a robot?" before we go into why typical American homeowners might want one, and shell out quite a bit of money for one, like the Neo. I grew up in Silicon Valley back when it was all orchards and the farmers taught me "pick the low hanging fruit first." What is the low hanging fruit in the American home? Laundry? Cleaning the toilets or your home? Watering the plants? Bringing you a beer? Nope. It is the preparation of food. Yesterday I got a new Posha robot, and the attached video with founder Raghav Gupta gets into depth about what the $1,500 robot does. Cooks meals. Far more time in the home is spent cooking meals than the other tasks and is far more complex than, say, folding laundry. But there is something I think everyone is missing in the discussion of robots: "what is the goal?" I've been doing consumer research talking with many around the world about these things. People tend to have a few goals: 1. Improve their lives. 2. Save them time. 3. Save them money. 4. Enable a new business. What is the best way to improve your life? Upgrade your food. This is very hard to do when both parents in a family are working their butts off to try to improve their careers. It gets worse when a single parent is trying to keep everything going. How many times have you decided to go out to eat rather than spend an hour cooking food? Doing that for a family of four in Silicon Valley costs $100+. And guarantees your family will overeat. I've done that many times while raising my kids, and often I can't say no when desert comes around. It gets worse if you take the easy route out at home. Put a pre-processed meal into the microwave, or heat up a frozen pizza. Horrible for everyone's health. But what if you could have a robot at home that cooks your meals? Then costs go down to less than $20 and ingredients get way way better. It gets worse when you consider a $20,000 humanoid. They aren't safe enough to trust around stoves yet. And their hands aren't yet dexterous enough to do that. I doubt my Neo will be allowed to cook meals over an open flame, and if so I will have to watch it to make sure it doesn't do anything wrong. (The Neo that arrives next year will be teleoperated by a human remotely and the risks that person does something, or misses oil catching on fire is just way too high). While neither robot will be able to do all food preparation (cutting chicken up into cubes, or cutting carrots or other fruits, for instance) this robot dramatically reduces the time needed for a human to make a meal and dramatically reduces the costs to do so. And, as we discuss in the video, when the Neo does arrive the Neo will be able to use this machine too, reducing time even more (and will be able to set the table and wash the dishes, saving even more time so you can answer more emails or learn more AI programs or, even, pay attention to your kids and give them a few more minutes of quality time). The robot industry should focus on the low hanging fruit first. Cooking meals is the biggest one to improve your life, save you time, and make your family healthier. It's why I bought one. And they actually make two: Your money is way better spent getting one of these than buying a humanoid. And if you do get a humanoid, like I am, they go together like peanut butter and jelly. ⚪️ sierra catalina has been saying this for years that our focus on humanoids is overblown and that specialized robots (you see my Matic Robots in the background to prove this point) are way better for most families.

Robert Scoble

178,120 次观看 • 8 个月前

$I saw a glimpse of the robots at the ongoing International Cultural Industries Fair in Shenzhen — playing piano, competing at chess, dancing, boxing, playing ping-pong, and even brewing fresh coffee. Yet these entertaining performances are only a small fraction of what they can do. Robots are everywhere in Shenzhen — patrolling subway stations, serving popcorn in cinemas, moving materials on production lines, even folding laundry in eldercare. A robust industry powers such widespread adoption: over 1,000 core robotics companies, a 200 billion RMB industrial chain, and nearly all components available locally — half within 10 km, 80% within 40 km. Shenzhen even hosts the world’s first robot 6S stores. From production lines to culture fairs, robots have quietly woven into daily life — yet seeing them perform live, with creativity and charm, still amazes.$

I saw a glimpse of the robots at the ongoing International Cultural Industries Fair in Shenzhen — playing piano, competing at chess, dancing, boxing, playing ping-pong, and even brewing fresh coffee. Yet these entertaining performances are only a small fraction of what they can do. Robots are everywhere in Shenzhen — patrolling subway stations, serving popcorn in cinemas, moving materials on production lines, even folding laundry in eldercare. A robust industry powers such widespread adoption: over 1,000 core robotics companies, a 200 billion RMB industrial chain, and nearly all components available locally — half within 10 km, 80% within 40 km. Shenzhen even hosts the world’s first robot 6S stores. From production lines to culture fairs, robots have quietly woven into daily life — yet seeing them perform live, with creativity and charm, still amazes.

Aurora Tan

650,995 次观看 • 1 个月前

Robotics is almost solved! Pun intended. Sunday published a "Solve" standard. They declared the scope (garment types, lighting, surfaces, starting configurations) and adaptation cost (zero—no fine-tuning per home) before evaluating. The same model weights run everywhere. This might be the differentiation between a good demo and reliable work of robots. They presented ACT-2, a robot foundation model that folds laundry at 99.1% success across unseen homes. Different houses, lighting, bed types, garment sizes. According to them scaling pre-training closes the generalization gap. Train on massive, high-quality human data. When you do that right, what you learn in-house actually transfers to unseen homes. Even wilder is that the model learns new folding techniques from a single example and generalizes them to new environments. Memo goes to homes later this year. Would you adopt this cutie? 🥹 Read more here: Great work Tony Zhao and team! 🧡 ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Robotics is almost solved! Pun intended. Sunday published a "Solve" standard. They declared the scope (garment types, lighting, surfaces, starting configurations) and adaptation cost (zero—no fine-tuning per home) before evaluating. The same model weights run everywhere. This might be the differentiation between a good demo and reliable work of robots. They presented ACT-2, a robot foundation model that folds laundry at 99.1% success across unseen homes. Different houses, lighting, bed types, garment sizes. According to them scaling pre-training closes the generalization gap. Train on massive, high-quality human data. When you do that right, what you learn in-house actually transfers to unseen homes. Even wilder is that the model learns new folding techniques from a single example and generalizes them to new environments. Memo goes to homes later this year. Would you adopt this cutie? 🥹 Read more here: Great work Tony Zhao and team! 🧡 ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Lukas Ziegler

35,818 次观看 • 4 天前

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