Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Most robot demos are scripted. Generalist's GEN-1 is not. > GEN-1 was doing a task > Mid-task, an extra object was thrown into the bin > GEN-1 quickly adapted to the new environment > Still completed the task smoothly This is the difference between a scripted policy (seen in... show more

mete

44,429 subscribers

11,311 views • 1 month ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

That's one of the coolest robotics demos! 😮‍💨 Generalist showed GEN-1 handling box folding and screw packing during Automate Show. The boxes are cardboard with real variability: creasing, deformation, different configurations. GEN-1 retries when things go wrong. It adapts mid-task. This is a jump from GEN-0, shown 3 months ago during GTC. GEN-0 handled rigid, predictable boxes. GEN-1 handles deformable ones with chaotic behavior. The scaling is straightforward. GEN-1 trained on 500,000 hours of proprietary, real-world dexterous manipulation data. Scaling laws are working. More high-quality real-world data leads to better physics understanding, better retry behavior, better generalization. And the time to get a working demo is shrinking drastically. They're not even at scale yet. They're just starting to see what happens when you pre-train on massive amounts of actual robot data. ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

That's one of the coolest robotics demos! 😮‍💨 Generalist showed GEN-1 handling box folding and screw packing during Automate Show. The boxes are cardboard with real variability: creasing, deformation, different configurations. GEN-1 retries when things go wrong. It adapts mid-task. This is a jump from GEN-0, shown 3 months ago during GTC. GEN-0 handled rigid, predictable boxes. GEN-1 handles deformable ones with chaotic behavior. The scaling is straightforward. GEN-1 trained on 500,000 hours of proprietary, real-world dexterous manipulation data. Scaling laws are working. More high-quality real-world data leads to better physics understanding, better retry behavior, better generalization. And the time to get a working demo is shrinking drastically. They're not even at scale yet. They're just starting to see what happens when you pre-train on massive amounts of actual robot data. ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Lukas Ziegler

29,587 views • 1 month ago

In my first week at Generalist, I trained a robot to pour liquids using GEN-1 🤖💧 I wanted to challenge the robot with a non-rigid manipulation task, so liquid felt like the perfect choice. The task involved: - unscrewing the bottle cap - pouring liquid into espresso glasses - rebalancing uneven pours Best of all, the robot was able to complete the task fully autonomously 3 times in a row (out of 3)! Pour-fect 😉 Excited for the journey ahead and grateful to be building alongside such an incredible team!

In my first week at Generalist, I trained a robot to pour liquids using GEN-1 🤖💧 I wanted to challenge the robot with a non-rigid manipulation task, so liquid felt like the perfect choice. The task involved: - unscrewing the bottle cap - pouring liquid into espresso glasses - rebalancing uneven pours Best of all, the robot was able to complete the task fully autonomously 3 times in a row (out of 3)! Pour-fect 😉 Excited for the journey ahead and grateful to be building alongside such an incredible team!

Kylie Ying

37,327 views • 2 months ago

🤖What if a robot could perform a new task just from a natural language command, with zero demonstrations? Our new work, NovaFlow, makes it possible! We use pre-trained video generative model to create a video of the task, then translate it into a plan for real-world robot execution. 1/6 #Robotics #AI #ZeroShot #Manipulation

🤖What if a robot could perform a new task just from a natural language command, with zero demonstrations? Our new work, NovaFlow, makes it possible! We use pre-trained video generative model to create a video of the task, then translate it into a plan for real-world robot execution. 1/6 #Robotics #AI #ZeroShot #Manipulation

Hongyu Li

105,628 views • 10 months ago

Xiaomi-Robotics-1 just dropped on Hugging Face 🔥 A robot foundation model trained on 100,000 hours of real-world manipulation. They turned it loose in a real apartment: folding laundry, loading the washer, doing the dishes, packing a suitcase. Fully autonomous.

Xiaomi-Robotics-1 just dropped on Hugging Face 🔥 A robot foundation model trained on 100,000 hours of real-world manipulation. They turned it loose in a real apartment: folding laundry, loading the washer, doing the dishes, packing a suitcase. Fully autonomous.

Victor M

130,344 views • 18 days ago

We ran a live demo NVIDIA GTC last week, but the real story is how quickly we got it running. The system was up and running in days, not weeks. This is a step toward robots that can be deployed quickly without task-by-task programming. How we made it happen👇 🧵 (1/6)

We ran a live demo NVIDIA GTC last week, but the real story is how quickly we got it running. The system was up and running in days, not weeks. This is a step toward robots that can be deployed quickly without task-by-task programming. How we made it happen👇 🧵 (1/6)

Generalist

41,800 views • 4 months ago

Which gen NASCAR is best and why is it obviously not Gen 1, 2, 4, 5, or 6. And yeah, they really did rebuild an entire engine at 1 a.m. in a hotel room so it was ready for the race the next morning

Which gen NASCAR is best and why is it obviously not Gen 1, 2, 4, 5, or 6. And yeah, they really did rebuild an entire engine at 1 a.m. in a hotel room so it was ready for the race the next morning

PATINA RESEARCH

50,970 views • 7 months ago

GEN-1 plays the 🐚 shell game, trained on just 1 hr of robot data. It also generalizes to unseen objects, like berkay 's car keys. Physical AI models should be capable of benchmark tasks like this one. It's interesting for the all the reasons Rhoda AI calls out -- requires visual memory, and the model must track the cups from the very start, at high frame rates. Interestingly, GEN-1 appears to exhibit a degree of "active perception." It's subtle; the hands can sometimes appear to "follow" the cups, using its own movements to help attend to where it thinks the object should be. Read more about GEN-1 in our blog post in the comments below ↓

GEN-1 plays the 🐚 shell game, trained on just 1 hr of robot data. It also generalizes to unseen objects, like berkay 's car keys. Physical AI models should be capable of benchmark tasks like this one. It's interesting for the all the reasons Rhoda AI calls out -- requires visual memory, and the model must track the cups from the very start, at high frame rates. Interestingly, GEN-1 appears to exhibit a degree of "active perception." It's subtle; the hands can sometimes appear to "follow" the cups, using its own movements to help attend to where it thinks the object should be. Read more about GEN-1 in our blog post in the comments below ↓

Generalist

859,713 views • 3 months ago

Watching Generalist AI demonstrate real-time intuition is incredible! 🤖✨ Check out this "Physical Commonsense" in action: the robot flips a washer, picks it up and then double-nudges it perfectly into a tight foam slot. This isn't a scripted motion; it’s an emergent behavior from a model trained on 270,000+ hours of real-world interaction. We are seeing the era where robots are starting to intuit and react to physics.

Watching Generalist AI demonstrate real-time intuition is incredible! 🤖✨ Check out this "Physical Commonsense" in action: the robot flips a washer, picks it up and then double-nudges it perfectly into a tight foam slot. This isn't a scripted motion; it’s an emergent behavior from a model trained on 270,000+ hours of real-world interaction. We are seeing the era where robots are starting to intuit and react to physics.

Humanoids daily

14,762 views • 6 months ago

Disruptors, we have something for you today. Something no human has ever seen before. In order for us to be in this dimension, there was a need for a physical form. It was the Lord of the Dimensions who was responsible for this task. Below is a peek at the master at work.

Disruptors, we have something for you today. Something no human has ever seen before. In order for us to be in this dimension, there was a need for a physical form. It was the Lord of the Dimensions who was responsible for this task. Below is a peek at the master at work.

Synthesis

20,977 views • 3 years ago

'This has nothing to do with Brown, Cat!' Deputy National Commissioner Lt Gen Shadrack Sibiya has told the Ad Hoc Committee in Parliament that Brown Mogotsi and Cat Matlala knew about the disbanding of the SAPS Political Killings Task Team from the media, claiming the directive signed by Police Minister Senzo Mchunu was leaked on January 1, 2025.

'This has nothing to do with Brown, Cat!' Deputy National Commissioner Lt Gen Shadrack Sibiya has told the Ad Hoc Committee in Parliament that Brown Mogotsi and Cat Matlala knew about the disbanding of the SAPS Political Killings Task Team from the media, claiming the directive signed by Police Minister Senzo Mchunu was leaked on January 1, 2025.

IOL News

44,142 views • 9 months ago

GEN-1 still works with lights off, and generalizes under harsh lighting conditions. The model uses raw video pixels to make decisions, so strong lighting changes can drastically alter its input distribution. Yet performance still holds. Why? GEN-1 was pre-trained on a massive, diverse dataset of different lighting conditions—everywhere from outdoor farms, to warehouses, from grocery stores, to dimly lit homes—it's already seen it all, and transfers this knowledge to new tasks. This is a glimpse of what we call Mastery, and is part of the reason these models can cross a new performance threshold. Read more about it in our blog post in the comments below 👇

GEN-1 still works with lights off, and generalizes under harsh lighting conditions. The model uses raw video pixels to make decisions, so strong lighting changes can drastically alter its input distribution. Yet performance still holds. Why? GEN-1 was pre-trained on a massive, diverse dataset of different lighting conditions—everywhere from outdoor farms, to warehouses, from grocery stores, to dimly lit homes—it's already seen it all, and transfers this knowledge to new tasks. This is a glimpse of what we call Mastery, and is part of the reason these models can cross a new performance threshold. Read more about it in our blog post in the comments below 👇

Generalist

18,396 views • 3 months ago

Generalist AI has just broken the "impossible triangle" of speed, reliability, and intelligence in robotics. They’ve pushed their embodied foundation model forward in a big way: GEN-1 hits “Mastery” on simple physical tasks,99% average success rate, and it can improvise when things go wrong. Smooth, natural, fast(factory owners must be very interested)👇 >Folding boxes: 200 runs at 99% success in ~12 seconds ( ~34s before) >Packing phones, folding T-shirts, servicing robot vacuums,hundreds of reps with almost zero failures > Intelligent improvisation, handling uncertain tasks >Needs only about 1 hour of robot-specific data per task This feels like a real step toward humanoid robots that are actually reliable enough for factories, warehouse or homes. Physical AGI is accelerating

Generalist AI has just broken the "impossible triangle" of speed, reliability, and intelligence in robotics. They’ve pushed their embodied foundation model forward in a big way: GEN-1 hits “Mastery” on simple physical tasks,99% average success rate, and it can improvise when things go wrong. Smooth, natural, fast(factory owners must be very interested)👇 >Folding boxes: 200 runs at 99% success in ~12 seconds ( ~34s before) >Packing phones, folding T-shirts, servicing robot vacuums,hundreds of reps with almost zero failures > Intelligent improvisation, handling uncertain tasks >Needs only about 1 hour of robot-specific data per task This feels like a real step toward humanoid robots that are actually reliable enough for factories, warehouse or homes. Physical AGI is accelerating

CyberRobo

25,519 views • 4 months ago

CHINA JUST SOLVED THE PROBLEM THAT'S BEEN BREAKING ROBOT AI FOR A DECADE. and the fix wasn't a smarter model. for years, every robot AI failure got the same diagnosis. the model isn't smart enough. so everyone scaled intelligence. bigger models. more parameters. better reasoning. AGIBOT asked a different question: what if the reasoning was never the problem? there's a gap that runs through every traditional robot AI system. reasoning on one side & motor commands on the other. the brain decides but the body executes something different, because thinking and moving were never actually connected. GO-2 fixes this by reasoning INSIDE the action space, not above it. before moving, it runs a complete mental simulation of every step - like a basketball player mentally tracing the arc of a shot before releasing the ball. watch the demo and you'll see exactly what this means. the robot works through a task queue autonomously. classify toiletries. upright the drink bottle. place headphones in the leather box. mid-execution, a new instruction drops: "my phone's missing. help me find it." it doesn't pause. doesn't reset. it processes the new task and keeps moving. that's not a scripted sequence. that's real-time instruction following on top of an active task queue. that one architectural change is where the numbers come from. > #1 on LIBERO across Spatial, Object, Goal, and Long tasks → 98.5% average success > 86.6% zero-shot accuracy in active disturbance environments > 47.4 on VLABench → best-in-class on objects and textures it's never seen before > 82.9% success trained on simulation only, tested on real hardware sim-to-real is the graveyard of robotics research. models trained in simulation collapse the moment they touch the real world. 82.9% means that graveyard just got a lot smaller. it holds because of how GO-2 trains. deliberately fed imperfect reasoning conditions, then trained to execute robustly anyway. not a researcher assumption. a design decision from a team that ships hardware and knows exactly what breaks. then there's the infrastructure layer. Genie Studio. fleet-wide data collection. cloud training. online post-training in live environments. 10x improvement in training efficiency. task startup reduced to minutes. 2-4x better success rates with 50%+ less data. the model gets smarter every time a robot fails in the field. this isn't a benchmark story. it's a compounding moat. dual CVPR 2026 + ACL 2026 acceptance. computer vision AND natural language processing. top conferences. simultaneously. that doesn't happen with incremental research. the US-China robotics race has been framed as a compute race. a model quality race. it was always an execution race. the robot that wins won't be the smartest one in the lab. it'll be the most reliable one on the floor. full breakdown: is execution reliability the real bottleneck, or are we still underestimating how far reasoning needs to go?

CHINA JUST SOLVED THE PROBLEM THAT'S BEEN BREAKING ROBOT AI FOR A DECADE. and the fix wasn't a smarter model. for years, every robot AI failure got the same diagnosis. the model isn't smart enough. so everyone scaled intelligence. bigger models. more parameters. better reasoning. AGIBOT asked a different question: what if the reasoning was never the problem? there's a gap that runs through every traditional robot AI system. reasoning on one side & motor commands on the other. the brain decides but the body executes something different, because thinking and moving were never actually connected. GO-2 fixes this by reasoning INSIDE the action space, not above it. before moving, it runs a complete mental simulation of every step - like a basketball player mentally tracing the arc of a shot before releasing the ball. watch the demo and you'll see exactly what this means. the robot works through a task queue autonomously. classify toiletries. upright the drink bottle. place headphones in the leather box. mid-execution, a new instruction drops: "my phone's missing. help me find it." it doesn't pause. doesn't reset. it processes the new task and keeps moving. that's not a scripted sequence. that's real-time instruction following on top of an active task queue. that one architectural change is where the numbers come from. > #1 on LIBERO across Spatial, Object, Goal, and Long tasks → 98.5% average success > 86.6% zero-shot accuracy in active disturbance environments > 47.4 on VLABench → best-in-class on objects and textures it's never seen before > 82.9% success trained on simulation only, tested on real hardware sim-to-real is the graveyard of robotics research. models trained in simulation collapse the moment they touch the real world. 82.9% means that graveyard just got a lot smaller. it holds because of how GO-2 trains. deliberately fed imperfect reasoning conditions, then trained to execute robustly anyway. not a researcher assumption. a design decision from a team that ships hardware and knows exactly what breaks. then there's the infrastructure layer. Genie Studio. fleet-wide data collection. cloud training. online post-training in live environments. 10x improvement in training efficiency. task startup reduced to minutes. 2-4x better success rates with 50%+ less data. the model gets smarter every time a robot fails in the field. this isn't a benchmark story. it's a compounding moat. dual CVPR 2026 + ACL 2026 acceptance. computer vision AND natural language processing. top conferences. simultaneously. that doesn't happen with incremental research. the US-China robotics race has been framed as a compute race. a model quality race. it was always an execution race. the robot that wins won't be the smartest one in the lab. it'll be the most reliable one on the floor. full breakdown: is execution reliability the real bottleneck, or are we still underestimating how far reasoning needs to go?

Shruti

18,622 views • 3 months ago

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

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

Lukas Ziegler

25,901 views • 3 months ago

Hasan thinks Ethan was written about too favorably in his GQ article “This is the most favorable interpretation of Ethan’s behavior who has both called it a gen*cide and not called it a gen*cide..I didn’t bring this up..journalist did..I would have not”

Hasan thinks Ethan was written about too favorably in his GQ article “This is the most favorable interpretation of Ethan’s behavior who has both called it a gencide and not called it a gencide..I didn’t bring this up..journalist did..I would have not”

yeet

46,574 views • 11 months ago

At Axis, every trajectory submitted by our community undergoes a strict replay validation process. We run each submission through checker to verify whether the target task was successfully completed. To see how strict it is, check this demo (Task: Place The Toy Train On The Board Game Box). Real human data passes smoothly (Video 1). However, bots or manually altered data will fail (Video 2). Why? Faking numbers breaks the simulation's physics causality. Even tiny tweaks cause error accumulation, resulting in failed movements. This invalid data is automatically rejected. Because of this mechanism, data submitted via bots will ultimately fail our replay verification. Invalid data is strictly excluded from model training, and the task slot is reopened to the community to collect genuine, high-quality trajectories. Furthermore, we actively monitor for duplicated data. Trajectories that are identical lack the diversity required for robot learning and will not be credited by our scoring system. If we detect accounts submitting a massive volume of identical trajectories, all associated addresses will be permanently banned. For Axis, the quality and diversity of data are the only ways to solve the robotics generalization gap. They will always be our absolute top priorities.

At Axis, every trajectory submitted by our community undergoes a strict replay validation process. We run each submission through checker to verify whether the target task was successfully completed. To see how strict it is, check this demo (Task: Place The Toy Train On The Board Game Box). Real human data passes smoothly (Video 1). However, bots or manually altered data will fail (Video 2). Why? Faking numbers breaks the simulation's physics causality. Even tiny tweaks cause error accumulation, resulting in failed movements. This invalid data is automatically rejected. Because of this mechanism, data submitted via bots will ultimately fail our replay verification. Invalid data is strictly excluded from model training, and the task slot is reopened to the community to collect genuine, high-quality trajectories. Furthermore, we actively monitor for duplicated data. Trajectories that are identical lack the diversity required for robot learning and will not be credited by our scoring system. If we detect accounts submitting a massive volume of identical trajectories, all associated addresses will be permanently banned. For Axis, the quality and diversity of data are the only ways to solve the robotics generalization gap. They will always be our absolute top priorities.

Axis Robotics

30,876 views • 3 months ago

I placed 🥈 2nd in the LeHome Challenge (at IEEE ICRA 2026) earlier this month, and before that I was 🥇 1st of 62 teams in the simulation round. Now I am sharing my solution — with a detailed logic walkthrough and open-source code. The task was to teach a cheap two-armed robot to fold different garments in simulation and on a real robot. I trained a VLA policy with an RL loop to make it work. Let's break it down 👇

I placed 🥈 2nd in the LeHome Challenge (at IEEE ICRA 2026) earlier this month, and before that I was 🥇 1st of 62 teams in the simulation round. Now I am sharing my solution — with a detailed logic walkthrough and open-source code. The task was to teach a cheap two-armed robot to fold different garments in simulation and on a real robot. I trained a VLA policy with an RL loop to make it work. Let's break it down 👇

Ilia

21,948 views • 1 month ago

🚨 BREAKING: this new video gen model is so cracked Hollywood-quality shots 1 minute long from a single prompt i've tested every major video model that dropped this year. sora, kling, runway, all of them. none of them made me stop and rewatch the same clip 6 times. (genuinely thought it was real footage for a second)

🚨 BREAKING: this new video gen model is so cracked Hollywood-quality shots 1 minute long from a single prompt i've tested every major video model that dropped this year. sora, kling, runway, all of them. none of them made me stop and rewatch the same clip 6 times. (genuinely thought it was real footage for a second)

Robin Delta

25,168 views • 4 months ago

Gen Z protesters in Mexico are demanding an end to impunity for gang violence and calling for justice for the mayor of Uruapan, Carlos Manzo, who was assassinated at a Day of the Dead event on November 1.

Gen Z protesters in Mexico are demanding an end to impunity for gang violence and calling for justice for the mayor of Uruapan, Carlos Manzo, who was assassinated at a Day of the Dead event on November 1.

Al Jazeera English

27,540 views • 8 months ago

🚨 OpenAI’s GPT-5.5 / GPT-5.5 Pro is being A/B tested inside ChatGPT! I created a 1:1 Windows OS clone and it delivered a surprisingly accurate recreation - the UI layout, structure, and interactions were all on point. This is one of the best real-world demos I’ve seen so far from this model. Watch the full breakdown + demo here:

🚨 OpenAI’s GPT-5.5 / GPT-5.5 Pro is being A/B tested inside ChatGPT! I created a 1:1 Windows OS clone and it delivered a surprisingly accurate recreation - the UI layout, structure, and interactions were all on point. This is one of the best real-world demos I’ve seen so far from this model. Watch the full breakdown + demo here:

WorldofAI

256,233 views • 3 months ago