Please open Telegram to view this post
VIEW IN TELEGRAM
π7π1π1
Please open Telegram to view this post
VIEW IN TELEGRAM
π21π4π€£2
This media is not supported in your browser
VIEW IN TELEGRAM
An AgiBot Lingxi X2 robot pulled off a silky-smooth dance to a Chinese pop hit in Guangzhou, China. From tech to rhythm, innovation meets fun.
Please open Telegram to view this post
VIEW IN TELEGRAM
π9β€7π5
Please open Telegram to view this post
VIEW IN TELEGRAM
β€12π7π―1
Media is too big
VIEW IN TELEGRAM
Elon Musk says working will be 'optional' in less than 20 years because of AI and robotics.
π€ @techdaily
Please open Telegram to view this post
VIEW IN TELEGRAM
π10π7π4π€2π¨βπ»2π1π€1
This media is not supported in your browser
VIEW IN TELEGRAM
Checked out the Sanitation Robot Competition in Shenzhen last week.
It was amazing to see so many types of cleaning robots for streets and roads, a perfect use case for service robots.
π€ @techdaily
It was amazing to see so many types of cleaning robots for streets and roads, a perfect use case for service robots.
Please open Telegram to view this post
VIEW IN TELEGRAM
β€12π5
This media is not supported in your browser
VIEW IN TELEGRAM
Multi-robot learning is getting a serious boost! π
Researchers have extended Isaac Lab to train heterogeneous multi-agent robotic policies at scale. The new framework supports high-resolution physics, GPU-accelerated simulation, and both homogeneous and heterogeneous agents working together on coordination tasks.
They benchmarked different approaches (MAPPO: Multi-Agent Proximal Policy Optimization and HAPPO: Heterogeneous Agent PPO) across six challenging scenarios and showed that large-scale multi-robot training is not only feasible, but efficient. Itβs an important step for real-world robotic collaboration, where teams of robots need to coordinate, split tasks, adapt roles, and interact dynamically, not just operate as identical clones.
The code is open-source, and it pushes Isaac Lab closer to what robotics actually needs: scalable, physics-driven environments where many different robots can learn to work together.
π€ @techdaily
Researchers have extended Isaac Lab to train heterogeneous multi-agent robotic policies at scale. The new framework supports high-resolution physics, GPU-accelerated simulation, and both homogeneous and heterogeneous agents working together on coordination tasks.
They benchmarked different approaches (MAPPO: Multi-Agent Proximal Policy Optimization and HAPPO: Heterogeneous Agent PPO) across six challenging scenarios and showed that large-scale multi-robot training is not only feasible, but efficient. Itβs an important step for real-world robotic collaboration, where teams of robots need to coordinate, split tasks, adapt roles, and interact dynamically, not just operate as identical clones.
The code is open-source, and it pushes Isaac Lab closer to what robotics actually needs: scalable, physics-driven environments where many different robots can learn to work together.
Please open Telegram to view this post
VIEW IN TELEGRAM
β€11π3
Elon Musk expects Tesla's Optimus robots to build themselves in self replicating factories and make most work optional within 20 years
π€ @techdaily
Please open Telegram to view this post
VIEW IN TELEGRAM
Please open Telegram to view this post
VIEW IN TELEGRAM
π10 4β€2π1π€1
This media is not supported in your browser
VIEW IN TELEGRAM
This print appears to be happening in midair⦠a new method of 3D printing:
It combines direct ink writing with up-conversion particles-assisted photopolymerization. With this method of printing ceramics, the need for support structures is eliminated.
π€ @techdaily
It combines direct ink writing with up-conversion particles-assisted photopolymerization. With this method of printing ceramics, the need for support structures is eliminated.
Please open Telegram to view this post
VIEW IN TELEGRAM
π13β€4π4
Please open Telegram to view this post
VIEW IN TELEGRAM
β€17π2π2π1π1
This media is not supported in your browser
VIEW IN TELEGRAM
Please open Telegram to view this post
VIEW IN TELEGRAM
β€13π3π2
This media is not supported in your browser
VIEW IN TELEGRAM
Xiaomi just realized a smartphone with a detachable pro-grade camera lens.
It is quite brilliantβ¦
π€ @techdaily
It is quite brilliantβ¦
Please open Telegram to view this post
VIEW IN TELEGRAM
β€13π6π4π€©3π€£1
Media is too big
VIEW IN TELEGRAM
Samsung Galaxy TriFold hands on video π₯
- 10β 120Hz inner display, 6.5β cover
- 3.9 mm (unfolded), 12.9 mm (folded)
- 309g weight
- Snapdragon 8 Elite
- 5600mAh + 45W
The multitasking on that big inner screen looks crazy good. Honestly, it looks great!
π€ @techdaily
- 10β 120Hz inner display, 6.5β cover
- 3.9 mm (unfolded), 12.9 mm (folded)
- 309g weight
- Snapdragon 8 Elite
- 5600mAh + 45W
The multitasking on that big inner screen looks crazy good. Honestly, it looks great!
Please open Telegram to view this post
VIEW IN TELEGRAM
β€24π₯8π2π1
Media is too big
VIEW IN TELEGRAM
A 17-year-old just built a mind-controlled prosthetic arm for $300.
Yes, $300. For something that usually costs $450,000. A teenager, working from home, used AI, cheap materials, and 23,000 lines of code to build a device that reads brain signals without surgery, without implants, and without a $450K price tag.
π€ @techdaily
Yes, $300. For something that usually costs $450,000. A teenager, working from home, used AI, cheap materials, and 23,000 lines of code to build a device that reads brain signals without surgery, without implants, and without a $450K price tag.
Please open Telegram to view this post
VIEW IN TELEGRAM
π31β€11π10π3π2
This media is not supported in your browser
VIEW IN TELEGRAM
China has officially entered the era of flying taxis. Two Chinese companies have obtained a commercial operation certificate for autonomous passenger drones from the CAAC, the Civil Aviation Administration of China.
π€ @techdaily
Please open Telegram to view this post
VIEW IN TELEGRAM
π13β€4π2
Please open Telegram to view this post
VIEW IN TELEGRAM
π14β€9π3π2
Trainium3 UltraServers:
β Up to 4.4Γ higher performance
β ~4Γ more memory bandwidth
β 4Γ better performance-per-watt vs Trainium2 UltraServers
β Scale to 144 chips (362 FP8 PFLOPs) per server and 1M+ chips in UltraClusters
AWS is positioning Trainium3 as a lower-cost alternative to Nvidia GPUs. It says Trainium-based instances can cut training and often inference costs by as much as 50% vs comparable GPU clusters on large models, for customers willing to adopt its Neuron software stack and Trainium instances.
Please open Telegram to view this post
VIEW IN TELEGRAM
π13β€7β€βπ₯1
This media is not supported in your browser
VIEW IN TELEGRAM
The drone is integrated directly into the vehicleβs roof. It can ascend to 200 meters and put out fires in areas unreachable for conventional fire engines. The copter deploys in under a minute and sprays water up to 45 meters. According to the developers, the system will allow for much faster response to fires in tall structures.
Deployment of these vehicles will begin this month.
Please open Telegram to view this post
VIEW IN TELEGRAM
π24β€1
Media is too big
VIEW IN TELEGRAM
Nvidia CEO Jensen Huang said:
"We create technology out of nothing. Our final product is pure technology"
π€ @techdaily
"We create technology out of nothing. Our final product is pure technology"
Please open Telegram to view this post
VIEW IN TELEGRAM
π13β€2
Media is too big
VIEW IN TELEGRAM
Nvidia CEO Jensen Huang predicts 90% of the world's knowledge will be AI-generated within 2-3 years .
π€ @techdaily
Please open Telegram to view this post
VIEW IN TELEGRAM
π11 7π2π€1