Omar Technoprince
@robogeek_omar
مدمن التقنية، متحمس للذكاء الاصطناعي والذكاء الاصطناعي العام 🤖💡
As AI evolves, so do its storage demands. Unlike HPC’s large-file workflows, AI inference depends on moving millions of small files fast. Part 4 of our Future of Storage series explores object stores, metadata, and AI-ready storage. Learn more here: ow.ly/KN8h50XqhLb
Dunking robots in 2025! 🏀 Imagine playing with humanoids and one of them dunks and shouts 'KOBEEE'. Sounds like from the future, but looks like it might be closer than we think. That's a Unitree Robotics humanoid bot dunking, shooting, and blocking. Personally, I think that…
What makes Matic actually intelligent? +5 cameras + pure vision (no LiDAR) + Nvidia Jetson Orin Nano + Trained on infinite simulated homes + Eight years of relentless work = A robot that finally lives up to the dream. 🤖
Live 3D mapping at 1 cm³ resolution using ordinary cameras, with the entire autonomy stack running fully on-device. Here’s how we do it, in 60 seconds:
Nano Banana Pro 🤩 you can swap out a subject in a real photo with an animated character. check the replies for a reliable way to do this. also, just curious.. who did it best? 🏀😤
you can even transform iconic photos with this technique. 😁 nano banana is unbelievable. check out the post below for more tips:
btw,their maximum joint(PhyArc150)torque peak is 590 N·m. Tesla Model Y 2025 AWD: 559 N·m.
After reviewing PHYBOT's cycloidal joint tests, it's easy to understand why their humanoid robots possess such explosive power. One of their large joints boasts a peak torque of 530 N·m, exceeding even human strength. Even the smaller joints exhibit an astonishing torque density…
Update ComfyUI it’s already there. We are just finalizing the example workflows.
We are excited to unveil HunyuanVideo 1.5, the strongest open-source video generation model. Built upon DiT architecture, it redefines the open-source SOTA for accessibility and performance.🚀🚀🚀 HunyuanVideo 1.5 delivers state-of-the-art visual quality and motion coherence…
One of my favorite Nano Banana Pro tricks 🍌 It is so good with text that you can ask it to draw animation notes right on the image. Then you can send it to VEO 3.1 on Google Flow, ask it to follow the notes and clean them from the first frame. Perfect little workflow.
How is the motion capability? Can it compete with the Unitree G1? 🏹 MagicLab's Z1 agility is fully upgraded; the movements are smoother and more nimble! 🤸
Check out the MagicLab Z1 bipedal humanoid robot going "full throttle" on its dance moves. This bot is showcasing some seriously smooth motion control. The Z1 is a compact, smaller humanoid, roughly similar in size to the Unitree G1. It’s exciting to see another player pushing…
At the beginning of the summer we sit down and decided to do something bold: get visual sim2real to work on a humanoid Loco manipulation task. After 6 month of toil, here it is:
Zero teleoperation. Zero real-world data. ➔ Autonomous humanoid loco-manipulation in reality. Introducing VIRAL: Visual Sim-to-Real at Scale. We achieved 54 autonomous cycles (walk, stand, place, pick, turn) using a simple recipe: 1. RL 2. Simulation 3. GPUs Website:…
The real moats in 2025: specific workflows, proprietary data with real switching costs, distribution, and UX that makes AI disappear into the job-to-be-done. Simultaneously: we are early (only a % are using AI properly) so this is an amazing time to start a startup.
Benedict Evans' new presentation just dropped: "AI eats the world" 90 slides on macro and strategic trends in tech. His biannual overview is always worth the time: ben-evans.com/presentations
I always believe robot learning should happen both in position and impedance control, since that is how human does motion control, explained by the original author of impedance control Neville Hogan: m.youtube.com/watch?v=SHnCIu…
youtube.com
YouTube
Mastering the Skeleton and its Muscles by Neville Hogan
What if a humanoid robot could choose how to interact with the environment 🤖 — soft when it needs compliance, stiff when it needs precision, and force-aware when it must push/pull? That’s exactly what our Heterogeneous Meta-Control (HMC) framework enables. Our new framework…
Most robot learning has focused on simple position control. But think about how a human uses a wrench 🔧: you’re not just rotating in one direction—you’re continuously shaping the forces, pushing and pulling differently as you move. Our robot can do exactly that now.
What if a humanoid robot could choose how to interact with the environment 🤖 — soft when it needs compliance, stiff when it needs precision, and force-aware when it must push/pull? That’s exactly what our Heterogeneous Meta-Control (HMC) framework enables. Our new framework…
Well done to the SAM3D team. I haven't benchmarked it properly yet but this is definitely going to be the default backbone of choice for all 3d tasks by the back half of next year in the literature. What a time to be alive 🌞
1.5 years ago: - Google Robotics friends quit - "What is Tony Zhao doing?" - DM'ed - had lunch in Palo Alto downtown and did 3 hr walk (I talked about geopolitics/Japan, video models, why I left robotics but bullish) - wrote an angel cheque Zero robot data. Congrats @tonyzzhao!
Today, we present a step-change in robotic AI @sundayrobotics. Introducing ACT-1: A frontier robot foundation model trained on zero robot data. - Ultra long-horizon tasks - Zero-shot generalization - Advanced dexterity 🧵->
ACT-1 is a special model, it can read the map of your house to figure out where your furniture and appliances are positioned (e.g. dining table, dishwasher). this is crucial for zero-shot deployment into unseen homes. and it's all end-to-end!
It is even more fun to see how Memo reacts to unseen environments. We deploy it to 6 unseen Airbnbs and task the robot with fine-grained tasks such as picking up utensils from the plate. Because we train on data from over 500 homes, the new home is instantly familiar to Memo.
Some of my favorites on reward modeling, RL, and robust VLA from the community: arxiv.org/abs/2505.10911 from @Jesse_Y_Zhang arxiv.org/abs/2510.14830 from @kunlei15 arxiv.org/abs/2509.25358 from @QianzhongChen and pi0.6 from @physical_int today: physicalintelligence.company/blog/pistar06
KAUST Unveils Shaheen III as Middle East’s Most Powerful Supercomputer at SC25 @KAUST_News #SC25 #HPC ow.ly/h9c650XtAG1
The wait is over! The 2025 HPCwire Readers' and Editors' Choice Awards winners were just announced! See all the winners here: ow.ly/mhsN50XsZsR #SC25 #HPCwire #HPCwireRCA25
Return of the King. #FairyTaleTuesday #aivideo #aivideos #grokimagine #googlegemini #perchanceai #fantasy
Ousama AlHaj, General Manager and Head of Human Pharma at Boehringer Ingelheim – Near East & United Arab Emirates (NEUAE); Yosra Badr, Managing Director of Seha Healthcare Group & Co-Founder of Grinta (Egypt); and Dr Mahmoud Farrag, CEO & Co-Founder of Parkville Pharmaceuticals…
United States Trends
- 1. Luka 59.4K posts
- 2. Clippers 17.3K posts
- 3. Lakers 46.1K posts
- 4. #DWTS 94.2K posts
- 5. Dunn 6,464 posts
- 6. #LakeShow 3,438 posts
- 7. Robert 133K posts
- 8. Kawhi 6,036 posts
- 9. Reaves 11.1K posts
- 10. Jaxson Hayes 2,332 posts
- 11. Ty Lue 1,526 posts
- 12. Alix 15K posts
- 13. Collar 42K posts
- 14. Elaine 45.9K posts
- 15. Jordan 115K posts
- 16. Zubac 2,272 posts
- 17. Ezra 14.8K posts
- 18. NORMANI 6,191 posts
- 19. Dylan 34.2K posts
- 20. Colorado State 2,379 posts
Something went wrong.
Something went wrong.