Kai speaks 14 languages at a diplomatic reception
Kai was invited to a diplomatic reception at the Fairmont. He spoke 14 of his more than 50 languages and also practiced some new material from a stand-up routine. No one told him the dress code.
Kai was invited to a diplomatic reception at the Fairmont. He spoke 14 of his more than 50 languages and also practiced some new material from a stand-up routine. No one told him the dress code.
Dongfang Precision and Leju Robotics have joined forces to build China's first 10,000-unit-class humanoid robot production line; details of the line were disclosed in the news section of Dongfang Precision's official website. Dongfang Precision lists embodied intelligent robot equipment as one of its main businesses, alongside its intelligent packaging equipment and intelligent marine equipment.
Greatness is in the details—precision, balance, and feel. The best products seem simple but are built to exacting standards. At Hexagon, Matthias Wieser leads our Design Center, creating technology experiences that are both practical and delightful. Learn more: https://hxgn.biz/4v5L1au
Memo appeared on Nasdaq today!
Sourccey is coming soon! April 2026 😊
Let him cook.
Physical Intelligence has added an RL token output to its latest model, π-0.6, which is learned rapidly through reinforcement learning by a very small actor and critic, allowing the model to be fine-tuned for precise tasks in hours or even minutes without training the entire model.
At Your Service, final episode: dinner
At Your Service, Episode 4: Bathroom cleaning + folding clothes
SOSV's HAX will host a Deep Tech Showcase in New York on Wednesday, April 1, 18:00–20:00 ET, featuring more than 15 seed and Series A startups across robotics, energy, advanced computing, health, manufacturing, and space.
At Your Service, Episode 3: Kitchen Edition
At Your Service, Episode 2: Washer/Dryer
18,000+ people. Nine countries. One question: what will spark the next leap in robotics? 🤖💥 Our Robotics Generation study shows that attitudes toward robots differ around the world. What will determine how quickly robots enter our lives is people, not cost or technology. https://hxgn.biz/4ugEYQd
'At Your Service' Episode 1: Good Morning
Dong Mengmeng, founder and CEO of Hangzhou Dewang Nano Technology, won first prize in the finals of the 5th "Hai Ju Ying Cai" Global Innovation and Entrepreneurship Competition with the "PSE.IonTech Platform—a platform-level industrial mother machine for high-performance surface engineering technology equipment built on modular design."
Today, IIT Director #GiorgioMetta's interview in @Corriere. https://www.corriere.it/150-anni/26_marzo_05/giorgio-metta-mio-corriere-c6db070c-1439-41c8-9526-244468a11xlk.shtml?refresh_ce
Hope you enjoy our spin on your latest ad
Physical Intelligence proposed a method called Multi-scale Embodied Memory (MEM) that provides vision-language-action models (VLAs) with both short-term and long-term memory, enabling the model to complete tasks lasting up to 15 minutes, such as cleaning an entire kitchen or making a grilled cheese sandwich from scratch.
Why it matters: The article identifies memory mechanisms as the bottleneck for long-horizon tasks and presents a dual-scale design that stores video for the short term and language for the long term, an approach that can be transferred to other work.
Reflex headquarters hit hard by a blizzard
You saw the speed. You saw the skill. You saw the greatness. We saw it too. We also saw the measurement data behind it—precision in every track, every skill, every turn. What separates champions from dreamers is those tiny, minute distances. Do you see how much #measurementmatters?
The 28th ROBO-ONE Light will be held on March 7–8, 2026! https://www.robo-one.com/lights/index/82
A new article in Wired Italy, written by journalist Marta Abbà, recounts the evolution of JOiiNT LAB and its role in the dialogue between research and industry. ⬇️ https://www.wired.it/gallery/joiint-lab-20-laboratorio-robotica/ __ #meetIIT
repeat(pick, scan, place, scan)
"Street Dance of China" — turning lightness into weight, and rhythm into impact. This is a head-on collision between metal and beat. 🤖🦾 This Spring Festival, watch PNDbotics Adam bring a different kind of heat.
At year-end, Linear Capital and its founder and CEO Wang Huai received industry awards from several institutions. Wang Huai was named to Global Entrepreneurship Week's "2025 Angel Investors TOP30," CVSource's "2025 Best Early-Stage Investors TOP50," and Jiazi Guangnian's "Early-Stage Tech Catchers TOP30."
🔥Day 2 at CES! Thanks to everyone who visited the M.AX booth! If you see the ROBOTIS AI Worker, don't miss the chance to shake hands with it 🤝 and experience embodied intelligence up close! 📍LVCC North Hall, Booth 8871 #ROBOTIS #CES #AIWorker #PhysicalAI #Robot #Humanoid #Robotics
Please continue to support us this year as well! Wishing everyone a pleasant, peaceful, and happy year.
Growth and new frontiers: a PNDbotics chronicle
Happy holidays from the Sunday team! 🎄 Bonus points if you find the hidden Memos
Merry Christmas, from Santa 🎄🎅 #ai #robot #christmas #christmastime #holidays
Santa Claus, but make it scalable 📈🎅
It's looking very Christmassy already 🎄 An early holiday greeting from IntBot: wishing you warmth and good cheer, with hot cocoa by your side! #christmas #christmaseve #robot #humanoid #ai
Although only some photos are available, photos from the 32nd KONDO BATTLE have been uploaded recently. Please take a look! https://kondo-robot.com/archives/22308
We found an emergent property of VLAs such as π0/π0.5/π0.6: as pre-training scale increases, the model learns to align human videos with robot data. This gives us a simple way to use human videos. Once π0.5 has learned to control a robot, it can naturally learn from human videos.
Physical Intelligence research found that in generalization scenarios where only human demonstrations appear, fine-tuning π0.5 with human first-person videos mixed into robot data alone improved policy performance by about 2x, without using any explicit transfer learning mechanism.
Why it matters: Experiments show that merely mixing human videos into fine-tuning improves the policy by about 2x on scenarios with purely human demonstrations, and this capability emerges as pretraining robot data scale grows.
This is the best compliment an orange robot could receive. Your daughter is welcome to come see the Memo unit in person anytime.
Sourccey is going to college! 😊
Mentee Robotics released a behind-the-scenes video showing how two MenteeBots combine perception, planning, and control in real time during a pick-and-place box logistics workflow; the left side of the frame shows the head-camera view, alongside the robot's "thoughts" and decision outputs. The clip comes from a complete 18-minute unedited autonomous run by two MenteeBots in the same setup.
If a robot doesn't have a bit of OCD, can it still be called a robot?