Image created with gemini-2.5-flash-image with claude-sonnet-4-5-20250929. Image prompt: A polished chrome humanoid robot sitting at a modern table, blowing out two glowing candles on a sleek metallic birthday cake, surrounded by floating confetti in rich reds, deep blues, and whites against a slate grey background with dramatic high-contrast lighting reflecting off its surface. Cinematic photograph with sharp focus and celebratory energy.
Helix is exceeding what I thought was possible in my home To put a robot in every household, we need massive amounts of data – today we’re launching Project Go-Big This is accelerated with Brookfield who owns over 100,000 residential units https://x.com/adcock_brett/status/1968703218482839987
We just wrapped up a huge week at Figure – here’s what we announced: 1/ Announced >$1B in Series C. This gives us the strongest balance sheet in humanoid robotics – critical to scaling Helix (our AI) and BotQ (robot manufacturing) 2/ Launched a partnership with Brookfield, who https://x.com/adcock_brett/status/1969781540805914856
“Figure announces Project Go-Big: Internet-Scale Humanoid Pretraining and Direct Human-to-Robot Transfer The company has achieved a new milestone for the Helix AI model: after training exclusively on egocentric human data, Figure robots can now navigate real-world spaces from https://x.com/TheHumanoidHub/status/1968720346141651390
How does a (then) 2-year-old startup with NO commercial product raise $675M from Jeff Bezos, OpenAI & NVIDIA? 🤯 This is the untold story of @Figure_robot. And @adcock_brett. It’s a masterclass in building a reality-distortion field that convinced the world’s smartest https://x.com/IlirAliu_/status/1969748812437549265
The Information: “Humanoid robotics startup 1X has told investors and employees it is trying to raise as much as $1 billion, according to three people who spoke to CEO Bernt Børnich. The decade-old company is aiming for a valuation of at least $10 billion, Børnich said, or https://x.com/TheHumanoidHub/status/1970493306401652883
Skild AI’s omni-bodied brain, trained on 100,000 diverse simulated robots for 1000 years, enables remarkable real-world adaptability. In-context adaptation allows the brain to discern the robot form and adapt to extreme changes like chopped limbs or walking on stilts. https://x.com/TheHumanoidHub/status/1970981739200909811
Gemini Robotics 1.5 brings AI agents into the physical world – Google DeepMind https://deepmind.google/discover/blog/gemini-robotics-15-brings-ai-agents-into-the-physical-world/
New Gemini Robotics 1.5 models will enable robots to better reason, plan ahead, use digital tools like Search, and transfer learning from one kind of robot to another. Our next big step towards general-purpose robots that are truly helpful — you can see how the robot reasons as https://x.com/sundarpichai/status/1971244716046872577
Talk to robots! Today we’re releasing our SOTA Gemini Robotics 1.5 model showing the power of using our multimodal Gemini models as a base, so it can understand & reason about the physical world. Robotics will be massive in the future – super excited by our pioneering work here!”” / X https://x.com/demishassabis/status/1971292365592854602
We’re making robots more capable than ever in the physical world. 🤖 Gemini Robotics 1.5 is a levelled up agentic system that can reason better, plan ahead, use digital tools such as @Google Search, interact with humans and much more. Here’s how it works 🧵 https://x.com/GoogleDeepMind/status/1971243947792925005
This is Ray3. The world’s first reasoning video model, and the first to generate studio-grade HDR. Now with an all-new Draft Mode for rapid iteration in creative workflows, and state of the art physics and consistency. Available now for free in Dream Machine. https://x.com/LumaLabsAI/status/1968684330034606372
Veo 3 = Zero-shot video reasoner • Trained on web-scale video, shows broad zero-shot skills (perception → physics → manipulation → reasoning) • New “Chain-of-Frames” reasoning = visual analogue of CoT • Big jump Veo2 → Veo3: edits, memory, symmetry, mazes, analogies • https://x.com/arankomatsuzaki/status/1971042970800701809
What used to take hours in After Effects now takes just ONE prompt. Nano Banana, Seedream 4, Wan 2.2, Runway Aleph et al are pioneering instruction-based editing — collapsing complex VFX pipelines into a single, implicit step. Here’s everything you need to know in 10 mins: https://x.com/bilawalsidhu/status/1970915228536947026
Did you see that the Agent Research Environment is MCP compatible? -> using any MCP tools with any agent is now completely trivial! Check it out! We’ve used an LLM agent to 1) move a robot arm remotely 2) depending on real time web search results! 😀 How to in thread ^^ https://x.com/clefourrier/status/1970394602592182627
VisualMimic enables humanoid loco-manipulation via sim-to-real. Using egocentric vision, a low-level keypoint tracker and high-level policy achieve zero-shot transfer for tasks like box pushing, generalizing across locations and lighting conditions. https://x.com/TheHumanoidHub/status/1970945814840410494
EngDesign: Benchmarking LLMs toward Engineering AGI • 101 tasks across 9 domains (OS, circuits, robotics, etc.) • Simulation-based eval (SPICE, FEA, etc.), not just Q&A • Iterative refinement boosts pass rate (o3 → ~60%) • Error analysis: 111 failure modes identified https://x.com/arankomatsuzaki/status/1970326076271513805
Building the largest humanoid pretraining set”” / X https://x.com/adcock_brett/status/1969097615946985574
It’s been incredible to see neural networks working so well on our humanoid robots Humanoids are crazy complex – an individual motor can rotate 360 degrees and you have 40+ joints. If you do the math, that means more possible robot states than atoms in the universe Figure has https://x.com/adcock_brett/status/1970538945663640011
Old pic, probably 10 years ago! If I had to summarize the most important attribute for hiring / good talent it’s deeply caring Matters much more than intelligence, agency, aptitude. Somebody who cares deeply will find a way to solve for all this”” / X https://x.com/adcock_brett/status/1969977690284584988
We’re now moving beyond models that react to single instructions and creating systems that can truly tackle problems in a general way – on the path towards solving AGI in the physical world. Developers can now use Gemini Robotics-ER 1.5 via the Gemini API in @GoogleAIStudio. https://x.com/GoogleDeepMind/status/1971243970953879643
🤯 A robot learned to handle a raw egg…using just 15 minutes of human video and a $10 glove: [📍bookmark for later] Feel the Force (FTF); a framework for teaching robots force-sensitive manipulation using natural human demos. Most policies treat the gripper like an on/off https://x.com/IlirAliu_/status/1968786117836132420
Adam humanoids from PNDbotics are growing in numbers. https://x.com/TheHumanoidHub/status/1969093688744686038
Anti-Gravity”” mode engaged on Unitree G1: great stability and fall recovery while performing action sequences. Also, some impressive backflips and landings at the end. https://x.com/TheHumanoidHub/status/1970137910637998244
Any2Track tracks diverse motions under various disturbances. The two-stage RL framework learns to track contact-rich human motions from motion capture datasets and adds adaptability to disturbances via dynamics-aware model prediction. https://x.com/TheHumanoidHub/status/1969276765337239690
Calvin-40 humanoid – developed by French company Wandercraft. https://x.com/TheHumanoidHub/status/1968539670566027509
Chinese Optimus knockoff, Kepler Robotics, claims to have commenced mass production of the K2 Bumblebee humanoid. https://x.com/TheHumanoidHub/status/1969106087942017137
For decades, robotics has been broken. It’s been dominated by academics in ivory towers (think MIT, Boston Dynamics) building hyper-complex machines for controlled labs. The result? 90% of robotics startups fail to scale. They build cool tech, not viable businesses. [Hyundai https://x.com/IlirAliu_/status/1969748816153702724
From demos to mass production. Thousands are already on order… @BotKepler just started shipping the K2 “Bumblebee.” Based in Shanghai, focuses on humanoids for industrial and general-purpose use. Their latest model pushes the field toward real commercial deployment. ✅ K2 https://x.com/IlirAliu_/status/1969290076631830727
HDMI (HumanoiD iMitation for Interaction) is a framework enabling humanoid robots to learn whole-body object interaction skills from monocular RGB human videos. It extracts and retargets human poses and object trajectories using GVHMR and LocoMujoco, building reference datasets https://x.com/TheHumanoidHub/status/1970584868431228933
HMND 01 Alpha, developed by UK-based Humanoid. Height can extend up to 7’3″” | 7.2 km/h speed | 15 kg payload | 29 active DOF, excluding end-effectors. https://x.com/TheHumanoidHub/status/1968750404054736962
interestingly these scratch-your-back deals have apparently become commonplace. an overvalued robotics company (that everyone’s confused as to how they’re raising so much) apparently raised a gazillion dollars through similar deal structure”” / X https://x.com/soumithchintala/status/1970465637110612477
K-Scale delivers the first K-Bot humanoid. They’ve paused accepting new orders since the second batch of K-Bots has also sold out. https://x.com/TheHumanoidHub/status/1970905390478131484
Most learning-based robot policies fail on hardware. They send slow or choppy commands. 📍That’s where CRISP comes in: CRISP is a set of compliant ROS2 controllers that make diffusion and VLA models run smoothly on real robots. ✅ Torque-based Cartesian and joint-space https://x.com/IlirAliu_/status/1968941773801009607
New Episode: Hao Cheng is Founder and CEO of Booster Robotics, a Beijing-based startup focused on affordable humanoid robots. Hao believes, like the personal computer, the availability of affordable and developer-friendly humanoid robots will drive innovation. We sit down at https://x.com/TheHumanoidHub/status/1970528352193982669
Phantom humanoid’s RL controller put through some stability tests. https://x.com/TheHumanoidHub/status/1968789672517124473
The HMND 01, featuring bipedal legs, is slated to “”come online”” later this year. Humanoid, the UK-based robotics company, showed a glimpse of the bipedal robot in a video released today. https://x.com/TheHumanoidHub/status/1968754751199387665
Unitree G1 performs a Webster flip, trained by researchers at the Beijing Institute for General Artificial Intelligence. https://x.com/TheHumanoidHub/status/1968722547999928492
Unitree G1 shenanigans https://x.com/TheHumanoidHub/status/1970177027325190385
Why do robotics teams rebuild the same tools over and over? Data pipelines, sync systems, logging… all from scratch. @stepjamUK saw this pattern for over a decade (Imperial, Berkeley, Dyson Lab). Every team spent months on plumbing before touching the real robotics problem. https://x.com/IlirAliu_/status/1970924031429005414
Why should a robot have to choose between legs, wheels, or arms? A new hexapod design switches seamlessly between all three. It even reuses legs as manipulators. Researchers present a multimodal hexapod robot for extreme environments like disaster response and planetary https://x.com/IlirAliu_/status/1970754221059215456
Most humanoid trackers break when reality pushes back. ❗️Not this one: Uneven terrain, external forces, or sudden changes throw them off. It is a two-stage RL framework that tracks diverse humanoid motions while adapting online to real-world disturbances. ✅ AnyTracker: a https://x.com/IlirAliu_/status/1969458971321782664
Why can robots do backflips but still struggle to open a drawer??? [📍 Link to project] Precise grasping and whole-body coordination make it harder than acrobatics. DreamControl takes a step toward solving this. It combines diffusion models and reinforcement learning to teach https://x.com/IlirAliu_/status/1970539603368042823
Demis Hassabis of Google DeepMind: ⦿ Humanoid form factors could be important for general and personal use in human-centric environments, while specialized robots optimized for specific tasks will be useful in production lines and laboratories. ⦿ There’ll be a sort of real https://x.com/TheHumanoidHub/status/1968507472064430330
Ran a world model robot learning reading group alongside @djkesu1 this Thursday, covering V-JEPA and V-JEPA 2 by @AIatMeta, here are some of my thoughts: JEPA is an interesting idea and borrows many ideas from the representation learning literature (a combination of SimCLR, https://x.com/stevengongg/status/1969387819920736396
