
Humanoids daily
@humanoidsdaily • 10,322 subscribers
Humanoids Daily brings you the latest developments in robotics, with a special focus on humanoid robots and intelligent machines. Newsletter for weekly updates.
Shorts
Videos

Meta's Chief AI Scientist Yann LeCun offers a critical take on the humanoid robot boom. Speaking at MIT, LeCun claimed the "big secret" of the industry is that current companies "have no idea" how to make their robots "smart enough to be generally useful." He argues that while humanoids can be trained for narrow manufacturing tasks, a truly autonomous domestic robot is impossible without fundamental AI breakthroughs. For LeCun, this means moving beyond current generative models and toward "world model planning-type architectures"—systems that can learn to understand and predict the physical world. The future of these billion-dollar startups, he says, depends entirely on this next wave of AI research.
Humanoids daily1,443,530 просмотров • 9 месяцев назад

The response to Boston Dynamics is here. 🤖🍻 BD's Atlas gripped the fridge directly, managing "complex contact points." AGIBOT's X2, powered by their new AGILE Foundation Model, achieved the same result by adding custom handles and custom end effectors. Different approaches, same cold beverage delivered!
Humanoids daily64,113 просмотров • 1 месяц назад

The man who put 50 million Roombas in our homes is back with a new vision for "Artificial Life". iRobot co-founder Colin Angle’s new startup, Familiar Machines & Magic, just unveiled "Familiars"—expressive, soft-coated quadrupeds designed for human connection rather than factory labor. With 23 degrees of freedom and strictly on-device AI, these robots aim to conquer the $2.5 trillion "emotional work" frontier without the privacy risks of the cloud.
Humanoids daily82,014 просмотров • 2 месяцев назад

Figure CEO Brett Adcock says his "major competitors" are using tele-op (human operators) in their videos and calls it "deceiving." His analogy: "It's like if a self-driving car pulled up... and we found out there was some guy from Tennessee driving it."
Humanoids daily174,114 просмотров • 8 месяцев назад

A notable strategic pivot from Xpeng's AI Day: The 'Iron' humanoid is being designed for deep customization, including "different body shapes and sexes". CEO He Xiaopeng detailed plans for "bionic muscles," "full coverage soft skin" , and options to "choose the sex", comparing it to customizing a car. The stated goal is to make the robot "warmer and also more intimate". A significant departure from the neutral, sexless designs seen from most competitors.
Humanoids daily104,932 просмотров • 8 месяцев назад

High-speed running at 13 km/h and human-like agility: KAIST’s v0.7 is a masterclass in in-house hardware optimization. 🧵
Humanoids daily54,454 просмотров • 4 месяцев назад

Reliability is a ground-up problem. 🤖 Figure CEO Brett Adcock demonstrates why the Figure 03 needs to master the "burpee". From low-battery safety to emergency recovery, the ability to transition to the floor and back up safely is a cornerstone of Figures "Never Fall protocol". Achieving this requires a significant range of motion in the hips and legs.
Humanoids daily37,160 просмотров • 2 месяцев назад

PNDbotics has released new footage of their 'Adam' humanoid, claiming it to be the first full-size humanoid robot to successfully scale a 1-meter (approx. 3.3 feet) wooden box. It's interesting to note that PNDbotics leans directly into a "future warrior" angle in their messaging, which certainly fits the slightly intimidating, raw mechanical aesthetic on display here.
Humanoids daily14,517 просмотров • 29 дней назад

This clip shows 1X's core design philosophy. Instead of the heavy, rigid harmonic drives common in robotics, NEO uses a complex, tendon-driven system. You can see the layers: from metal actuators, to the "3D Lattice Polymer" mesh, to the final knit suit. This architecture is the key to its 66 lb weight, 22 dB noise level, and 154 lb lift capacity.
Humanoids daily85,587 просмотров • 8 месяцев назад

Disney is doing what only Disney can: spending massive R&D budgets to make robots "lovable." 💸🤖 By solving the physics of character animation with Olaf, they are creating the blueprint for mainstream acceptance. Here is how startups like Cartwheel, Sunday & Tangible are following that lead to get robots into our homes. 👇
Humanoids daily60,224 просмотров • 7 месяцев назад

Would you like one of these? 🤔 A Chinese company called Luvbotics just announced "RUMI" (小伴), a 65cm humanoid companion robot. According to their site, the specs include: ✦ Full-Body Touch: Supposedly differentiates between a hug, a tickle, and a hit. ✦ Unique Voices: Claims every single bot gets a distinctly generated voice. ✦ Evolving Personality: Behaviors and skills unlock over time based on your interactions. ✦Autonomy: Indoors it roams independently; outdoors it follows you.
Humanoids daily14,269 просмотров • 1 месяц назад

Sharpa Robotics officially launched its dexterous hand, the SharpaWave, on October 16, moving it into mass production. The hand, which impressed many at ICRA 2025, is built around what the company calls a "Dynamic Tactile Array" (DTA) for high-fidelity sensing. Key specs from the official brochure are noteworthy: • Human-like design: 22 degrees of freedom in a 1:1 human-sized form factor. • Advanced sensing: The DTA features >1000 tactile pixels per fingertip, 20 N of fingertip strength with a movement speed of >4 Hz across all gestures. With out-of-the-box support for ROS and MuJoCo simulation, Sharpa is making a clear play for the robotics R&D community.
Humanoids daily66,083 просмотров • 9 месяцев назад

Robotics has a massive, silent bottleneck. It isn’t just data collection—it’s the brutal 1x speed of the physical world. Genesis AI Genesis AI just unveiled Genesis World 1.0, and they are attempting to turn the notorious Sim2Real gap into a pure compute problem. Evaluating a robotics foundation model across edge cases usually means hundreds of hours of physical lab testing. With Genesis World 1.0, what traditionally takes nearly a week of continuous, real-world operation is being compressed into 30 minutes in simulation. What makes this different from just dropping a robot model into an off-the-shelf game engine? 1️⃣ Nyx Renderer: A custom, real-time path-traced engine rendering noise-free 1080p frames in under 4ms. Game engines use rasterization tricks that confuse AI; Nyx uses physically accurate multi-bounce lighting so the model's "eyes" see exactly what real sensors see. 2️⃣ Quadrants Compiler: A custom Python-to-GPU compiler to run heavily parallelized multi-physics simulations (rigid bodies, fluids, deformables) natively across architectures. 3️⃣ Evaluation First: They aren't rushing to train on synthetic data. They are using this purely for closed-loop evaluation to perfect the physics first, currently claiming an impressive 89% correlation with real-world hardware tests. If the industry can accurately evaluate models in simulation without the physical world bottleneck, humanoid development stops moving at wall-clock time and starts scaling with compute.
Humanoids daily17,240 просмотров • 1 месяц назад

The biggest bottleneck for General Purpose Robots isn't hardware, it's data. Enter X-Humanoid: A pipeline from Mike Shou and team that "robotizes" huge human video datasets for training. The Key Insight: It solves the "visual embodiment gap." 🧠 Robots learn faster when the training pixels match their own hardware (metal & joints) rather than human skin and fluid motion. Read more here:
Humanoids daily45,350 просмотров • 7 месяцев назад

A rare glimpse under the textile "clothes": 1X’s new factory reveal offers a peek at the slim metallic skeleton powering NEO. While competitors often lean into rigid, heavy frames, 1X is betting on this lightweight, tendon-driven anatomy to keep NEO at ~30kg for "meaningfully safer" home interaction.
Humanoids daily19,994 просмотров • 2 месяцев назад

The "sim-to-real" gap just got a massive 2,300m² bridge. 🤖🏗️ NEURA Robotics and TUM are launching the TUM RoboGym at Munich Airport—Europe’s largest training center for Physical AI. Starting mid-2026, a fleet of humanoid robots will begin real-world training to fuel the Neuraverse ecosystem with high-quality, sovereign data.
Humanoids daily25,560 просмотров • 4 месяцев назад