Arithmancy

Agility Robotics co-founder doesn't believe in the teleop data story. Jonathan Hurst laid it out on The Cube at the Raise Summit in Paris. 🔵 LLMs learned from internet, robots have no equivalent for movement 🔵 Teleoperation alone would take millions people and centuries to scale. His answer is simulation, not teleop. Build a digital twin, train with RL, then transfer to the real robot. He also talked about Digit's future, and how V5 drops the cage. 🔴 Digit V5 ships by the end of this year, safety is the headline feature 🔴 V5 is built to walk into unmodified, as-built human spaces without one

Everyone in the Bay is building a robot company. Almost nobody is talking about the real bottleneck. After 2 years running the Robotics Center of Silicon Valley, seeing hundreds of frontier labs and stealth startups, one pattern is undeniable: The blocker isn't model capability. It’s the sim-to-real loop. 1. Simulation ≠ Hardware Reality A VLA trained in Isaac Sim or MuJoCo hits a wall when it encounters real-world actuator limits: thermal throttling, joint backlash, and sensor drift. Even two units of the exact same robot model have completely different dynamics. 2. Teleop data is inherently off-policy Human demonstrations bootstrap your model, but your model isn't human. That distribution gap doesn't disappear with more of the same data. It only closes through real-world rollouts, failure capture, and targeted data collection. 3. "Working" is undefined until deployment Teams optimize for lab success rates, then fail on real-world edge cases: a wrinkled shirt, 4 PM glare, or a pallet 3cm off spec. Without a real-world evaluation harness, you are just guessing. 4. The solution? Treat RL² as your core loop Internally, we call this Reinforcement Learning in Real Life. Deploy ➔ capture failures ➔ evaluate against real criteria ➔ collect targeted data ➔ redeploy. Continuous learning on real hardware. 5. The Takeaway The winners in Physical AI won't just have the biggest models. They'll be the teams that can close the real-world learning loop the fastest. Closing the sim-to-real gap is a 2-year infrastructure problem. That’s why we built our stack at the Robotics Center. 6. Come hang out If you're hitting these exact walls, let’s compare notes. We host robotics builders at 90 Welsh St (SF) every Friday. Drop by.

Click-and-Drag Gripper Control is now LIVE We've been simplifying high-hardware-demand, complex simulation teleoperation through web-based keyboard-and-mouse control — and now we're taking it one step further. Direct click-and-drag gripper control is officially here. This eliminates a significant number of intermediate steps. Simply reorient your viewing plane by adjusting the camera perspective, and focus only on gripper-object contact. See it in action 👇

This is Festo BionicKangaroo. A robotic kangaroo built by Festo’s Bionic Learning Network to study one idea from nature: how a kangaroo stores energy when it lands and reuses it in the next jump. How it works: • An elastic rubber tendon acts like an artificial Achilles tendon • Pneumatic cylinders trigger the jump • Electric motors control the hips and tail • The tail balances the body during flight • On landing, the tendon absorbs kinetic energy • That stored energy is reused for the next jump The robot weighs about 7 kg, stands around 1 meter upright and can jump up to 40 cm high and 80 cm forward. It was not built as a consumer robot. It was built to test energy recovery, lightweight mechanics, mobile pneumatics and the smart combination of pneumatic and electric drives. Nature already solved efficient jumping. Festo turned that biology into hardware. Official source: Festo’s BionicKangaroo page and brochure.