NEW SKILL: This team take humanoid parkour to the next level! This project involves @zhenkirito123, @Yuanhang__Zhang, @pabbeel, @carlo_sferrazza, @GuanyaShi, and others. Called Perceptive Humanoid Parkour (PHP), it lets a Unitree G1 humanoid (29 DoF, 1.3 m) run long-horizon, vision-based parkour and decide on its own whether to step over, climb onto, vault, or roll off obstacles, using only onboard depth sensing (a 30 Hz camera) plus a discrete 2D velocity command from the operator. It builds the motions by motion matching, a nearest-neighbor search in a 27-dimensional feature space (imported from the video-game industry) that stitches retargeted atomic human clips into long-horizon kinematic trajectories through a shared Locomotion to Skill to Locomotion manifold. It then trains per-motion RL tracking experts (with a privileged height scan) and distills them into one depth-based multi-skill student via DAgger combined with RL. It replaces hand-scripted skill sequencing and operator-triggered skill selection. The onboard depth-only student holds 0.95 to 1.00 success across obstacle heights in simulation, while a naive end-to-end depth policy collapses from 0.95 to 0.07/0.08 at 58 and 76 cm. The upside comes from distilling privileged-height-scan experts and from motion-matching composition, not from training onboard from scratch. The operator sends only a coarse 2D velocity (W/A/D keys), and the robot itself picks which skill to run from the depth image. The training data contains only single-obstacle traversals, so the ability to chain skills across a multi-obstacle course is generalization the policy was never explicitly taught. Transitions are constructed in kinematics, not rediscovered by learning. Every skill enters and exits through a shared locomotion manifold, so no hand-captured skill-to-skill transition clip is needed. Motion matching densifies a sparse library into smooth long-horizon trajectories before any RL. The entire parkour library is only about 66 seconds of mocap, and each skill is just a few seconds. Ablating approach-distance variety ("Extreme Distances") drops the 76 and 94 cm climbs to 0.62 and 0.64, and "Half Density" drops the 76 cm climb to 0.32. Motion matching's job is to manufacture stride-phase and approach variety out of seconds of data. It clears a 1.25 m wall (96 percent of its 1.3 m height) in 3.63 s, and a cat-vault peaks at 3.41 m/s clearing a 0.4 m by 0.5 m obstacle in 0.8 s. A single 30 Hz depth stream drives both traversal and runtime re-planning, shown by displacing obstacles about 0.5 m mid-run.

Destin Daniel Cretton says they looked at parkour GoPro footage as inspiration for Spider-Man's web swinging in ‘SPIDER-MAN: BRAND NEW DAY’. “We were able to now look at YouTube footage of wingsuiters & parkourers… and analyze what angles make your heart drop, that make you feel like you’re right there with them.” (Source:

JUST IN: FED RATES REMAIN UNCHANGED AT 3.50–3.75% Kevin Warsh just looked America dead in the eyes and said “nah, we’re good” while oil’s doing parkour off Iranian missiles, three regional Fed presidents are screaming for a hike like the building’s on fire, and Trump is somewhere yelling “LOWEST RATES IN THE WORLD” into a gold-plated megaphone. Nine-to-three. Three grown adults voted to raise rates and got overruled so the official statement could still end with the purest corporate-copium sentence ever written: “The Committee will deliver price stability.” Deliver it where? FedEx? Because it’s not showing up here. We’re in a society so advanced that “strong productivity” means robots are writing the earnings calls while the rest of us fight over $6 gas and a 30-year mortgage that now requires a second kidney as down payment. Capital markets are having a full nervous breakdown, the 10-year is climbing like it’s late for a meeting with God, and the Fed’s response is putting on noise-canceling headphones and humming. Geopolitics is just vibes now. A shooting war can jack energy prices, the statement shrugs “supply shocks,” and everyone goes back to arguing about whether the AI boom is going to save us or just make the unemployment charts look prettier while we all starve in climate-controlled server farms. This is peak late-empire comedy. The people who run the money printer are playing chicken with reality, the people who run the country are live-tweeting about GDP targets that exist only in PowerPoints, and the rest of us are refreshing the chart like it’s going to start making sense. We’re not a serious country. We’re a reality show with a central bank. And the central bank just voted to keep the laugh track running

Jean Guesdon, a lead developer behind Assassin’s Creed Unity, says the game is “one of the most underrated games in the Assassin’s Creed series,” even though its launch was a disaster. When Unity released in 2014, it was full of bugs, glitches, and performance problems. Ubisoft released several updates to fix it, apologized to players, and made the Dead Kings DLC free. Looking back, Guesdon says the team tried to do too much at once. -They built a huge version of Paris -added thousands of NPCs on screen -created a new parkour system -let players enter many buildings -introduced four-player co-op. He admitted that combining all those features made development very difficult. Now that most technical issues have been fixed, many fans see Unity as one of the best-looking Assassin’s Creed games Via:gamesradar

HappyOyster 1.0 is now live! Happy Oyster 1.0 is an open-ended world model product for real-time world creation and interaction. Create your world now at — let's explore together! Directing: Real-time interactions: Chat with virtual companions—every prompt changes the experience. Rewrite story: Pause, rewind, and generate a new path whenever you want. More ways to play: Virtual pets, dress-up, mystery boxes, and hidden interactions waiting to be discovered! Adventure: Explore extraordinary places: From deep ocean floors ruins to oil paintings or surreal dreamscapes. Feel the freedom of movement: Skate, parkour, and wingsuit through dynamic worlds. Open-world interaction: Move freely with WASD controls, jump, hide and battle enemies—just like playing a game! Limited-time rewards: Get FREE credits daily until July 17! Start exploring: The world is your oyster. Open it.

NEW RESEARCH: You can play soccer with your Unitree G1 Humanoid! This project involves @PeterYChong, @YuhangLu2000, @sephy_li, and others. It is a collaboration between OpenDriveLab, the University of Hong Kong, the Chinese University of Hong Kong and Archon Robotics. Called RoboNaldo, it is a reinforcement-learning policy that makes a Unitree G1 humanoid (29 DoF, 35 kg) shoot a soccer ball, both stationary free-kicks and one-touch shots on a moving ball. It is trained by a three-stage motion-guided curriculum in Isaac Lab (4096 parallel environments): - Stage 1 imitates a single human-kick reference (retargeted from a human video via GVHMR and GMR) with no ball or task reward - Stage 2 adds the ball, a target and shooting rewards while randomizing the ball spawn - Stage 3 adds a high-level locomotion command and a kick trigger so it can walk up and volley Perception is onboard and egocentric: a head-mounted Livox MID-360 LiDAR and a chest RealSense D435 track a retro-reflective ball (Kalman-fused), and an AprilTag (or a fixed coordinate) marks the target, all at 50 Hz with no motion capture. What is interesting imho: if you remove Stage 1 (imitating one retargeted human-kick video), the alive rate collapses to 1.2 percent, it topples on nearly every attempt. The human prior's real job is whole-body stability through a violent, high-impulse motion, not style; the hard part of a robot kick is not aiming but not tipping over from your own swing The onboard, no motion capture perception still needs a retro-reflective ball and an AprilTag goal. Ball localization is a head Livox MID-360 LiDAR plus a chest RealSense IR camera tracking a retro-reflective ball, fused by a Kalman filter. The target is read from an AprilTag or hardcoded as a fixed coordinate. It shoots hard but sub-human and only roughly accurate. Peak ball velocity is 13.10 m/s, which it reports as 59 to 71 percent of a professional open-play shot (about two-thirds of a real Ronaldo, fitting the pun). Stationary accuracy is 0.73 m average error from 3 m, with 31.5 percent of shots inside 0.5 m and 80.6 percent inside 1.0 m. The entire motion prior is one human video. There is no kick dataset and no mocap suit: a single human-kick clip retargeted to the 29-DoF G1 via GVHMR and GMR, then curriculum RL across 4096 Isaac Lab environments deviates from it for aim, contact and timing. It is the same one-demonstration-plus-sim-compute recipe as LadderMan and the perceptive-parkour projects I talked about previously. Also fun fact: it is an autonomous-driving lab pivoting to humanoid soccer. RoboNaldo comes from OpenDriveLab (the HKU end-to-end self-driving / UniAD group) with CUHK and the startup Archon Robotics, all Hong Kong based.