Gemini 4 Argon от Google DeepMind обучила робота-гуманоида игре в корнхол по видеозаписи

Gemini 4 Argon taught a robot to play cornhole just by watching a video. It mapped the physics, coded its own simulation, and taught

Google DeepMind researchers gave it one clip of a person tossing bean bags. That was the whole lesson. Argon mapped the scene, worked out the physics, and wrote its own simulation to practise in. Then it took control of Apptronik's Apollo humanoid and kept training on the real robot. Advertisement Advertisement Advertisement Advertisement After a few warm-up throws, Apollo sank seven in a row. Google says this video-to-sim-to-real pipeline squeezes weeks of manual simulation work into a few hours. Asked to explain itself, the robot even came up with its own gestures. And it's not a one-off. From a single tabletop photo, Argon built a digital twin, generated over 4,500 simulated practice runs, and lifted a robot's real-world success rate from 56 to 75 percent. [...] Skip to main Home News Finance Sports Mail Advertisement Advertisement Return to Homepage ## Top Stories: Cornell rape case latest Katie Zacharia Social Security's 2027 COLA Roller coaster tragedy Co

# Google's Gemini 4 Argon Taught a Humanoid Robot to Play Cornhole ## Channel: Interesting Engineering Explains 39.5K subscribers 174 likes ### Description 6,994 views Posted: 2026-10-14 Google has released Gemini 4 Argon, its new frontier model, and showed it teaching Apptronik's Apollo humanoid to play cornhole from a single video of a person. Argon mapped the scene, built its own physics simulation and refined the throws on the real robot, which then sank seven shots in a row. This is one of 30+ #defense & #engineering videos we drop weekly. Get them all in the IE App: ⏩iOS: ⏩Android: 🚀 Discover deep-dive engineering stories and breakthrough technologies on Interesting Engineering: / @interestingengineeringie [...] ### Transcript [0:00] This humanoid learned cornhole from a single video and its coach was an AI meet Gemini 4 argon [music] Google's new [0:09] frontier model released worldwide on October 13th. Google deep mind researchers gave it one clip of a person [0:16] to

# Teaching a humanoid robot to play cornhole with Gemini 4 Argon ## Channel: Google DeepMind (verified) 933K subscribers 4.6K likes ### Description 798,995 views Posted: 2026-10-13 In an internal test, Google DeepMind robotics researchers instructed Gemini 4 Argon to watch a video of a human playing cornhole. Using its video understanding and advanced reasoning capabilities, Gemini 4 built a virtual simulation in MuJoCo and accurately calculated the physics and whole-body control required for Apptronik’s Apollo 2 humanoid to sink seven consecutive shots in the real world. This Video-to-Sim-to-Real pipeline compresses weeks of manual modeling into a few hours, transforming ordinary videos into an infinite training ground for physical AI. Learn more about Gemini 4 Argon: ___ [...] [0:58] Second, I looked through my two eye cameras to measure the exact 3D position of the board and built a physics simulation of my body and the board. [1:07] Third, I practiced the throw in simulation, w

We taught a robot to play cornhole using Gemini 4 Argon Just by watching a video of humans playing, Argon mapped the physics, coded its

@GoogleDeepMind Google DeepMind@GoogleDeepMind Gemini 4 Argon taught a robot to play cornhole just by watching a video. It mapped the physics, coded its own simulation, and taught @Apptronik's Apollo humanoid to sink 7 throws in a row. 00:00 6:41 PM · Oct 13, 2026·266.2K Views 284 @GoogleDeepMind Google DeepMind@GoogleDeepMind Oct 13 By turning video into a training ground for physical AI, Argon compresses weeks of manual modeling into just a few hours. ↓ Google for Developers Blog - News about Web, Mobile, AI and CloudFrom developers.googleblog.com 16 @EntryStateGod Adam Shepherd@EntryStateGod Oct 13 [...] Log inSign up ## Post Log inSign up # Google DeepMind on X: "Gemini 4 Argon taught a robot to play cornhole just by watching a video. It mapped the physics, coded its own simulation, and taught @Apptronik's Apollo humanoid to sink 7 throws in a row." @GoogleDeepMind Google DeepMind@GoogleDeepMind Gemini 4 Argon taught a robot to