super-intelligence.lol Domain for sale: $2,000 → email
Share unit 003 · October 14, 2026

One video, then seven in a row

Google's new model learned cornhole from a raw video of a human playing, then wrote its own simulation code.

Flat cartoon of a backyard. A large TV on a stand shows a paused video frame of a simple cartoon human mid-throw. On the lawn, a two-legged humanoid robot stands in the same pose with a bean bag in its hand. Beside it is a wooden cornhole board with four bean bags in the hole, one more bean bag lies on the grass, and a small chalkboard shows the digit 7.
Illustration of the Oct 13 demo. The television is invented; the one-video demo and the seven shots are Google's reported test. (Illustration: gpt-image-2.5-flare)

Google's Gemini 4 Argon learned cornhole by simply watching one raw video of a human playing. In an internal test, Argon wrote its own practice simulator from that video and worked out the body movements. Google DeepMind says Argon then taught Apptronik's Apollo humanoid to play.

Argon is the model that learned, and Apollo is the robot that threw. In a real-world trial, Apollo quickly sank seven consecutive shots, according to Google's blog. We have watched plenty of cornhole on TV and still miss the hole.

Sources: Google blog (Oct 13), Google DeepMind on X (Oct 13)

Current as of Oct 14, 2026. The image is an AI-generated illustration. Not affiliated with anyone.

← super-intelligence.lol