Stanford's HomeBody gives frontier VLMs spatial memory and composable humanoid skills
CyberRobooo · x · 2026-09-28
The official writeup of HomeBody (Caltech + Stanford; Gio Huh, Cayden Gu, et al., advised by Karen Liu and Guy Tevet): a system that equips frontier VLMs with persistent spatial memory and a composable skill library for long-horizon humanoid tasks. In an unseen kitchen it can tidy across the room and retrieve remembered objects from underspecified requests, with no environment-specific training data. Paper and code are available.
More from Embodied
- University of Tokyo grows self-healing living human skin on robotic finger — TinfoilTricorn · 2026-09-28
- Delta Intelli launches Δ₀, a humanoid foundation model with 69-DoF whole-body control — jeasinema · 2026-09-28
- Meta's Muse AI assistant watches everything, critics call it a parasite on human life — CarissaVeliz · 2026-09-28
- UK Startup Turns Gaming Inputs Into Training Data for World Models, With LeCun Advising — nordicinst · 2026-09-28
- British Startup Turns Gamer Inputs Into Training Data for Physical-World AI — Wired AI · 2026-09-28
- AUAR ships robotic microfactories in containers to build homes up to six stories — lukas_m_ziegler · 2026-09-28