ACT-2 Achieves 99.1% Success in Unseen Homes via One-Shot Fine-Tuning
Olivier__OG · x · 2026-08-02
Sunday Robotics introduced its ACT-2 Preview and Memo robot, demonstrating the ability to learn new behaviors from a single fine-tuning example. The robot successfully applied these behaviors in unfamiliar home environments, with the company reporting a 99.1% success rate.
The core challenge in home robotics is generalization: models often fail when room layouts, objects, or lighting change. If robust, this demo signifies a major breakthrough in rapid learning and cross-scenario generalization, paving the way for home robots that can continuously adapt to new tasks.
More from Embodied
- Berkeley researchers sample humanoid robot policies directly from weight space with diffusion — berkeley_ai · 2026-10-16
- OneWM: unified causal training shows world and action modeling improve each other — IAAI233 · 2026-10-16
- Tesla's hover car isn't a flying car — compressed air can't sustain flight — tinyfool · 2026-10-16
- Scobleizer relays claim that Optimus demo will be even more mind-blowing than the road-flipping Roadster — Scobleizer · 2026-10-16
- Indie dev's AI PASSPORT wearable firmware hits 1,200+ downloads in days — ezshine · 2026-10-16
- 1X Opens California Factory for Mass Production of NEO Humanoid Robots — emmanuelvivier · 2026-10-16