Video: New household humanoid robot trained on 10 million chores unveiled in US

Sunday, a Mountain View startup founded by Stanford-trained roboticists Tony Zhao and Cheng Chi, has unveiled Memo, a household robot designed to assist with everyday chores by learning directly from real-world domestic behavior. Unlike many home robots trained in controlled lab settings or industrial environments, Memo’s development is grounded in an extensive dataset collected from over 500 households using Sunday’s patented Skill Capture Glove. This wearable device recorded nearly 10 million episodes of actual household routines, enabling Memo to perform complex, multi-step tasks such as clearing tables, running dishwashers, folding laundry, sorting shoes, and brewing espresso—activities that typically challenge robots trained only in structured environments. Memo’s design emphasizes stability and approachability over humanoid realism. It features a wheeled base for mobility and balance, a central column that adjusts torso height, two arms, and a soft, silicone-clad exterior with a retro-futuristic look reminiscent of Baymax from Big Hero 6. This design prioritizes safety and blending into home