Human Archive taps Indias gig economy to collect data for physical AI

Human Archive, a Silicon Valley based startup, partners with companies in India's home services, hotel, and restaurant sectors to have workers wear caps equipped with cameras. The devices collect egocentric video data from a first person point of view of everyday tasks that could be used to train robots. The company has more than 1,000 active headsets deployed across multiple locations. The startup raised 8.2 million dollars in funding from Wing Venture Capital, NVP Capital, Y Combinator, and angels from OpenAI, Nvidia, Google, Mercor, AfterQuery, BAIR, SAIL, Brad Boa, and Meta. It was founded by Samay Maini, Rushil Agarwal, Shloke Patel, and Raj Patel, who is chief executive. The founders, three from UC Berkeley and one from Stanford, have research backgrounds in robotics, hardware, and tactile data. Human Archive develops tactile gloves, a full body motion capture suit, and wrist cameras to capture synchronized motion, tactile force, and RGB-D data. It has more than 50 devices deployed and more than seven hardware products. Workers receive a base rate of 1 dollar per hour for data collection. The company works with smaller startups to offer discounted services in exchange for customer consent to recordings. Its contracts comply with India's Digital Personal Data Protection Act, with data anonymized and faces blurred. It is expanding into Southeast Asia and the United States while building a platform for wider participation in data collection. Zach DeWitt, a partner at Wing Venture Capital, said the startup has a unique advantage in synchronizing data from multiple sensors at scale and that major labs and universities are interested in experiments with the data.