First look: This weird wearable device turns human workers into robot data collectors

Instawork is making waves in the gig economy with the introduction of a groundbreaking wearable camera system designed to transform its workers into essential data collectors for robotics training. This new device, known as Instacore, was showcased to Business Insider and features five strategically placed cameras on the head, chest, and wrists, all linked to a durable compute backpack built for an entire work shift. The primary objective of Instacore is to equip Instawork's gig workers, referred to as 'Pros,' with the ability to document a variety of tasks such as food preparation in commercial kitchens or stocking shelves in retail environments. The collected footage will serve as valuable training material for robotics companies and AI labs aiming to enhance machines' capabilities in real-world commercial settings. Instawork, which has secured over $150 million in funding from investors like Benchmark, Greylock, and Spark Capital, is keeping its customer details under wraps but asserts that it is collaborating with top-tier research institutions. This initiative comes amid a growing demand for authentic data as the focus of AI expands from virtual interactions to physical operations. Industry leaders, including Sam Altman, have identified robotics as a key area for future development, with major tech firms like Nvidia, Meta, and Tesla increasing their investment in robotics. With a community of approximately 10 million Pros at its disposal, Instawork has a unique advantage in gathering diverse data across various tasks and environments, enhancing its appeal to AI and robotics developers. "We are rolling out Instacore globally, aiming to compile what we believe will be the largest and most varied dataset for robotics training," stated Aaron Bromberg, Instawork's head of robotics and a former Amazon technologist involved in the development of the Astro home robot. During a recent tour of Instawork's Mountain View facility, Bromberg noted that workers were already donning the Instacore systems to record their assembly of more units. "It's like Instacore inception — they're capturing footage of themselves as they build the devices," he remarked. The transition into hardware marks a significant shift for Instawork, which has evolved from a labor marketplace over nearly a decade. The CEO, Sumir Meghani, recognized the potential for robotics within the gig economy after observing a trend of robotics companies utilizing Instawork's platform for staffing. "When users adapt your product in unexpected ways, it’s crucial to follow that signal," Meghani explained. He believes that robotics presents a unique intersection of physical labor, an area in which Instawork excels, and artificial intelligence. A major challenge for the company was to create a lightweight, comfortable system suitable for various job functions. The initial prototypes were too cumbersome and uncomfortable for workers. After extensive feedback, the final version strikes a balance between user comfort and the high-quality data required by robotics firms. The system's design includes two head cameras for depth perception, a chest camera for environmental context, and wrist cameras to monitor hand movements, all while keeping the total weight under three pounds. "It takes some getting used to," shared Aaron Kerchner, an Instawork Pro and current on-site supervisor at the Mountain View location. "You have to be mindful of the cables, but it becomes easier with time." Instawork is in the process of assembling hundreds of these systems on-site, with plans to ramp up production overseas. Components are sourced from both China and the US, with some parts being 3D printed in-house. Other companies are also leveraging gig workers for real-world data collection. A recent startup gained attention for offering complimentary apartment cleaning services to workers willing to wear head-mounted cameras while completing household tasks. Instawork intends to deploy thousands of Pros equipped with the Instacore system, ensuring that data collection remains voluntary and anonymized. Meghani envisions a future where physical AI, a concept popularized by Jensen Huang to describe AI systems capable of real-world interactions, will foster new job opportunities, including roles for robot trainers and technicians. "The rise of physical AI will generate new employment avenues, and our mission is to ensure our Pros remain relevant in a robotic future," Meghani concluded.