The next great tech leap may be built on videos of people folding laundry
This image was generated by AI and may not depict real events.
Companies and researchers are gathering data for robot training by recording videos of people doing household chores, with gig workers earning up to $25 an hour. This data is crucial for adapting artificial intelligence to new tasks and enabling robots to help with domestic work.
Robots need data to learn household chores. Start-ups, including those backed by Tesla CEO Elon Musk, are gathering videos of people doing tasks like folding clothes. DoorDash has joined this effort, paying gig workers $25 an hour to record themselves doing chores. The data is used to train AI models, which get better with more data. Researchers hope that 'scaling laws' apply to robotics, just like they do to text and image AI models. Currently, there's no large repository of relevant robot data, making human videos a valuable resource. Researchers are exploring ways to use this data to teach robots, including combining it with more expensive 'teleoperation data'.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.