AI Trains on Human Experience: Robots Next?

A company named Human Archive has recently raised $8.2 million in funding, signaling Silicon Valley's growing interest in egocentric data collection. This technology involves equipping workers with cameras, similar to a viral video that recently emerged from India. Human Archive attaches cameras to workers across various environments, including homes, restaurants, and construction sites, to build extensive datasets.

These datasets are categorized into two main types: a 3-D dataset captured by visor-like cameras and a 2-D dataset focusing on hand movements from wrist-mounted cameras. The company's co-founders state the goal is to develop foundational datasets for modeling human sensimotor intelligence and understanding human interaction in the physical world at scale.

While the potential application for replacing human workers with robots is evident, Human Archive suggests its ambition extends to a broader understanding of 'embodied intelligence.' This concept integrates the brain, body, and environment to create a synergistic view of intelligence. The company believes its technology will become essential infrastructure for automating labor and advancing the study of human intelligence.