Data for the Real World

AI is superhuman on code, language and images, but the physical world is still measured by sparse sensors built for humans. Build new ways to collect dense physical-world data, then model and control real systems.

Better foundation models and falling sensor costs make dense physical-world data collection feasible. Examples: Gecko Robotics inspects hard-to-reach places with robots; Sorcerer flies autonomous weather balloons whose data improves US weather forecasts.

The opportunity: Energy, agriculture, logistics and construction rely on limited data and intuition-based models. More real-world data enables precise models, and once you can model a system you can control it, from steering hurricanes to reversing desertification.