I keep thinking we are just dumping piles of “junk” in space, learning nothing from our past. At least, super interesting possibilities.
By 2019 [Planet Labs] are likely to be taking a million high-resolution photographs a day, covering anywhere on the globe that anyone might want to photograph, every 24 hours.
Now, machine-learning means that the sudden surge in the volume of images from private CubeSat fleets is manageable. Algorithms can ask questions such as how many of those Brazilian favela dwellings have new roofs? What percentage of roads in Uganda are in good shape? Economists are now frantically scrambling to sharpen their skills in handling these tools.