Would love to see someone have access to the "raw" photo data and try and find common combo objects (taxis parked, people with hats, people on roofs, animals in

Would love to see someone have access to the "raw" photo data and try and find common combo objects (taxis parked, people with hats, people on roofs, animals in buildings etc.) using ML like the Where's Waldo example a long time back[1] [1]: https://news.ycombinator.com/item?id=3367865 https://news.ycombinator.com/item?id=3367865