It's a nice trick sure, using machine learning (I assume) to build 3d geometry from 2d satellite images, but it's nothing compared to the quality and fidelity of the images taken by low flying airplanes that Birds Eye in Bing actually used. Made house/apartment hunting a dream!
Looking at 3d google maps a bit more closely, it does seem they are capturing the sides of buildings that are not visible on the top down image. Still nowhere near the quality of Birds Eye, but I'd be interested in what other image sources they are using besides top down, and how they are captured.
It isn't machine learning, it's aerial photography shot at an angle with depth sensing.
For example, take an airplane and mount two cameras on it, one on each side, angled down at perhaps 45 degrees to capture ground imagery in each direction.
Add a LIDAR device next to each camera to capture a "point cloud" of the same area the camera is imaging.
Now you have photographic images with distance data for each pixel. You can use that to construct a 3D image that can be viewed from various angles.
(Disclosure: I work at an unrelated Alphabet company but have no personal knowledge of any of this, it's just my semi-educated guess.)
You probably don't even need the LIDAR. Since the airplane is moving, you can use the parallax between successive images to determine depth information. The technique is called "structure from motion"
I thought it had to be accomplished by satellite somehow. I looked at a very remote island in the southern Indian Ocean where I was stationed and the amount of detail in Bing Bird's eye view of the geography was astonishing. This was circa 2013. I would be very surprised if they flew a plane out there just for those islands, there are no flights or shipping routes passing by there.
Looking at 3d google maps a bit more closely, it does seem they are capturing the sides of buildings that are not visible on the top down image. Still nowhere near the quality of Birds Eye, but I'd be interested in what other image sources they are using besides top down, and how they are captured.