[andrewljohnson]

Well, I can definitely understand why you would make your remarks anonymously. I still don't think I'm uninformed to suggest this is not a coincidence, despite you berating me.

I was chatting with my colleague at work today, who actually stood up our landsat servers and does all of our tiling work, and he thinks that the pixel-averaging technique to de-cloud the images was an innovation that was pretty unique to MapBox. According to this article, they noticed a guy doing it in February 2013, and hired him right away to do it for them: http://www.wired.com/design/2013/05/a-cloudless-atlas/.

MapBox got to market with it in just a few months, and it's a particularly absurd argument to say this was too hard for Google's hundreds of maps engineers to implement. I heard at the State of the Map conference that it was just a few guys working on it at MapBox. Given Google already has the imagery, and the processing pipeline, I think adopting this declouding technique was actually a piece of cake.

Also, while Google has been removing clouds from their images for a long time, I think their technique has been to pick the best of many images. Their imagery used to have many more clouds in it, until all of a sudden.