Statistics from the 4.7 development cycle

If anything this proves the point that academia doesn't contribute directly to OSS. Obviously they do generate new algorithms and other good ideas, but unless someone else picks up the slack it won't turn into something useful. The fact that Linaro chose to hire someone who previously worked in academia to do the work doesn't change that, it is still someone outside of academia who picks up the academic idea and turns it into something OSS can use.