State sponsored fundamental research for one. Pretty much all of the actual computing innovation was a direct result of fundamental research that was sponsored by the state either directly or indirectly. Heck, modern compilers have a direct line back to Grace Hopper working at the naval research lab.
Commercializing stuff is great and all but there is no point in pretending it is the same or equal to the kind of innovative work that consistently happens at research universities and arguably a lot of the future is invented there even if it is not commercialized.
Also what are these dozens of other projects Facebook is currently innovating on? I do see a lot of duplicated effort across google, facebook, apple, and microsoft when it comes to AI research though. This is why I don't quite get how people can argue for commercial innovation and market efficiency all in the same sentence or paragraph when there is obvious and clear evidence to the contrary.
