If you take the average Google AI project and the average AI researcher in academia, the Google project is going to be more useful to humanity 99/100. Maybe at

If you take the average Google AI project and the average AI researcher in academia, the Google project is going to be more useful to humanity 99/100. Maybe at the very high end application is "beneath" exploration, but even then the puritans in the theoretical world seem the have the most impact subsequently applying their work (academia or not). Google and Facebook have the money to fund interesting projects with long time horizons. Does academia? You could argue not as much, it's about writing incremental papers as opposed to real world impact unless you have grinded long enough (or are at the top) to earn long-term measurement.