Ironically, Yann is kind of known in academic circles for being the guy working on boring, unsexy stuff (deep learning) when all the other academics were doing sexy cutting edge SVMs and stuff. That’s probably why he got the Meta job in the first place.

Yeah, being a trendsetter pretty much means you had to be working on something boring and unsexy at first. I can't really remember the last time a major innovation came because someone knew exactly what to invent, everyone agreed it was important, and we all ran full speed ahead on it.

Perhaps the atomic bomb? Not being alarmist, but that seems like one of the few instances in modern history. I also thought of the space race, but even rockets as an innovative concept started centuries beforehand as a tool of war.