That is one of the very important points a lot of people are missing here, and a key argument of the article: a fast rendering engine was a requirement for quak

That is one of the very important points a lot of people are missing here, and a key argument of the article: a fast rendering engine was a requirement for quake to be as popular as it was and it was a product of a brilliant engineer doing something innovative. More often then not, innovation comes from a reduced number of people doing things differently, and it's very hard to assemble a team of engineers which is willing or even capable of doing this sort of work. This also comes back to the "decision shielding" from working different hours: it's very common for managers or people averse to risks to try to stick to solutions they know, even if they are suboptimal or do not fit the problem very nicely. Anyone capable of doing more will soon realize they are spending most of their time solving mismatches between what solution was used/implemented and the problem they were trying to solve. If they don't find an alternative or a way around these limitations, they will leave.