1) Paying workers as little as possible isn't the best way to maxmize business profits.
2) Paying workers way above market rates doesn't attract the best talent.
3) You should pay the fabled 10X productive engineer 10 times the salary.
So... I don't know, I think the author was trying to say to pay in the middle somewhere, but got lost on the way to the point?
Paying people very high wages (he says) retains the best talent, which turns out to be a really good investment.
The businesses he mentions that overpay mediocre engineers presumably aren't good at finding the best talent, nor at firing talent that turns out not to be good.
At least, that's how I read it.
You should create a tight-knit team of highly competent people in an environment that favors productivity. This will attract the best people. Then share the profits with them (i.e. pay them well) to make sure they stay.
Of course, this may be easier said than done, and ignores the fact that not all work is interesting.